Discussion:
Can Lady Heather Keep Computer Clock On Time?
Brooke Clarke
2010-07-29 18:04:07 UTC
Permalink
Hi:

I have a friend who's setting up an observatory and the PC that controls
the telescope needs to have an accurate clock.
The telescope can point to within arc seconds of a star and that implies
that the computer clock needs to be within 50 ms.
If he does a Windows NTP sync first thing in the evening after a few
hours there's too big an error. Running the observatory is like flying
a 747, i.e. there's a lot to do and the time needs to be handled
automatically, not by manual NTP updates.

I know that TAC32 and a Motorola GPS will do this and you have a lot of
options of how it does it. For example at a specified time interval or
when the computer clock differs from GPS by a specified amount of time
(50 ms in this case).

What are the options in LH? I ask because it's a lower cost option then
the Motorola GPS plus TAC32 option or building a NTP time server.

--
Have Fun,

Brooke Clarke
http://www.PRC68.com
Bob Camp
2010-07-29 18:13:18 UTC
Permalink
Hi

My observation is that the displayed clock on LH can be off by > 10 seconds.
That's with the current beta code and Windows 7 on a quad processor machine
or under XP on a dual core machine.

Bob

-----Original Message-----
From: time-nuts-bounces-***@public.gmane.org [mailto:time-nuts-bounces-***@public.gmane.org] On
Behalf Of Brooke Clarke
Sent: Thursday, July 29, 2010 2:04 PM
To: Discussion of precise time and frequency measurement
Subject: [time-nuts] Can Lady Heather Keep Computer Clock On Time?

Hi:

I have a friend who's setting up an observatory and the PC that controls
the telescope needs to have an accurate clock.
The telescope can point to within arc seconds of a star and that implies
that the computer clock needs to be within 50 ms.
If he does a Windows NTP sync first thing in the evening after a few
hours there's too big an error. Running the observatory is like flying
a 747, i.e. there's a lot to do and the time needs to be handled
automatically, not by manual NTP updates.

I know that TAC32 and a Motorola GPS will do this and you have a lot of
options of how it does it. For example at a specified time interval or
when the computer clock differs from GPS by a specified amount of time
(50 ms in this case).

What are the options in LH? I ask because it's a lower cost option then
the Motorola GPS plus TAC32 option or building a NTP time server.

--
Have Fun,

Brooke Clarke
http://www.PRC68.com



_______________________________________________
time-nuts mailing list -- time-nuts-***@public.gmane.org
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Eric Garner
2010-07-29 18:22:18 UTC
Permalink
with the assumption that he seems to be fine right after a sync,
(seemingly supported in the OP) reconfiguring the windows time service
to update frequently seems to me to be the easiest answer.
under xp you can do so thus:
http://support.microsoft.com/kb/314054

_eric


On Thu, Jul 29, 2010 at 11:13 AM, Bob Camp <lists-***@public.gmane.org> wrote:
> Hi
>
> My observation is that the displayed clock on LH can be off by > 10 seconds.
> That's with the current beta code and Windows 7 on a quad processor machine
> or under XP on a dual core machine.
>
> Bob
>
> -----Original Message-----
> From: time-nuts-bounces-***@public.gmane.org [mailto:time-nuts-bounces-***@public.gmane.org] On
> Behalf Of Brooke Clarke
> Sent: Thursday, July 29, 2010 2:04 PM
> To: Discussion of precise time and frequency measurement
> Subject: [time-nuts] Can Lady Heather Keep Computer Clock On Time?
>
> Hi:
>
> I have a friend who's setting up an observatory and the PC that controls
> the telescope needs to have an accurate clock.
> The telescope can point to within arc seconds of a star and that implies
> that the computer clock needs to be within 50 ms.
> If he does a Windows NTP sync first thing in the evening after a few
> hours there's too big an error.  Running the observatory is like flying
> a 747, i.e. there's a lot to do and the time needs to be handled
> automatically, not by manual NTP updates.
>
> I know that TAC32 and a Motorola GPS will do this and you have a lot of
> options of how it does it.  For example at a specified time interval or
> when the computer clock differs from GPS by a specified amount of time
> (50 ms in this case).
>
> What are the options in LH?  I ask because it's a lower cost option then
> the Motorola GPS plus TAC32 option or building a NTP time server.
>
> --
> Have Fun,
>
> Brooke Clarke
> http://www.PRC68.com
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



--
--Eric
_________________________________________
Eric Garner
Brooke Clarke
2010-07-29 19:10:04 UTC
Permalink
Hi Eric:

That seems to be for a time server. He is just using an ordinary PC and
NTP to set it's time.

Have Fun,

Brooke Clarke
http://www.PRC68.com


Eric Garner wrote:
> with the assumption that he seems to be fine right after a sync,
> (seemingly supported in the OP) reconfiguring the windows time service
> to update frequently seems to me to be the easiest answer.
> under xp you can do so thus:
> http://support.microsoft.com/kb/314054
>
> _eric
>
>
> On Thu, Jul 29, 2010 at 11:13 AM, Bob Camp<lists-***@public.gmane.org> wrote:
>
>> Hi
>>
>> My observation is that the displayed clock on LH can be off by> 10 seconds.
>> That's with the current beta code and Windows 7 on a quad processor machine
>> or under XP on a dual core machine.
>>
>> Bob
>>
>> -----Original Message-----
>> From: time-nuts-bounces-***@public.gmane.org [mailto:time-nuts-bounces-***@public.gmane.org] On
>> Behalf Of Brooke Clarke
>> Sent: Thursday, July 29, 2010 2:04 PM
>> To: Discussion of precise time and frequency measurement
>> Subject: [time-nuts] Can Lady Heather Keep Computer Clock On Time?
>>
>> Hi:
>>
>> I have a friend who's setting up an observatory and the PC that controls
>> the telescope needs to have an accurate clock.
>> The telescope can point to within arc seconds of a star and that implies
>> that the computer clock needs to be within 50 ms.
>> If he does a Windows NTP sync first thing in the evening after a few
>> hours there's too big an error. Running the observatory is like flying
>> a 747, i.e. there's a lot to do and the time needs to be handled
>> automatically, not by manual NTP updates.
>>
>> I know that TAC32 and a Motorola GPS will do this and you have a lot of
>> options of how it does it. For example at a specified time interval or
>> when the computer clock differs from GPS by a specified amount of time
>> (50 ms in this case).
>>
>> What are the options in LH? I ask because it's a lower cost option then
>> the Motorola GPS plus TAC32 option or building a NTP time server.
>>
>> --
>> Have Fun,
>>
>> Brooke Clarke
>> http://www.PRC68.com
>>
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>>
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>>
>
>
>
Eric Garner
2010-07-29 19:26:22 UTC
Permalink
see the section of the article "Configuring the Windows Time service
to use an external time source" that details how to reconfigure the
winXP time sync service so that it polls and external server for time
at a specified interval.

On Thu, Jul 29, 2010 at 12:10 PM, Brooke Clarke <brooke-***@public.gmane.org> wrote:
> Hi Eric:
>
> That seems to be for a time server.  He is just using an ordinary PC and NTP
> to set it's time.
>
> Have Fun,
>
> Brooke Clarke
> http://www.PRC68.com
>
>
> Eric Garner wrote:
>>
>> with the assumption that he seems to be fine right after a sync,
>> (seemingly supported in the OP) reconfiguring the windows time service
>> to update frequently seems  to me to be the easiest answer.
>> under xp you can do so thus:
>> http://support.microsoft.com/kb/314054
>>
>> _eric
>>
>>
>> On Thu, Jul 29, 2010 at 11:13 AM, Bob Camp<lists-***@public.gmane.org>  wrote:
>>
>>>
>>> Hi
>>>
>>> My observation is that the displayed clock on LH can be off by>  10
>>> seconds.
>>> That's with the current beta code and Windows 7 on a quad processor
>>> machine
>>> or under XP on a dual core machine.
>>>
>>> Bob
>>>
>>> -----Original Message-----
>>> From: time-nuts-bounces-***@public.gmane.org [mailto:time-nuts-bounces-***@public.gmane.org] On
>>> Behalf Of Brooke Clarke
>>> Sent: Thursday, July 29, 2010 2:04 PM
>>> To: Discussion of precise time and frequency measurement
>>> Subject: [time-nuts] Can Lady Heather Keep Computer Clock On Time?
>>>
>>> Hi:
>>>
>>> I have a friend who's setting up an observatory and the PC that controls
>>> the telescope needs to have an accurate clock.
>>> The telescope can point to within arc seconds of a star and that implies
>>> that the computer clock needs to be within 50 ms.
>>> If he does a Windows NTP sync first thing in the evening after a few
>>> hours there's too big an error.  Running the observatory is like flying
>>> a 747, i.e. there's a lot to do and the time needs to be handled
>>> automatically, not by manual NTP updates.
>>>
>>> I know that TAC32 and a Motorola GPS will do this and you have a lot of
>>> options of how it does it.  For example at a specified time interval or
>>> when the computer clock differs from GPS by a specified amount of time
>>> (50 ms in this case).
>>>
>>> What are the options in LH?  I ask because it's a lower cost option then
>>> the Motorola GPS plus TAC32 option or building a NTP time server.
>>>
>>> --
>>> Have Fun,
>>>
>>> Brooke Clarke
>>> http://www.PRC68.com
>>>
>>>
>>>
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>>> To unsubscribe, go to
>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow the instructions there.
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>>> To unsubscribe, go to
>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow the instructions there.
>>>
>>>
>>
>>
>>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



--
--Eric
_________________________________________
Eric Garner
b***@public.gmane.org
2010-07-29 18:38:28 UTC
Permalink
Or even better download a real NTP for windows.

http://www.meinberg.de/english/sw/ntp.htm#ntp_nt_stable

--

Björn

> with the assumption that he seems to be fine right after a sync,
> (seemingly supported in the OP) reconfiguring the windows time service
> to update frequently seems to me to be the easiest answer.
> under xp you can do so thus:
> http://support.microsoft.com/kb/314054
>
> _eric
>
>
> On Thu, Jul 29, 2010 at 11:13 AM, Bob Camp <lists-***@public.gmane.org> wrote:
>> Hi
>>
>> My observation is that the displayed clock on LH can be off by > 10
>> seconds.
>> That's with the current beta code and Windows 7 on a quad processor
>> machine
>> or under XP on a dual core machine.
>>
>> Bob
>>
>> -----Original Message-----
>> From: time-nuts-bounces-***@public.gmane.org [mailto:time-nuts-bounces-***@public.gmane.org] On
>> Behalf Of Brooke Clarke
>> Sent: Thursday, July 29, 2010 2:04 PM
>> To: Discussion of precise time and frequency measurement
>> Subject: [time-nuts] Can Lady Heather Keep Computer Clock On Time?
>>
>> Hi:
>>
>> I have a friend who's setting up an observatory and the PC that controls
>> the telescope needs to have an accurate clock.
>> The telescope can point to within arc seconds of a star and that implies
>> that the computer clock needs to be within 50 ms.
>> If he does a Windows NTP sync first thing in the evening after a few
>> hours there's too big an error.  Running the observatory is like flying
>> a 747, i.e. there's a lot to do and the time needs to be handled
>> automatically, not by manual NTP updates.
>>
>> I know that TAC32 and a Motorola GPS will do this and you have a lot of
>> options of how it does it.  For example at a specified time interval or
>> when the computer clock differs from GPS by a specified amount of time
>> (50 ms in this case).
>>
>> What are the options in LH?  I ask because it's a lower cost option then
>> the Motorola GPS plus TAC32 option or building a NTP time server.
>>
>> --
>> Have Fun,
>>
>> Brooke Clarke
>> http://www.PRC68.com
>>
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>>
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>
>
>
> --
> --Eric
> _________________________________________
> Eric Garner
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
Mark Sims
2010-07-29 22:57:04 UTC
Permalink
Yes,  there is a time sync command that causes the program to set the system clock at periodic intervals or whenever it differs from GPS by a given amount.
I'm still out on the "Project From Hell Mark II" and I don't remember all the gory details...  I think /TSA on the command line  says sync the time whenever the clocks differ by a millisecond.  There is also /TSO /TSD /TSH /TSM /TSS to set it once, hourly, daily, every minute or second.  There is also a /TSX command to specify the time offset from the Tbolt serial command to GPS (which is usually around 45 milliseconds).  TS from the keyboard just syncs the clock once.
Leigh L. Klotz, Jr WA5ZNU
2010-08-01 03:49:05 UTC
Permalink
On 07/29/2010 03:57 PM, Mark Sims wrote:
> Yes, there is a time sync command that causes the program to set the system clock at periodic intervals or whenever it differs from GPS by a given amount.
> I'm still out on the "Project From Hell Mark II" and I don't remember all the gory details... I think /TSA on the command line says sync the time whenever the clocks differ by a millisecond. There is also /TSO /TSD /TSH /TSM /TSS to set it once, hourly, daily, every minute or second. There is also a /TSX command to specify the time offset from the Tbolt serial command to GPS (which is usually around 45 milliseconds). TS from the keyboard just syncs the clock once.
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

I realize this request is for Windows, but I thought I'd mention this
for the benefit of others.

I'm using tboltd by Ralph Smith on Linux. It provides the time to ntpd
and LH 3.00 beta can talk to tboltd via the TCP connection, so
monitoring can happen at the same time as the ntpd. It works great on
Linux, as well as BSD (for which it was designed):
http://www.mail-archive.com/time-nuts-***@public.gmane.org/msg26129.html

It runs well under wine like this:
wine heather.exe /IP=localhost:45000 /TW=250

The /etc/ntp.conf config file stanza looks like this. (I admit I've
guessed at the 0.0275 delay.)

# for attached GPS
tos mindist 0.030
server 127.127.28.0 minpoll 4 maxpoll 4 prefer
fudge 127.127.28.0 time1 0.0275 stratum 1 refid GPS

Serious use would probably call for a Soekris box, but not so serious
use lets me use the tbolt time signal for free, as I have it on for the
10MHz reference.

Leigh/WA5ZNU
Hal Murray
2010-07-29 23:29:28 UTC
Permalink
> The telescope can point to within arc seconds of a star and that implies
> that the computer clock needs to be within 50 ms.

> If he does a Windows NTP sync first thing in the evening after a few hours
> there's too big an error.

That looks like the classic time vs frequency problem.

What is the primary goal? Pointing or tracking? Pointing requires time.
Tracking requires frequency.

If you are otherwise happy with "Windows NTP sync", you may be able to solve
your problem by doing a sync before pointing at another object if it's been
more than N hours since the last sync.


Long song and dance.... [Remember, I don't run Windows so I may screwup
anything that's Windows specific.]


The typical PC has 2 crystals. One runs at 32 KHz. The other is usually
14.xxx (from early PC days) that gets PLLed up to make clocks for the CPU and
PCI and USB and ...

The 32KHz crystal runs the battery backed RTC/TOY/CMOS clock. It's a watch
crystal so it should be pretty good. But it's not very convenient for
keeping time at the microsecond level.

The 14 MHz crystal is stable, but typically not very accurate. (Remember low
cost.) That's accurate at the PPM level, it will be fine if you just put a
scope on it. It may be off by 50 PPM. Even if the hardware is good, the
software can screw things up. (Linux is good at this. Current kernels don't
get a consistent answer on the same hardware. Jumps by 200 PPM from boot to
boot are not uncommon.)

[Network and audio and ??? cards typically have a separate crystal. They are
usually not convenient for timekeeping but if you do serious audio work you
can measure it's actual frequency.]

Let's see if I can do the math right...

3 hours is 10,000 seconds. 50 PPM times 10,000 seconds is 500,000
microseconds. So if the clock is off by 50 PPM, it will drift 1/2 second in
3 hours. Even 5 PPM will drift 50 ms in 3 hours.


The main reason for running real ntpd rather than just setting the time
occasionally is that ntpd will figure out how far off the frequency is and
correct for it. ntpd calls that fudge factor "drift" and prints it out in
PPM with 3 digits to the right of the decimal point.

If all you need is 50 ms, you should be able to get that most of the time by
just running ntpd over the net. It's sure worth a try. It may not be good
enough if you have a crappy net connection or change from no-load to
uploading tons of data from observations earlier in the evening. (Contact me
off list if you want help in monitoring a ntp server and/or setting up and
interpreting its log files.)



Odds and ends to keep in mind...

Modern PCs use spread spectrum clocking. That fudges things by 1 or 1/2 % or
so which is huge in terms of PPM. The point is that you have to measure it.
Just doing the math from the nominal CPU frequency isn't good enough.

The actual frequency is temperature dependent, so things will change if you
open the roof and let the cold air in or the CPU changes from idle (or off)
to working hard. The ballpark is 1 PPM per 10 F.

One of the classic ways to screwup timekeeping is to miss interrupts,
typically because some other interrupt routine is running too long. This was
easy to tickle on (very) old Linux systems that used PIO rather than DMA for
disk transfers. I only mention it because you might have some strange
hardware with buggy interrupt routines.

Normal Windows clocks tick every 10 ms. Windows has a multimedia mode that
does much better. There is a switch in the Registry or something. It may
help to use that mode. I think you want to leave it on. (The Meinberg
ntpd-installer package turns it on.)

ntpd is both a client and server. A system will act as a client to get time
from lower stratum servers and act as a server to provide time to higher
stratum servers.



--
These are my opinions, not necessarily my employer's. I hate spam.
Bob Camp
2010-07-29 23:51:23 UTC
Permalink
Hi

There is a standard NTP driver that "talks Thunderbolt". The full blown NTP package is pretty easy to set up. It will hold ms accuracy slaved to a GPS.

Bob



On Jul 29, 2010, at 7:29 PM, Hal Murray <hmurray-8cQiHa/C+6Go9G/***@public.gmane.org> wrote:

>
>> The telescope can point to within arc seconds of a star and that implies
>> that the computer clock needs to be within 50 ms.
>
>> If he does a Windows NTP sync first thing in the evening after a few hours
>> there's too big an error.
>
> That looks like the classic time vs frequency problem.
>
> What is the primary goal? Pointing or tracking? Pointing requires time.
> Tracking requires frequency.
>
> If you are otherwise happy with "Windows NTP sync", you may be able to solve
> your problem by doing a sync before pointing at another object if it's been
> more than N hours since the last sync.
>
>
> Long song and dance.... [Remember, I don't run Windows so I may screwup
> anything that's Windows specific.]
>
>
> The typical PC has 2 crystals. One runs at 32 KHz. The other is usually
> 14.xxx (from early PC days) that gets PLLed up to make clocks for the CPU and
> PCI and USB and ...
>
> The 32KHz crystal runs the battery backed RTC/TOY/CMOS clock. It's a watch
> crystal so it should be pretty good. But it's not very convenient for
> keeping time at the microsecond level.
>
> The 14 MHz crystal is stable, but typically not very accurate. (Remember low
> cost.) That's accurate at the PPM level, it will be fine if you just put a
> scope on it. It may be off by 50 PPM. Even if the hardware is good, the
> software can screw things up. (Linux is good at this. Current kernels don't
> get a consistent answer on the same hardware. Jumps by 200 PPM from boot to
> boot are not uncommon.)
>
> [Network and audio and ??? cards typically have a separate crystal. They are
> usually not convenient for timekeeping but if you do serious audio work you
> can measure it's actual frequency.]
>
> Let's see if I can do the math right...
>
> 3 hours is 10,000 seconds. 50 PPM times 10,000 seconds is 500,000
> microseconds. So if the clock is off by 50 PPM, it will drift 1/2 second in
> 3 hours. Even 5 PPM will drift 50 ms in 3 hours.
>
>
> The main reason for running real ntpd rather than just setting the time
> occasionally is that ntpd will figure out how far off the frequency is and
> correct for it. ntpd calls that fudge factor "drift" and prints it out in
> PPM with 3 digits to the right of the decimal point.
>
> If all you need is 50 ms, you should be able to get that most of the time by
> just running ntpd over the net. It's sure worth a try. It may not be good
> enough if you have a crappy net connection or change from no-load to
> uploading tons of data from observations earlier in the evening. (Contact me
> off list if you want help in monitoring a ntp server and/or setting up and
> interpreting its log files.)
>
>
>
> Odds and ends to keep in mind...
>
> Modern PCs use spread spectrum clocking. That fudges things by 1 or 1/2 % or
> so which is huge in terms of PPM. The point is that you have to measure it.
> Just doing the math from the nominal CPU frequency isn't good enough.
>
> The actual frequency is temperature dependent, so things will change if you
> open the roof and let the cold air in or the CPU changes from idle (or off)
> to working hard. The ballpark is 1 PPM per 10 F.
>
> One of the classic ways to screwup timekeeping is to miss interrupts,
> typically because some other interrupt routine is running too long. This was
> easy to tickle on (very) old Linux systems that used PIO rather than DMA for
> disk transfers. I only mention it because you might have some strange
> hardware with buggy interrupt routines.
>
> Normal Windows clocks tick every 10 ms. Windows has a multimedia mode that
> does much better. There is a switch in the Registry or something. It may
> help to use that mode. I think you want to leave it on. (The Meinberg
> ntpd-installer package turns it on.)
>
> ntpd is both a client and server. A system will act as a client to get time
> from lower stratum servers and act as a server to provide time to higher
> stratum servers.
>
>
>
> --
> These are my opinions, not necessarily my employer's. I hate spam.
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
Brooke Clarke
2010-07-29 23:59:08 UTC
Permalink
Hi Hal:

The key thing is pointing, not tracking (where a guide star is commonly
used).
A TPoint model is made by pointing to known stars and the quality of the
model depends on good computer time.

Have Fun,

Brooke Clarke
http://www.PRC68.com


Hal Murray wrote:
>
>> The telescope can point to within arc seconds of a star and that implies
>> that the computer clock needs to be within 50 ms.
>>
>
>> If he does a Windows NTP sync first thing in the evening after a few hours
>> there's too big an error.
>>
> That looks like the classic time vs frequency problem.
>
> What is the primary goal? Pointing or tracking? Pointing requires time.
> Tracking requires frequency.
>
> If you are otherwise happy with "Windows NTP sync", you may be able to solve
> your problem by doing a sync before pointing at another object if it's been
> more than N hours since the last sync.
>
>
> Long song and dance.... [Remember, I don't run Windows so I may screwup
> anything that's Windows specific.]
>
>
> The typical PC has 2 crystals. One runs at 32 KHz. The other is usually
> 14.xxx (from early PC days) that gets PLLed up to make clocks for the CPU and
> PCI and USB and ...
>
> The 32KHz crystal runs the battery backed RTC/TOY/CMOS clock. It's a watch
> crystal so it should be pretty good. But it's not very convenient for
> keeping time at the microsecond level.
>
> The 14 MHz crystal is stable, but typically not very accurate. (Remember low
> cost.) That's accurate at the PPM level, it will be fine if you just put a
> scope on it. It may be off by 50 PPM. Even if the hardware is good, the
> software can screw things up. (Linux is good at this. Current kernels don't
> get a consistent answer on the same hardware. Jumps by 200 PPM from boot to
> boot are not uncommon.)
>
> [Network and audio and ??? cards typically have a separate crystal. They are
> usually not convenient for timekeeping but if you do serious audio work you
> can measure it's actual frequency.]
>
> Let's see if I can do the math right...
>
> 3 hours is 10,000 seconds. 50 PPM times 10,000 seconds is 500,000
> microseconds. So if the clock is off by 50 PPM, it will drift 1/2 second in
> 3 hours. Even 5 PPM will drift 50 ms in 3 hours.
>
>
> The main reason for running real ntpd rather than just setting the time
> occasionally is that ntpd will figure out how far off the frequency is and
> correct for it. ntpd calls that fudge factor "drift" and prints it out in
> PPM with 3 digits to the right of the decimal point.
>
> If all you need is 50 ms, you should be able to get that most of the time by
> just running ntpd over the net. It's sure worth a try. It may not be good
> enough if you have a crappy net connection or change from no-load to
> uploading tons of data from observations earlier in the evening. (Contact me
> off list if you want help in monitoring a ntp server and/or setting up and
> interpreting its log files.)
>
>
>
> Odds and ends to keep in mind...
>
> Modern PCs use spread spectrum clocking. That fudges things by 1 or 1/2 % or
> so which is huge in terms of PPM. The point is that you have to measure it.
> Just doing the math from the nominal CPU frequency isn't good enough.
>
> The actual frequency is temperature dependent, so things will change if you
> open the roof and let the cold air in or the CPU changes from idle (or off)
> to working hard. The ballpark is 1 PPM per 10 F.
>
> One of the classic ways to screwup timekeeping is to miss interrupts,
> typically because some other interrupt routine is running too long. This was
> easy to tickle on (very) old Linux systems that used PIO rather than DMA for
> disk transfers. I only mention it because you might have some strange
> hardware with buggy interrupt routines.
>
> Normal Windows clocks tick every 10 ms. Windows has a multimedia mode that
> does much better. There is a switch in the Registry or something. It may
> help to use that mode. I think you want to leave it on. (The Meinberg
> ntpd-installer package turns it on.)
>
> ntpd is both a client and server. A system will act as a client to get time
> from lower stratum servers and act as a server to provide time to higher
> stratum servers.
>
>
>
>
Hal Murray
2010-11-28 07:17:13 UTC
Permalink
brooke-***@public.gmane.org said:
> The telescope has around arc second pointing capability so I think he needs
> a hundredth of a second or slightly better.

That doesn't sound right. What's the field of view of the telescope?


> I've found some NMEA software that will set a PC clock, but it expects the
> data on a COM port. Don't know how it might work with USB data.

I don't use Windows, but most likely it will magically show up as COMx when
you plug it in.


>> It is not clear whether the message is fixed of if it will reply
>> to queries.

Normally, NMEA devices send a clump of "sentences" each second. Usually, you
can turn off the ones you don't want.



--
These are my opinions, not necessarily my employer's. I hate spam.
Bruce Griffiths
2010-11-28 07:50:42 UTC
Permalink
Hal Murray wrote:
> brooke-***@public.gmane.org said:
>
>> The telescope has around arc second pointing capability so I think he needs
>> a hundredth of a second or slightly better.
>>
> That doesn't sound right. What's the field of view of the telescope?
>
>
Worst case (declination = 0), 1 arc second of pointing is equivalent to
about 1/15 (~67ms) second of time.
The telescope field of view isn't really relevant (as long as its
greater than 1 arcsec) if one needs to maintain 1 arc second pointing.
In this case a timing accuracy of 10ms isn't overly conservative.
If one merely needs to ensure that the telescope can acquire the
object of interest so that it is well within the field of view of an
eyepiece, then 1 arc second pointing isnt necessary.
However if one needs to place the object of interest directly on a
spectrometer slit or fibre then a somewhat lower pointing error is
desirable.

To achieve 1 arc second pointing accuracy usually requires correcting
for such effects as
Tube flexure
The orthogonality error between axes
drive train eccentricities
The value of dUT1.
etc

One also requires high resolution encoders (or equivalent on each axis).
Alternative a stellar compass system like those used to gage the
orientation of spacecraft can be employed.
Despite what Meade would have one believe the application of such
devices to telescope pointing did not originate with them.


>
>> I've found some NMEA software that will set a PC clock, but it expects the
>> data on a COM port. Don't know how it might work with USB data.
>>
> I don't use Windows, but most likely it will magically show up as COMx when
> you plug it in.
>
>

>
>>> It is not clear whether the message is fixed of if it will reply
>>> to queries.
>>>
> Normally, NMEA devices send a clump of "sentences" each second. Usually, you
> can turn off the ones you don't want.
>
>
>

Bruce
Brooke Clarke
2010-11-29 00:50:23 UTC
Permalink
Hi Hal:

FOV is 0.2 min x 0.3 min @0.47arcsec/pixel when using 1x1 binning. The
seeing doesn't often justify 1x1 binning.

Mount is the Paramount ME.
http://www.bisque.com/help/paramountme/performance_specifications.htm
There's a software package called TPoint where you manually point to
known stars and that data is used to fit a model of the common mount
errors and so correct them. For example if the RA and DEC axis are not
at exactly 90 degrees. After doing many dozens of stars this correction
(and the PEC correction) allow making images without a guide star for
say 10 minutes, something that's impossible on a Meade or Celestron type
setup.

The scope is a Planewave 17":
http://www.planewave.com/index.php?page=1&id0=0&id=1

Camera is SBIG STL-11000M
http://www.sbig.com/sbwhtmls/online.htm

Have Fun,

Brooke Clarke
http://www.PRC68.com


Hal Murray wrote:
> brooke-***@public.gmane.org said:
>
>> The telescope has around arc second pointing capability so I think he needs
>> a hundredth of a second or slightly better.
>>
> That doesn't sound right. What's the field of view of the telescope?
>
>
>
>> I've found some NMEA software that will set a PC clock, but it expects the
>> data on a COM port. Don't know how it might work with USB data.
>>
> I don't use Windows, but most likely it will magically show up as COMx when
> you plug it in.
>
>
>
>>> It is not clear whether the message is fixed of if it will reply
>>> to queries.
>>>
> Normally, NMEA devices send a clump of "sentences" each second. Usually, you
> can turn off the ones you don't want.
>
>
>
>
Ralph Smith
2010-11-29 02:37:33 UTC
Permalink
OK, after dropping the amount of coin on the mount, scope, and camera, (my guess in the neighborhood of $40,000 US) a few hundred bucks for a timing solution shouldn't be too bad. Would an external NTP box, perhaps a Soekris Net4501 driven by a Thunderbolt or another GPS receiver with PPS work? My Windows boxes typically track my Soekris NTP server within a few ms, my Macs to about 1 ms, and my FreeBSD boxes to about 30 microseconds.

Ralph

On Nov 28, 2010, at 7:50 PM, Brooke Clarke wrote:

> Hi Hal:
>
> FOV is 0.2 min x 0.3 min @0.47arcsec/pixel when using 1x1 binning. The seeing doesn't often justify 1x1 binning.
>
> Mount is the Paramount ME.
> http://www.bisque.com/help/paramountme/performance_specifications.htm
> There's a software package called TPoint where you manually point to known stars and that data is used to fit a model of the common mount errors and so correct them. For example if the RA and DEC axis are not at exactly 90 degrees. After doing many dozens of stars this correction (and the PEC correction) allow making images without a guide star for say 10 minutes, something that's impossible on a Meade or Celestron type setup.
>
> The scope is a Planewave 17":
> http://www.planewave.com/index.php?page=1&id0=0&id=1
>
> Camera is SBIG STL-11000M
> http://www.sbig.com/sbwhtmls/online.htm
>
> Have Fun,
>
> Brooke Clarke
> http://www.PRC68.com
>
>
> Hal Murray wrote:
>> brooke-***@public.gmane.org said:
>>
>>> The telescope has around arc second pointing capability so I think he needs
>>> a hundredth of a second or slightly better.
>>>
>> That doesn't sound right. What's the field of view of the telescope?
>>
>>
>>
>>> I've found some NMEA software that will set a PC clock, but it expects the
>>> data on a COM port. Don't know how it might work with USB data.
>>>
>> I don't use Windows, but most likely it will magically show up as COMx when
>> you plug it in.
>>
>>
>>
>>>> It is not clear whether the message is fixed of if it will reply
>>>> to queries.
>>>>
>> Normally, NMEA devices send a clump of "sentences" each second. Usually, you
>> can turn off the ones you don't want.
>>
>>
>>
>>
> <23Apr2010Pauls.jpg>_______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
Chris Albertson
2010-11-29 03:34:46 UTC
Permalink
I'm new on this list but This is something I know a little about.

Pointing a telescope acuratly is more then just knowing the time. The
air refracts, or bends the line of sight depending on the angle from
zenith and no matter how strong the structure of the telescope sags
under it's own weight. I'd not bring this up except that I know
there are people here concerned about very small errors.

The solution is a closed loop system. You measure the location of a
star on the image plain periodically and compute it's motion and then
send small correction signals to the motors driving the scope to drive
the star's apparent velocity to zero. In general if things are set up
well you can measure a star's location to about 1/10th of a pixel
because the star is blurred and you can fit a function. There is more
to it but you get the point, you only need to know the time well
enough to center an object in the field of view. Typically 15 arc
seconds of very good. For that you need to know the time to just
better than one second. I can get that without owning a GPS receiver
using NTP servers on the Internet.

For other uses you might need much better time. For example high
speed photometry (occupations) knowing current time to 10 ms would be
good Some people will record audio from WWV simultaneous with the
optical data.

If you already have NTP as 1 ms youhave what you need already

On Sun, Nov 28, 2010 at 6:37 PM, Ralph Smith <ralph-***@public.gmane.org> wrote:
> OK, after dropping the amount of coin on the mount, scope, and camera, (my guess in the neighborhood of $40,000 US) a few hundred bucks for a timing solution shouldn't be too bad. Would an external NTP box, perhaps a Soekris Net4501 driven by a Thunderbolt or another GPS receiver with PPS work? My Windows boxes typically track my Soekris NTP server within a few ms, my Macs to about 1 ms, and my FreeBSD boxes to about 30 microseconds.
>
> Ralph
>
> On Nov 28, 2010, at 7:50 PM, Brooke Clarke wrote:
>
>> Hi Hal:
>>
>> FOV is 0.2 min x 0.3 min @0.47arcsec/pixel when using 1x1 binning.  The seeing doesn't often justify 1x1 binning.
>>
>> Mount is the Paramount ME.
>> http://www.bisque.com/help/paramountme/performance_specifications.htm
>> There's a software package called TPoint where you manually point to known stars and that data is used to fit a model of the common mount errors and so correct them.  For example if the RA and DEC axis are not at exactly 90 degrees.  After doing many dozens of stars this correction (and the PEC correction) allow making images without a guide star for say 10 minutes, something that's impossible on a Meade or Celestron type setup.
>>
>> The scope is a Planewave 17":
>> http://www.planewave.com/index.php?page=1&id0=0&id=1
>>
>> Camera is SBIG  STL-11000M
>> http://www.sbig.com/sbwhtmls/online.htm
>>
>> Have Fun,
>>
>> Brooke Clarke
>> http://www.PRC68.com
>>
>>
>> Hal Murray wrote:
>>> brooke-***@public.gmane.org  said:
>>>
>>>> The telescope has around arc second pointing capability so I think he  needs
>>>> a hundredth of a second or slightly better.
>>>>
>>> That doesn't sound right.  What's the field of view of the telescope?
>>>
>>>
>>>
>>>> I've found some NMEA software that will set a PC clock, but it expects  the
>>>> data on a COM port.  Don't know how it might work with USB data.
>>>>
>>> I don't use Windows, but most likely it will magically show up as COMx when
>>> you plug it in.
>>>
>>>
>>>
>>>>> It is not clear whether the message is fixed of if it will reply
>>>>> to queries.
>>>>>
>>> Normally, NMEA devices send a clump of "sentences" each second.  Usually, you
>>> can turn off the ones you don't want.
>>>
>>>
>>>
>>>
>> <23Apr2010Pauls.jpg>_______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



--
=====
Chris Albertson
Redondo Beach, California
Hal Murray
2011-01-16 21:26:58 UTC
Permalink
> Also how about using Aluminum Organic Polymer caps (where there are
> replacements) instead of the plain electrolytic caps?

One problem with using low ESR caps is that some regulators depend upon the
ESR for stability.



> The plan is to replace all the electrolytic and tantalum caps since the
> ones that are now OK may fail in the near future.

Why replace the tantalums? Do they dry out like electrolytics?

I poked around a bit and found a couple of web pages that said no-problems,
but both specifically referred to solid tantalums. Are there versions using
a liquid?



--
These are my opinions, not necessarily my employer's. I hate spam.
Dr. Frank Stellmach
2011-01-16 21:58:49 UTC
Permalink
Hi Brooke,

Our electronics designers (@automotive electronics supplier) have to
take lot of care concerning voltage stabilizers, that the cap provides
sufficiently low ESR, which electrolytics deliver only at much higher
capacitance value; and a series resistor is introduced to limit loading
currents of the Tantalum. Otherwise, the Tantalums might fail, either
open, or they might burn, which already caused total damage of vehicles
(Mercedes).

I assume, latter aspect has been overseen in your device. This is the
main fault modus of Tantalums.

Electrolytics dry out over the years, especially when stored
non-operating, Tantalums do not at all.

You have to chose higher cap value and higher environmental temperature
grade (>= 105°C) for the ElCo, to get a reliable replacement.
I commonly do not recommend polymer elcos, they have some further
disadvantages, one being the price, the other (perhaps) being CN
compounds in the polymer, which may be poisonous in case of fire.

High cap value Ceramics are available since years, therefore , if price
is no argument, I'd prefer to take them for values up to ~ 100µF; X5R
ceramics should be good enough. With some luck, they (SMD type) fit to
the existing pads on the solder side.

Otherwise, you have to analyse the (misdesigned) circuitry for possible
high pulse currents, and replace the bad tantalums by new ones in series
with a limiting resistor. Several tantalums in parallel might also
improve withstanding to current pulses. Check the tantalums datasheets.
epcos has got some detailed application hints, as far as I remember.

regards Frank
Poul-Henning Kamp
2011-01-16 22:15:09 UTC
Permalink
In message <4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:

>High cap value Ceramics are available since years, [...]

Can you clarify one thing for me: When I studied datasheets for these
it looked like they drop 50% of their capacitance at a DC voltage
of 10-20V.

Doesn't that make them a so-so bargain for power supply bulk capacitance ?

--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk-***@public.gmane.org | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Chuck Harris
2011-01-16 22:43:18 UTC
Permalink
I have found that such high dielectric constant capacitors have other
problems in some circuits. While the capacitance value is dropping,
as much as 50% when you apply voltage, their physical volume is
changing. They behave as piezoelectric transducers.

I have used them inappropriately, and found them singing loudly. If
that were to happen in a switching regulator, I would think the ultrasonic
vibrations could do damage to the capacitor, the PCB, or the solder joints.

-Chuck Harris

Poul-Henning Kamp wrote:
> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>
>> High cap value Ceramics are available since years, [...]
>
> Can you clarify one thing for me: When I studied datasheets for these
> it looked like they drop 50% of their capacitance at a DC voltage
> of 10-20V.
>
> Doesn't that make them a so-so bargain for power supply bulk capacitance ?
>
s***@public.gmane.org
2011-01-20 11:35:08 UTC
Permalink
Also they are almost impossible to solder by hand without setting high internal stress which sometimes results in cracks, not necessarily immediately. Even on professional wave solder equipment, high delayed failure rates are not uncommon, which sometimes results in many units having to be recalled. Don't ask me how I know, I was lucky this was not my design, but it could have been.

Yet, I still use them, under the assurance that our manufacturing folks have resolved the problem, which for now seems true.

Didier

Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: Chuck Harris <cfharris-***@public.gmane.org>
Sender: time-nuts-bounces-***@public.gmane.org
Date: Sun, 16 Jan 2011 17:43:18
To: Discussion of precise time and frequency measurement<time-nuts-***@public.gmane.org>
Reply-To: Discussion of precise time and frequency measurement
<time-nuts-***@public.gmane.org>
Subject: Re: [time-nuts] GC-1000 Clock Cap Choice?

I have found that such high dielectric constant capacitors have other
problems in some circuits. While the capacitance value is dropping,
as much as 50% when you apply voltage, their physical volume is
changing. They behave as piezoelectric transducers.

I have used them inappropriately, and found them singing loudly. If
that were to happen in a switching regulator, I would think the ultrasonic
vibrations could do damage to the capacitor, the PCB, or the solder joints.

-Chuck Harris

Poul-Henning Kamp wrote:
> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>
>> High cap value Ceramics are available since years, [...]
>
> Can you clarify one thing for me: When I studied datasheets for these
> it looked like they drop 50% of their capacitance at a DC voltage
> of 10-20V.
>
> Doesn't that make them a so-so bargain for power supply bulk capacitance ?
>

_______________________________________________
time-nuts mailing list -- time-nuts-***@public.gmane.org
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brooke Clarke
2011-01-20 15:29:29 UTC
Permalink
Hi Didier:

I know CMAX had a real problem with their LF time receiver boards. The
high value caps used to resonate the loop-stick antenna permanently
changed value so much when heated to soldering temperature they no
longer worked. Now they use radial lead caps shrink wrapped to the
loop-sticks instead of having the cap on the receiver PCB.

Have Fun,

Brooke Clarke
http://www.PRC68.com


shalimr9-***@public.gmane.org wrote:
> Also they are almost impossible to solder by hand without setting high internal stress which sometimes results in cracks, not necessarily immediately. Even on professional wave solder equipment, high delayed failure rates are not uncommon, which sometimes results in many units having to be recalled. Don't ask me how I know, I was lucky this was not my design, but it could have been.
>
> Yet, I still use them, under the assurance that our manufacturing folks have resolved the problem, which for now seems true.
>
> Didier
>
> Sent from my Verizon Wireless BlackBerry
>
> -----Original Message-----
> From: Chuck Harris<cfharris-***@public.gmane.org>
> Sender: time-nuts-bounces-***@public.gmane.org
> Date: Sun, 16 Jan 2011 17:43:18
> To: Discussion of precise time and frequency measurement<time-nuts-***@public.gmane.org>
> Reply-To: Discussion of precise time and frequency measurement
> <time-nuts-***@public.gmane.org>
> Subject: Re: [time-nuts] GC-1000 Clock Cap Choice?
>
> I have found that such high dielectric constant capacitors have other
> problems in some circuits. While the capacitance value is dropping,
> as much as 50% when you apply voltage, their physical volume is
> changing. They behave as piezoelectric transducers.
>
> I have used them inappropriately, and found them singing loudly. If
> that were to happen in a switching regulator, I would think the ultrasonic
> vibrations could do damage to the capacitor, the PCB, or the solder joints.
>
> -Chuck Harris
>
> Poul-Henning Kamp wrote:
>
>> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>>
>>
>>> High cap value Ceramics are available since years, [...]
>>>
>> Can you clarify one thing for me: When I studied datasheets for these
>> it looked like they drop 50% of their capacitance at a DC voltage
>> of 10-20V.
>>
>> Doesn't that make them a so-so bargain for power supply bulk capacitance ?
>>
>>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
>
>
paul swed
2011-01-20 15:38:57 UTC
Permalink
Good tidbit on the cmax

On Thu, Jan 20, 2011 at 10:29 AM, Brooke Clarke <brooke-***@public.gmane.org> wrote:

> Hi Didier:
>
> I know CMAX had a real problem with their LF time receiver boards. The
> high value caps used to resonate the loop-stick antenna permanently changed
> value so much when heated to soldering temperature they no longer worked.
> Now they use radial lead caps shrink wrapped to the loop-sticks instead of
> having the cap on the receiver PCB.
>
> Have Fun,
>
> Brooke Clarke
> http://www.PRC68.com
>
>
> shalimr9-***@public.gmane.org wrote:
>
>> Also they are almost impossible to solder by hand without setting high
>> internal stress which sometimes results in cracks, not necessarily
>> immediately. Even on professional wave solder equipment, high delayed
>> failure rates are not uncommon, which sometimes results in many units having
>> to be recalled. Don't ask me how I know, I was lucky this was not my design,
>> but it could have been.
>>
>> Yet, I still use them, under the assurance that our manufacturing folks
>> have resolved the problem, which for now seems true.
>>
>> Didier
>>
>> Sent from my Verizon Wireless BlackBerry
>>
>> -----Original Message-----
>> From: Chuck Harris<cfharris-***@public.gmane.org>
>> Sender: time-nuts-bounces-***@public.gmane.org
>> Date: Sun, 16 Jan 2011 17:43:18
>> To: Discussion of precise time and frequency measurement<
>> time-nuts-***@public.gmane.org>
>> Reply-To: Discussion of precise time and frequency measurement
>> <time-nuts-***@public.gmane.org>
>> Subject: Re: [time-nuts] GC-1000 Clock Cap Choice?
>>
>> I have found that such high dielectric constant capacitors have other
>> problems in some circuits. While the capacitance value is dropping,
>> as much as 50% when you apply voltage, their physical volume is
>> changing. They behave as piezoelectric transducers.
>>
>> I have used them inappropriately, and found them singing loudly. If
>> that were to happen in a switching regulator, I would think the ultrasonic
>> vibrations could do damage to the capacitor, the PCB, or the solder
>> joints.
>>
>> -Chuck Harris
>>
>> Poul-Henning Kamp wrote:
>>
>>
>>> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>>>
>>>
>>>
>>>> High cap value Ceramics are available since years, [...]
>>>>
>>>>
>>> Can you clarify one thing for me: When I studied datasheets for these
>>> it looked like they drop 50% of their capacitance at a DC voltage
>>> of 10-20V.
>>>
>>> Doesn't that make them a so-so bargain for power supply bulk capacitance
>>> ?
>>>
>>>
>>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>> _______________________________________________
>> time-nuts mailing list -- time-nuts-***@public.gmane.org
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>>
>>
>>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
Bob Camp
2011-01-16 22:59:03 UTC
Permalink
Hi

The other thing to watch out for is the temperature coefficient. Some of the high K materials move a *lot* with modest changes in temperature. There are indeed industry standards on what a given dielectric code should be. In some cases there have ben liberties taken interpreting the codes. You really need to go back to the original data sheet on each part to see what the temp co actually is.

In a modestly warm box (say 60 C) the net effect between voltage and temperature may be that you have < 1/4 your original capacitance.

Bob


On Jan 16, 2011, at 5:15 PM, Poul-Henning Kamp wrote:

> In message <4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>
>> High cap value Ceramics are available since years, [...]
>
> Can you clarify one thing for me: When I studied datasheets for these
> it looked like they drop 50% of their capacitance at a DC voltage
> of 10-20V.
>
> Doesn't that make them a so-so bargain for power supply bulk capacitance ?
>
> --
> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
> phk-***@public.gmane.org | TCP/IP since RFC 956
> FreeBSD committer | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by incompetence.
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
Poul-Henning Kamp
2011-01-16 23:03:32 UTC
Permalink
In message <A9103007-9686-4310-A9E5-AA8B193D1CCF-***@public.gmane.org>, Bob Camp writes:

>In a modestly warm box (say 60 C) the net effect between voltage
>and temperature may be that you have < 1/4 your original capacitance.

That was sort of my conclusion too, not exactly the first thing that
springs to mind when you read the datasheet either, is it ?

Thanks for confirming my suspicion.

--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk-***@public.gmane.org | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
Magnus Danielson
2011-01-19 20:26:20 UTC
Permalink
On 01/16/2011 11:15 PM, Poul-Henning Kamp wrote:
> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>
>> High cap value Ceramics are available since years, [...]
>
> Can you clarify one thing for me: When I studied datasheets for these
> it looked like they drop 50% of their capacitance at a DC voltage
> of 10-20V.
>
> Doesn't that make them a so-so bargain for power supply bulk capacitance ?
>

High dielectric "constant" comes at a high price performance wise. :)

You gain some, you loose some. Temperature performance and linearity.
It should not come as a surprise if you start to think about what
happens on the atomic/molecular level of dielectrics, it's a combination
of electrostatics and "mechanical" stress.

Cheers,
Magnus
Brooke Clarke
2011-01-19 20:31:43 UTC
Permalink
Hi Magnus:

I've just placed an order for 16 different kinds of caps.
Most of them are either low leakage or low series resistance
electrolytic types but also a small number of plastic film caps. Should
be here in a few days.
Also have replacements for the tone decoder PLL ICs and the Op Amps that
drive them, but will replace the caps first and see what happens.

Have Fun,

Brooke Clarke
http://www.PRC68.com


Magnus Danielson wrote:
> On 01/16/2011 11:15 PM, Poul-Henning Kamp wrote:
>> In message<4D336A19.40709-***@public.gmane.org>, "Dr. Frank Stellmach" writes:
>>
>>> High cap value Ceramics are available since years, [...]
>>
>> Can you clarify one thing for me: When I studied datasheets for these
>> it looked like they drop 50% of their capacitance at a DC voltage
>> of 10-20V.
>>
>> Doesn't that make them a so-so bargain for power supply bulk
>> capacitance ?
>>
>
> High dielectric "constant" comes at a high price performance wise. :)
>
> You gain some, you loose some. Temperature performance and linearity.
> It should not come as a surprise if you start to think about what
> happens on the atomic/molecular level of dielectrics, it's a
> combination of electrostatics and "mechanical" stress.
>
> Cheers,
> Magnus
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
>
Magnus Danielson
2011-01-19 20:41:23 UTC
Permalink
Brooke,

On 01/19/2011 09:31 PM, Brooke Clarke wrote:
> Hi Magnus:
>
> I've just placed an order for 16 different kinds of caps.
> Most of them are either low leakage or low series resistance
> electrolytic types but also a small number of plastic film caps. Should
> be here in a few days.

It would be interesting to try these out with a capacitance meter at
various temperatures. See how capacitance and loss changes over time.
The memory effect would also be fun to test if it changes. Should be
enlightening. Another thing would be to see how it varies with applied
voltage.

> Also have replacements for the tone decoder PLL ICs and the Op Amps that
> drive them, but will replace the caps first and see what happens.

Sounds like a fun project. Have fun!

Cheers,
Magnus
Dr. Frank Stellmach
2011-01-16 22:57:27 UTC
Permalink
>Can you clarify one thing for me: When I studied datasheets for these
> it looked like they drop 50% of their capacitance at a DC voltage
> of 10-20V.

> Doesn't that make them a so-so bargain for power supply bulk capacitance ?

Poul-Henning,
Pls. check the spec for dielectric material.

X5R is automotive grade, interior applications (up to 85°C) ; X7R would be engine applications (>= 125°C).

Those ceramics don't have such high detoriation of cap vs. T and V.

For example, Y5V is commercial application range and might have such problems.

Frank
Dr. Frank Stellmach
2011-01-16 23:08:18 UTC
Permalink
>I have found that such high dielectric constant capacitors have other
> problems in some circuits. While the capacitance value is dropping,
> as much as 50% when you apply voltage, their physical volume is
> changing. They behave as piezoelectric transducers.

> I have used them inappropriately, and found them singing loudly. If
> that were to happen in a switching regulator, I would think the ultrasonic
> vibrations could do damage to the capacitor, the PCB, or the solder joints.

> -Chuck Harris

We use ceramics widely in our applications, also switching power supplies and PWM circuits.

Our customers always search intensivley for audible and visual problems before accepting our products, (we design instrument clusters).

(X5R) ceranic capacitors did not yet pose any audible problems, but "singing" inductors did.
It is also known, that ceramics'* piezo effect is problematic for audio amplifiers.

I have to admit, that we do not use very high value caps; several µF is oK, higher values are either too expensive compared to ElCo and Ta, or too bulky.


Anyhow, if our dear customers require us to avoid Tantalums for safety concerns, Ceramics are looked upon first.

Frank
Hal Murray
2011-03-18 19:29:34 UTC
Permalink
> From what I can learn about "seeing" the atmosphere instability is too
> great to allow making measurements optically in the 1 ms area.

There has been some interesting amateur work done on taking lots of
short-exposure pictures and sifting through them to find the good ones. (I
can't find a good URL right now.)

I don't know if any stars are bright enough to make that practical.


> The source has to be "not moving" (which I think leaves out using
> something like jupiter)

We know where Jupiter is. It should be possible to correct for that motion.
It's just another layer of software. :)


> As someone else has pointed out, measuring the earth surface position
> relative to spacecraft orbits, e.g. GPS, would be another technique.

That's an interesting idea, but I think all the orbit data for GPS satellites
is Earth relative rather than star relative. I wonder if the group that
drives the GPS satellites even knows their location relative to the stars.
I'll bet not.



--
These are my opinions, not necessarily my employer's. I hate spam.
Chris Albertson
2011-03-18 22:37:45 UTC
Permalink
On Fri, Mar 18, 2011 at 12:29 PM, Hal Murray <hmurray-8cQiHa/C+6Go9G/***@public.gmane.org> wrote:
>
>> From what I can learn about "seeing" the atmosphere instability is too
>> great to allow making measurements optically in the 1 ms area.
>
> There has been some interesting amateur work done on taking lots of
> short-exposure pictures and sifting through them to find the good ones.  (I
> can't find a good URL right now.)

Video is used for imaging planets. They are bright because like earth
they are lit by the sun. But measurement is different from imaging.
For this job we don't need pictures. The method using by FASST is
called "drift scan". The images of the sky is allowed to move over a
large CCD sensor. The telescope does not track the sky so the stars
images drift. CCD sensors are read out one row at a time then the
entire images is shifted down and the next ros is read. It a drift
scan camers the charge shifting in the CCD is synchronized to the
motion of the image. The result is a very. long image. Think of
it as like a flat bad scanner ony the image moves not the sensor.
This allows for some longer exposures and also covers a lot of the
sky. One BIG advantage of drift scan is that the telescope is bolted
down solid and never moves, so uncertainty in pointing is greatly
reduced.

The FASST camera continously a large part of the sky as it drifts
overheadand they match that data to a catalog of stars and then know
the rate the sky is drifting and thenthe rate the Earth turns

I worked on a drift scan camera project a few years back. We were
able to generate the wave forms to control the CCD in software. The
Earth does not really turn all that fast I had a prtottype camera
mounted on my Garage roof for a long time, It used a chaep telephoto
lens (135mm f/2.8) as the telescopes and we were getting surprizing
accuracy for a "junk box" optical system. An example of one of the
prototypes is the white tripple lens unit here
http://www.tass-survey.org/. Software ran under a rel-time version
of Linux with timming by NTP but later re-calibratd by the stars
themselves. Our system not not nearly as good as FASST, not even
close but used many of the same techniques



--
=====
Chris Albertson
Redondo Beach, California
jimlux
2011-03-19 03:31:26 UTC
Permalink
> That's an interesting idea, but I think all the orbit data for GPS satellites
> is Earth relative rather than star relative. I wonder if the group that
> drives the GPS satellites even knows their location relative to the stars.
> I'll bet not.
>

I'll bet they do. Lots of earth orbiting satellites use star trackers
for position determination.
Chris Albertson
2011-03-19 18:24:29 UTC
Permalink
On Fri, Mar 18, 2011 at 8:31 PM, jimlux <jimlux-***@public.gmane.org> wrote:
>
>> That's an interesting idea, but I think all the orbit data for GPS
>> satellites
>> is Earth relative rather than star relative.  I wonder if the group that
>> drives the GPS satellites even knows their location relative to the stars.
>> I'll bet not.
>>
>
> I'll bet they do. Lots of earth orbiting satellites use star trackers for
> position determination.

Star trackers or Sun sensors are used for pointing, They measure
rotation but not position. It is at least conceptually easy to
measure a GPS sat's location if you have a GPS receiver at a surveyed
location. And also they have access to the encrypted military band we
don't so they can correct for atmospheric effects. I think in
addition the GPS sats get observed optically too
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>



--
=====
Chris Albertson
Redondo Beach, California
Hal Murray
2011-03-28 01:03:03 UTC
Permalink
brooke-***@public.gmane.org said:
> John Mills (THE NTP guru) has written a number of papers on and built
> examples of a matched filter type receiver for the HF station WWV, but the
> ideas would also be applicable to a WWVB receiver. The performance he gets
> from WWV would knock your socks off so I expect a WWVB version would be
> even better. In addition there are some things that could be done to
> improve it.

I assume you mean Dave Mills rather than John.
http://www.ece.udel.edu/~mills/index.html

He's basically given up on WWVB. In his area, there is too much crap around
60 KHz, mostly from switching power supplies.
http://lists.ntp.org/pipermail/questions/2003-August/000106.html
http://lists.ntp.org/pipermail/questions/2006-April/009958.html



--
These are my opinions, not necessarily my employer's. I hate spam.
paul swed
2011-03-28 02:07:35 UTC
Permalink
Indeed those posts are from 2003 the car plant could be closed these days.
But I do indeed receive wwvb I think pretty well with a loop and preamp. Its
always been a challenge on the east coast and even Michigan when I first
started tinkering way to many years ago.
Pretty much before all these switching power supplies and cpfls etc.
Regards

On Sun, Mar 27, 2011 at 9:03 PM, Hal Murray <hmurray-8cQiHa/C+6Go9G/***@public.gmane.org> wrote:

>
> brooke-***@public.gmane.org said:
> > John Mills (THE NTP guru) has written a number of papers on and built
> > examples of a matched filter type receiver for the HF station WWV, but
> the
> > ideas would also be applicable to a WWVB receiver. The performance he
> gets
> > from WWV would knock your socks off so I expect a WWVB version would be
> > even better. In addition there are some things that could be done to
> > improve it.
>
> I assume you mean Dave Mills rather than John.
> http://www.ece.udel.edu/~mills/index.html
>
> He's basically given up on WWVB. In his area, there is too much crap
> around
> 60 KHz, mostly from switching power supplies.
> http://lists.ntp.org/pipermail/questions/2003-August/000106.html
> http://lists.ntp.org/pipermail/questions/2006-April/009958.html
>
>
>
> --
> These are my opinions, not necessarily my employer's. I hate spam.
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
Hal Murray
2012-02-20 20:32:27 UTC
Permalink
> You may be able to do a similar thing in your receivers. For example if the
> master node were to send a timing message at known times (say once at the
> top of every hour) the receivers could use that to determine their local
> clock offset and rate for those cases where the path was the same (or maybe
> even for a list of paths).

Nope. The problem is queuing delays in routers. Satellites don't have
queues.

I think those delays is what he is trying to measure.


--
These are my opinions, not necessarily my employer's. I hate spam.
Bob Bownes
2012-02-20 20:53:47 UTC
Permalink
Think trying to measure the distance between two distant moving spacecraft
with no idea what the gravitational gradient is between them or the ability
to measure the doppler.

Unless, of course, Bill is doing much different things than he was when I
last ran into him. :)



On Mon, Feb 20, 2012 at 3:32 PM, Hal Murray <hmurray-8cQiHa/C+6Go9G/***@public.gmane.org> wrote:

>
> > You may be able to do a similar thing in your receivers. For example if
> the
> > master node were to send a timing message at known times (say once at
> the
> > top of every hour) the receivers could use that to determine their local
> > clock offset and rate for those cases where the path was the same (or
> maybe
> > even for a list of paths).
>
> Nope. The problem is queuing delays in routers. Satellites don't have
> queues.
>
> I think those delays is what he is trying to measure.
>
>
> --
> These are my opinions, not necessarily my employer's. I hate spam.
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts-***@public.gmane.org
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
Hal Murray
2012-02-21 01:24:15 UTC
Permalink
brooke-***@public.gmane.org said:
> It was my understanding that the receiving station knows the path taken, is
> that the case? Or are you saying even when it's the same path the time
> delay has large variations?

Even if the path is stable, the delays vary due to queuing delays in routers.

The simple example is a router with 3 cables. Call them A, B, and C. Assume
they are all the same speed.

If B is idle, packets that come in on A can go out C right away.

But suppose A and B are both sending traffic out C. If packets arrive at
close to the same time, the second one will have to wait on a queue until the
first one is finished. If a clump of packets arrives the queue can get
longer.

There is another source of queuing. That's when the link speed changes.
Suppose the server has a 100 megabit connection, the backbone has gigabit
links, and the last hop is 1 megabit. (just to pick some round numbers) If
the server sends a clump of packets, they go out at 100 megabits, probably
won't hit much queuing delay in the backbone, but then they pile up on the
last hop.



--
These are my opinions, not necessarily my employer's. I hate spam.
Hal Murray
2013-08-05 22:13:16 UTC
Permalink
brooke-***@public.gmane.org said:
> This is the only low cost receiver I've found that has 10 Hz update rate and
> that's very important for aircraft flight control systems.

Is there real information in all of those updates, or is it just doing the
traditional 1 per second calculations and extrapolating the next 9 steps
until it gets new data?

One of Garmin's has a 5Hz update rate.

--
These are my opinions. I hate spam.
Hal Murray
2014-02-19 19:23:56 UTC
Permalink
brooke-***@public.gmane.org said:
> I wonder if NIST has one of the GWR gravitymeters on a pier and uses that to
> discipline their fountain clocks for the elevation change of the pier or if
> that's done for the GPS reference antennas?

Radio astronomers pay serious attention to earth tides. For VLBI, they need
to know where their antennas are located down to a fraction of a wavelength.

An example, from google:

Tidal deformations of the earth from VLBI observations
http://link.springer.com/article/10.1134%2FS1063773712050027


--
These are my opinions. I hate spam.
Hal Murray
2014-07-22 06:01:19 UTC
Permalink
brooke-***@public.gmane.org said:
> I expect that there's date and time information being sent in the header of
> every phone call, maybe even before the first ring along with the Caller ID
> info.

Wiki says CallerID is sent between the first and second ring, and includes
the date and time.
http://en.wikipedia.org/wiki/Caller_ID#Operation


--
These are my opinions. I hate spam.
David Malone
2014-07-22 08:25:32 UTC
Permalink
On Mon, Jul 21, 2014 at 11:01:19PM -0700, Hal Murray wrote:
> > I expect that there's date and time information being sent in the header of
> > every phone call, maybe even before the first ring along with the Caller ID
> > info.

> Wiki says CallerID is sent between the first and second ring, and includes
> the date and time.
> http://en.wikipedia.org/wiki/Caller_ID#Operation

Some old modems will happily decode it for you. I compared it to
NTP at some point last year. The time stamp was only given to the
nearest minute, and for my exchange it was pretty terrible - it was
slow by about 90s for a few months. I was considering adding it to
my leap second measurements, but there didn't seem to be much point.

David.
Hal Murray
2014-11-01 22:33:37 UTC
Permalink
***@pacific.net said:
> The click-click-click... is the self winding. A solenoid vibrates back and
> forth and a pawl and ratchet winds the main spring.

I don't think that's what I was referring to. It was a long time ago so my
memory may be buggy.

The click-click-click... that I remember was loud enough to distract the
class. It only happened occasionally, ballpark of once per month or less.
The minute hand was making a step with each click.

I assumed there was some mechanism to remotely set the clock. I think it
ended up roughly correct. (I didn't even have a watch back then.)

Was there any alternative to SWCC technology in the 1950s time frame?


> There's a heart shaped cam that forces the hour, minute and if the clock has
> one the second hand to 12:00 and holds them there until the sync pulse goes
> away. Note the second hand only was used on clocks in radio stations so they
> could join the network. For that 1 second was close enough.

Is that one cam that gets all the hands or a separate cam for each hand?



--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brooke Clarke
2014-11-01 23:09:43 UTC
Permalink
Hi Hal:

I think there were a number of slave clock systems and some of them could do DST/ST changes and/or catch up from a power
failure.
That very well might have been what you heard.

To me the winding sounds like a muffled air compressor.
The setting sounds like Thunk.
Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Hal Murray wrote:
> ***@pacific.net said:
>> The click-click-click... is the self winding. A solenoid vibrates back and
>> forth and a pawl and ratchet winds the main spring.
> I don't think that's what I was referring to. It was a long time ago so my
> memory may be buggy.
>
> The click-click-click... that I remember was loud enough to distract the
> class. It only happened occasionally, ballpark of once per month or less.
> The minute hand was making a step with each click.
>
> I assumed there was some mechanism to remotely set the clock. I think it
> ended up roughly correct. (I didn't even have a watch back then.)
>
> Was there any alternative to SWCC technology in the 1950s time frame?
>
>
>> There's a heart shaped cam that forces the hour, minute and if the clock has
>> one the second hand to 12:00 and holds them there until the sync pulse goes
>> away. Note the second hand only was used on clocks in radio stations so they
>> could join the network. For that 1 second was close enough.
> Is that one cam that gets all the hands or a separate cam for each hand?
>
>
>

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2014-11-04 19:31:52 UTC
Permalink
***@pacific.net said:
> UARTs are not mandatory, a simple transistor level shifting circuit is all
> that's needed for TTL/RS-232 levels.

I think you are confusing UARTs with level shifters.

In the old days, transistors were expensive enough that a 68000 class CPU was
all that could fit on a chip. All memory and peripherals were off chip.

While not quite mandatory, most of the time it made sense to use a UART chip.
If you had a FPGA, you could put a simple UART in there. (But back then,
FPGAs weren't very big either.) If the CPU was idle, you could bit-bang
things.


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brooke Clarke
2014-11-04 19:58:13 UTC
Permalink
Hi Hal:

Yes for level shifting. Most micro controllers have the UART built-in so that's all that's needed.
I haven't got my boxes so don't know what chips are in them.
Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Hal Murray wrote:
> ***@pacific.net said:
>> UARTs are not mandatory, a simple transistor level shifting circuit is all
>> that's needed for TTL/RS-232 levels.
> I think you are confusing UARTs with level shifters.
>
> In the old days, transistors were expensive enough that a 68000 class CPU was
> all that could fit on a chip. All memory and peripherals were off chip.
>
> While not quite mandatory, most of the time it made sense to use a UART chip.
> If you had a FPGA, you could put a simple UART in there. (But back then,
> FPGAs weren't very big either.) If the CPU was idle, you could bit-bang
> things.
>
>

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2015-03-28 21:07:11 UTC
Permalink
***@pacific.net said:
> PS GPS does not contain Daylight Saving Time or Leap Second bits and so
> can not be used to discipline domestic time

GPS does contain leap second data.


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2016-06-29 19:28:50 UTC
Permalink
***@pacific.net said:
> At one point they were looking into making a GPS time receiver where the
> cable length calibration would be built-in.

How would you do that?

The obvious way is to compare the time you get with a known-good time, but if
you had that, why would you want this new GPS with an unknown cable length.

You might be able to do it by measuring the DC drop. Getting enough accuracy
seems tough.


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brooke Clarke
2016-06-29 20:08:15 UTC
Permalink
Hi Hal:

I think the cal process is essentially a time domain reflection measure of cable length. The GPS receiver and the cable
cal hardware would be in the antenna unit.
The 1 PPS signal would be aligned at the output of the cable.

--
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
The lesser of evils is still evil.

-------- Original Message --------
> ***@pacific.net said:
>> At one point they were looking into making a GPS time receiver where the
>> cable length calibration would be built-in.
> How would you do that?
>
> The obvious way is to compare the time you get with a known-good time, but if
> you had that, why would you want this new GPS with an unknown cable length.
>
> You might be able to do it by measuring the DC drop. Getting enough accuracy
> seems tough.
>
>

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Scott McGrath
2016-06-30 08:19:43 UTC
Permalink
If the nominal velocity of propagation is known and length is known delay is easily determined mathematically

Time Domain Reflectrometry is the usual technique for finding cable length but even there the cables NVP is an essential parameter if you want to compute length but not essential in time nuts application because we are interested in delay which a TDR reads directly When using a TDR its best if cable is unterminated as the discontinuity at the end is helpful as a marker. Also most TDRs like the Tek 1502 can put 100v or more on the cable which will blow most GPS antennas


> On Jun 29, 2016, at 4:08 PM, Brooke Clarke <***@pacific.net> wrote:
>
> Hi Hal:
>
> I think the cal process is essentially a time domain reflection measure of cable length. The GPS receiver and the cable cal hardware would be in the antenna unit.
> The 1 PPS signal would be aligned at the output of the cable.
>
> --
> Have Fun,
>
> Brooke Clarke
> http://www.PRC68.com
> http://www.end2partygovernment.com/2012Issues.html
> The lesser of evils is still evil.
>
> -------- Original Message --------
>> ***@pacific.net said:
>>> At one point they were looking into making a GPS time receiver where the
>>> cable length calibration would be built-in.
>> How would you do that?
>>
>> The obvious way is to compare the time you get with a known-good time, but if
>> you had that, why would you want this new GPS with an unknown cable length.
>>
>> You might be able to do it by measuring the DC drop. Getting enough accuracy
>> seems tough.
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
David
2016-06-30 14:59:06 UTC
Permalink
The Tektronix 1502 uses a tunnel diode pulser to produce a 50
picosecond output step of about 200 millivolts. There is a misprint
in the theory section of the service manual which says "400 V" instead
of "400mV".

On Thu, 30 Jun 2016 04:19:43 -0400, you wrote:

>If the nominal velocity of propagation is known and length is known delay is easily determined mathematically
>
>Time Domain Reflectrometry is the usual technique for finding cable length but even there the cables NVP is an essential parameter if you want to compute length but not essential in time nuts application because we are interested in delay which a TDR reads directly When using a TDR its best if cable is unterminated as the discontinuity at the end is helpful as a marker. Also most TDRs like the Tek 1502 can put 100v or more on the cable which will blow most GPS antennas
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Scott McGrath
2016-06-30 23:40:05 UTC
Permalink
This is highly dependent on the TDR especially ones designed for long twisted pair runs where a high voltage pulse is used to overcome resistive losses

Content by Scott
Typos by Siri

> On Jun 30, 2016, at 10:59 AM, David <***@gmail.com> wrote:
>
> The Tektronix 1502 uses a tunnel diode pulser to produce a 50
> picosecond output step of about 200 millivolts. There is a misprint
> in the theory section of the service manual which says "400 V" instead
> of "400mV".
>
>> On Thu, 30 Jun 2016 04:19:43 -0400, you wrote:
>>
>> If the nominal velocity of propagation is known and length is known delay is easily determined mathematically
>>
>> Time Domain Reflectrometry is the usual technique for finding cable length but even there the cables NVP is an essential parameter if you want to compute length but not essential in time nuts application because we are interested in delay which a TDR reads directly When using a TDR its best if cable is unterminated as the discontinuity at the end is helpful as a marker. Also most TDRs like the Tek 1502 can put 100v or more on the cable which will blow most GPS antennas
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Charles Steinmetz
2016-07-01 02:31:13 UTC
Permalink
The 1503 was Tek's long-line (10km) TDR analyzer. IIRC, it put
haversine pulses of about 5v onto the line under test (~10v open
circuit). The 1502 (~600m line length) put pulses of about 200mV onto
the line under test (~400mV open circuit).

Best regards,

Charles


Scott wrote:

> This is highly dependent on the TDR especially ones designed for long twisted pair runs where a high voltage pulse is used to overcome resistive losses
>
>> David had written:
>>
>> The Tektronix 1502 uses a tunnel diode pulser to produce a 50
>> picosecond output step of about 200 millivolts. There is a misprint
>> in the theory section of the service manual which says "400 V" instead
>> of "400mV".
>>
>>> Scott had written:
>>>
>>> Time Domain Reflectrometry is the usual technique for finding cable length but even there the cables NVP is an essential parameter if you want to compute length but not essential in time nuts application because we are interested in delay which a TDR reads directly When using a TDR its best if cable is unterminated as the discontinuity at the end is helpful as a marker. Also most TDRs like the Tek 1502 can put 100v or more on the cable which will blow most GPS antennas


_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Charles Steinmetz
2016-06-30 19:26:35 UTC
Permalink
Scott wrote:

> When using a TDR its best if cable is unterminated as the discontinuity at the end is helpful as a marker.

Someone else previously asked how you're expected to do TDR on a GPS
antenna cable since everything is matched (therefore, presumably, no
reflections).

There are ALWAYS reflections. As one message in this thread noted, TDR
easily shows sharp bends in coax. Plus, a GPS antenna will be "matched"
only at a relatively narrow band of frequencies centered on the
frequency(ies) it receives (L1, L1+L2, etc.). Also note that GPS
antennas specified as having 50 ohm outputs are often (if not usually)
used with 75 ohm cable. (Trimble actually advises this.)

You can do the TDR situationally (when the receiver is powered up, and
whenever an antenna fault is cleared), or repetitively during operation
(e.g., once a minute). I see no reason why either strategy would not
work well on a GPS antenna cable.

There is no need to use signals that would damage a GPS antenna preamp.

Best regards,

Charles


_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Dr. David Kirkby (Kirkby Microwave Ltd)
2016-06-30 21:03:57 UTC
Permalink
On 30 June 2016 at 09:19, Scott McGrath <***@gmail.com> wrote:

> If the nominal velocity of propagation is known and length is known delay
> is easily determined mathematically
>

Except that coax does not have a uniform impedance or velocity factor. Both
will vary as a function of position and frequency. How relevant this is
depends on the accuracy you require, but since it is time-nuts, it is
reasonable to assume that such a simplistic method is not of the standard
expected on time-nuts.

Dave
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Scott McGrath
2016-06-30 22:05:43 UTC
Permalink
I'd say it depends on the Time Nut 😀. NVP X Distance will get you close and for a beginning time nut is a worthwhile exercise

To improve delay calculations now you need instrumentation that not all especially beginning time nuts own. I've got a 20 Ghz Agilent scope/TDR along with a 110 Ghz network analyzer and just putting a cable on the network analyzer and handling it you can see the characteristics change So yes you are correct that the simple NVPxDistance is not suitable for advanced time nuttery. In fact the delay will change with temperature and barometric pressure unless you are using hardline and that's subject to humidity unless filled with dry nitrogen under pressure.

Content by Scott
Typos by Siri

> On Jun 30, 2016, at 5:03 PM, Dr. David Kirkby (Kirkby Microwave Ltd) <***@kirkbymicrowave.co.uk> wrote:
>
>> On 30 June 2016 at 09:19, Scott McGrath <***@gmail.com> wrote:
>>
>> If the nominal velocity of propagation is known and length is known delay
>> is easily determined mathematically
>
> Except that coax does not have a uniform impedance or velocity factor. Both
> will vary as a function of position and frequency. How relevant this is
> depends on the accuracy you require, but since it is time-nuts, it is
> reasonable to assume that such a simplistic method is not of the standard
> expected on time-nuts.
>
> Dave
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Michael Wouters
2016-06-30 22:13:35 UTC
Permalink
As a practical matter, in the lab we seldom need a cable delay
measured to better than +/- 0.5 ns, which we usually do as a time
interval measurement, with a 1 pps into a tee on channel A of a TIC
and then the cable from the tee to channel B.

For cables up to 40 m or so, just measuring the physical length is as
accurate (experimentally determined!).

A while back, EURAMET ran a pilot where a few lengths of cable were
circulated amongst a number of NMIs who then had to measure the
delays. The reported scatter was about +/- 1 ns, as I recall. This
mainly came down to differences in test signals, trigger levels etc.

Cheers
Michael


On Fri, Jul 1, 2016 at 7:03 AM, Dr. David Kirkby (Kirkby Microwave
Ltd) <***@kirkbymicrowave.co.uk> wrote:
> On 30 June 2016 at 09:19, Scott McGrath <***@gmail.com> wrote:
>
>> If the nominal velocity of propagation is known and length is known delay
>> is easily determined mathematically
>>
>
> Except that coax does not have a uniform impedance or velocity factor. Both
> will vary as a function of position and frequency. How relevant this is
> depends on the accuracy you require, but since it is time-nuts, it is
> reasonable to assume that such a simplistic method is not of the standard
> expected on time-nuts.
>
> Dave
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Greg Dowd
2016-06-29 20:11:36 UTC
Permalink
We have a smart antenna where the receiver is in the antenna housing and the link to the timing receiver is digital over coax. On that, we can run a variant of DTI (DocSis Timing Interface) which calibrates the cable delay automagically. Not sure what the original question was but yes, it's possible.

-----Original Message-----
From: time-nuts [mailto:time-nuts-bounces+greg.dowd=***@febo.com] On Behalf Of Hal Murray
Sent: Wednesday, June 29, 2016 12:29 PM
To: Discussion of precise time and frequency measurement
Cc: ***@megapathdsl.net
Subject: [time-nuts] Cable length calibration

EXTERNAL EMAIL


***@pacific.net said:
> At one point they were looking into making a GPS time receiver where
> the cable length calibration would be built-in.

How would you do that?

The obvious way is to compare the time you get with a known-good time, but if you had that, why would you want this new GPS with an unknown cable length.

You might be able to do it by measuring the DC drop. Getting enough accuracy seems tough.


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Poul-Henning Kamp
2016-06-29 20:18:49 UTC
Permalink
--------
In message <***@ip-64-139-1-69.sjc.megapath.net>, Hal Mu
rray writes:

>> At one point they were looking into making a GPS time receiver where the
>> cable length calibration would be built-in.
>
>How would you do that?

TDR ?

If it wasn't behind a choke, the inrush current to the antenna
preamp power filtering capacitor could be measured, but the choke
ruins that.

The trouble is how to do it without frying the antenna preamp...


Seriously...

GPS antennas and receivers are cheap, I would just use two GPS antennas
with a known difference in cable-length.


--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
***@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Attila Kinali
2016-06-29 21:29:32 UTC
Permalink
On Wed, 29 Jun 2016 20:18:49 +0000
"Poul-Henning Kamp" <***@phk.freebsd.dk> wrote:

> If it wasn't behind a choke, the inrush current to the antenna
> preamp power filtering capacitor could be measured, but the choke
> ruins that.
>
> The trouble is how to do it without frying the antenna preamp...

That's rather "simple". The impedance of the LNA is anything but
constant over frequency, said capacitor with the choke are one of
those things that make it different than 50 Ohm. You only need to
find one frequency at which the impedance is close to a short
or to an open and then send down a sine at that frequency and
measure the reflection. This is probably easier than doing TDR,
but unfortunately it is still quite involved.

Another way would be to send down a sine at a known frequency,
couple it out at the LNA and inject sharp pulses into the
antenna, at the rate of the sine. This way the whole path
through the antenna and the LNA down to the receiver can
be measured. It even makes it possible to measure the phase
differences between different frequencies (think L1, L2).
But this would require a custom receiver.

Attila Kinali

--
Malek's Law:
Any simple idea will be worded in the most complicated way.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Gary E. Miller
2016-06-29 22:15:28 UTC
Permalink
Yo Poul-Henning!

On Wed, 29 Jun 2016 20:18:49 +0000
"Poul-Henning Kamp" <***@phk.freebsd.dk> wrote:

> >How would you do that?
>
> TDR ?
>
> If it wasn't behind a choke, the inrush current to the antenna
> preamp power filtering capacitor could be measured, but the choke
> ruins that.
>
> The trouble is how to do it without frying the antenna preamp...

The TDR would only put the same voltage on the cable that it already
does. So nothing gets fried. If you are worried about the antenna,
disconnect it before doing the TDR.

It works on live ethernet cables, so it can't be too harsh.

A real good TDR can shouw you every tight bend in a cable.

Or buy a tape measure.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
***@rellim.com Tel:+1 541 382 8588
Poul-Henning Kamp
2016-06-29 22:57:30 UTC
Permalink
--------
In message <***@spidey.rellim.com>, "Gary E. Miller" writes
:

>It works on live ethernet cables, so it can't be too harsh.

Ethernet is incredibly robust, to the tune of a couple hundred volts
pretty much any way you can connect it.

Most GPS antenna preamps will croak before you get up to 10V.


--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
***@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Gary E. Miller
2016-06-29 23:13:58 UTC
Permalink
Yo Poul-Henning!

On Wed, 29 Jun 2016 22:57:30 +0000
"Poul-Henning Kamp" <***@phk.freebsd.dk> wrote:

> Most GPS antenna preamps will croak before you get up to 10V.


The Fluke TDR is 4V max:

http://www.flukenetworks.com/datacom-cabling/installation-tools/ts100-pro-cable-fault-finder-powerbt-bridge-tap



RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
***@rellim.com Tel:+1 541 382 8588
Mike Cook
2016-07-01 07:40:28 UTC
Permalink
> Le 29 juin 2016 à 22:18, Poul-Henning Kamp <***@phk.freebsd.dk> a écrit :
>
> --------
> In message <***@ip-64-139-1-69.sjc.megapath.net>, Hal Mu
> rray writes:
>
>>> At one point they were looking into making a GPS time receiver where the
>>> cable length calibration would be built-in.
>>
>> How would you do that?
>
> TDR ?
>
> If it wasn't behind a choke, the inrush current to the antenna
> preamp power filtering capacitor could be measured, but the choke
> ruins that.
>
> The trouble is how to do it without frying the antenna preamp...
>
>
> Seriously...
>
> GPS antennas and receivers are cheap, I would just use two GPS antennas
> with a known difference in cable-length.
>
Sounds simple, but even after a days reflection I don’t see how you find the complete path delay. You would get the cable delay (OPs concern) provided they were the same antenna/cable type combinations, but not delay induced by the antenna electronics. From another post that delay seems to be non-negligable. I find it curious that antenna manafacturers don’t seem to give this parameter. I looked at some datasheets on Trimble and Leica sites but they don’t have it.
As for me, I measure cable delay by injecting a 1PPS into it through a T and for the RG174 attached to the patch antennas, just sacrificed one by cutting the head off and measuring that.

> --
> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
> ***@FreeBSD.ORG | TCP/IP since RFC 956
> FreeBSD committer | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by incompetence.
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

"The power of accurate observation is commonly called cynicism by those who have not got it. »
George Bernard Shaw

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bob Camp
2016-07-01 11:34:58 UTC
Permalink
Hi


> On Jul 1, 2016, at 3:40 AM, Mike Cook <***@sfr.fr> wrote:
>
>
>> Le 29 juin 2016 à 22:18, Poul-Henning Kamp <***@phk.freebsd.dk> a écrit :
>>
>> --------
>> In message <***@ip-64-139-1-69.sjc.megapath.net>, Hal Mu
>> rray writes:
>>
>>>> At one point they were looking into making a GPS time receiver where the
>>>> cable length calibration would be built-in.
>>>
>>> How would you do that?
>>
>> TDR ?
>>
>> If it wasn't behind a choke, the inrush current to the antenna
>> preamp power filtering capacitor could be measured, but the choke
>> ruins that.
>>
>> The trouble is how to do it without frying the antenna preamp...
>>
>>
>> Seriously...
>>
>> GPS antennas and receivers are cheap, I would just use two GPS antennas
>> with a known difference in cable-length.
>>
> Sounds simple, but even after a days reflection I don’t see how you find the complete path delay. You would get the cable delay (OPs concern) provided they were the same antenna/cable type combinations, but not delay induced by the antenna electronics. From another post that delay seems to be non-negligable. I find it curious that antenna manafacturers don’t seem to give this parameter. I looked at some datasheets on Trimble and Leica sites but they don’t have it.
> As for me, I measure cable delay by injecting a 1PPS into it through a T and for the RG174 attached to the patch antennas, just sacrificed one by cutting the head off and measuring that.

If you dig into the papers on calibrating the delay of the whole antenna, there are a bunch of things they run into. I suspect
the biggest one is quite simple: nobody but TimeNuts care. The survey guys are happy once they know the phase center. The
NIST category guys are going to calibrate it anyway. The cell phone outfits never did any real GPS time calibration, they did
it all on the other end of the system. Way to many cables in a cell site to measure them all ….

Bob

>
>> --
>> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
>> ***@FreeBSD.ORG | TCP/IP since RFC 956
>> FreeBSD committer | BSD since 4.3-tahoe
>> Never attribute to malice what can adequately be explained by incompetence.
>> _______________________________________________
>> time-nuts mailing list -- time-***@febo.com
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>
> "The power of accurate observation is commonly called cynicism by those who have not got it. »
> George Bernard Shaw
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brooke Clarke
2016-07-01 16:04:23 UTC
Permalink
Hi Mike:

For quite a while I was heavily into "chirp" transmissions. These are HF ionosphere radio transmissions that sweep from
2 to 30 MHz at 100 kHz/sec.
In order to "tune" the radio to a specific station (you can not tune by frequency) you need to know the start time
schedule for that specific station (time nuts content).
When GPS became popular the transmitters switched to GPS.
http://www.prc68.com/I/RCS-5A.shtml

You can use a pulse of RF to calibrate the time delay through your HF receiver to get a more accurate time of reception
value. That helps because with a GPS synchronized transmitter you can determine it's great circle distance from you.
Under some conditions you can see a transmission going around the Earth 2 or 3 times.

In a similar way if you used a pair non amplified versions of a GPS antenna back to back you could determine the time
delay of the pair and then divide by two.

--
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
The lesser of evils is still evil.

-------- Original Message --------
>> Le 29 juin 2016 à 22:18, Poul-Henning Kamp <***@phk.freebsd.dk> a écrit :
>>
>> --------
>> In message <***@ip-64-139-1-69.sjc.megapath.net>, Hal Mu
>> rray writes:
>>
>>>> At one point they were looking into making a GPS time receiver where the
>>>> cable length calibration would be built-in.
>>> How would you do that?
>> TDR ?
>>
>> If it wasn't behind a choke, the inrush current to the antenna
>> preamp power filtering capacitor could be measured, but the choke
>> ruins that.
>>
>> The trouble is how to do it without frying the antenna preamp...
>>
>>
>> Seriously...
>>
>> GPS antennas and receivers are cheap, I would just use two GPS antennas
>> with a known difference in cable-length.
>>
> Sounds simple, but even after a days reflection I don’t see how you find the complete path delay. You would get the cable delay (OPs concern) provided they were the same antenna/cable type combinations, but not delay induced by the antenna electronics. From another post that delay seems to be non-negligable. I find it curious that antenna manafacturers don’t seem to give this parameter. I looked at some datasheets on Trimble and Leica sites but they don’t have it.
> As for me, I measure cable delay by injecting a 1PPS into it through a T and for the RG174 attached to the patch antennas, just sacrificed one by cutting the head off and measuring that.
>
>> --
>> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
>> ***@FreeBSD.ORG | TCP/IP since RFC 956
>> FreeBSD committer | BSD since 4.3-tahoe
>> Never attribute to malice what can adequately be explained by incompetence.
>> _______________________________________________
>> time-nuts mailing list -- time-***@febo.com
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
> "The power of accurate observation is commonly called cynicism by those who have not got it. »
> George Bernard Shaw
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
jimlux
2016-07-01 16:53:43 UTC
Permalink
On 7/1/16 9:04 AM, Brooke Clarke wrote:
> Hi Mike:
>
> For quite a while I was heavily into "chirp" transmissions. These are
> HF ionosphere radio transmissions that sweep from 2 to 30 MHz at 100
> kHz/sec.
> In order to "tune" the radio to a specific station (you can not tune by
> frequency) you need to know the start time schedule for that specific
> station (time nuts content).
> When GPS became popular the transmitters switched to GPS.
> http://www.prc68.com/I/RCS-5A.shtml
>

I'm building a satellite (actually 2 of them) that is, among other
things, designed to receive these transmissions. It turns out that
accurately measuring the "propagation delay" through the receiver (as in
from "EM wavefront" to "time stamped samples in the output stream" is
non trivial.

There's some phase shift/time delay from the physical interaction with
the antenna and the load impedance presented by the LNA. Then there's
the filters and amplifiers in the analog chain. Finally, there's the ADC
sampling (pipeline delay between voltage at sampling instant to when
bits appear at the output) delay, and the various delays through the
digital signal processing (which is fortunately deterministic, but
non-trivial to actually "measure")

Fortunately, I only claim 10 microsecond timing accuracy so the 100
kHz/second chirp means that the downconverted stream might have a
frequency error of 1 Hz. (that is, if I tell the receiver that the
chirp starts at 12:34:56.001, and I actually start the ramp at
12:34:56.00101, I'll see a 1 Hz error in frequency.. if I have a several
kHz output bandwidth, it will still be in there.

Given that ionospheric delays and propagation delays are substantially
longer than 10 microseconds, this isn't an issue. 1km is 3 microseconds,
1000km is 3 milliseconds (or 300 Hz).



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Tim Shoppa
2016-07-01 00:52:45 UTC
Permalink
We are all time nuts, so there's an obvious answer: What you do, is raise
the GPS up to a height the same as the cable length. You then drop it,
measure the time until it hits the ground, and use d = 0.5 a * t * t to
calculate d. Then you correct for the velocity factor.

Tim N3QE

On Wed, Jun 29, 2016 at 3:28 PM, Hal Murray <***@megapathdsl.net> wrote:

>
> ***@pacific.net said:
> > At one point they were looking into making a GPS time receiver where the
> > cable length calibration would be built-in.
>
> How would you do that?
>
> The obvious way is to compare the time you get with a known-good time, but
> if
> you had that, why would you want this new GPS with an unknown cable length.
>
> You might be able to do it by measuring the DC drop. Getting enough
> accuracy
> seems tough.
>
>
> --
> These are my opinions. I hate spam.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
jimlux
2016-07-01 03:45:43 UTC
Permalink
On 6/30/16 5:52 PM, Tim Shoppa wrote:
> We are all time nuts, so there's an obvious answer: What you do, is raise
> the GPS up to a height the same as the cable length. You then drop it,
> measure the time until it hits the ground, and use d = 0.5 a * t * t to
> calculate d. Then you correct for the velocity factor.
>

or rig it as a pendulum, and correct for the gravitational pull of the
moon and sun<grin>

_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2017-12-08 10:59:29 UTC
Permalink
What's the Q of a Helmholtz Resonator? What do I do to make a high(er) Q
version?

With a narrow band filter, it might make a neat demo/toy to pull an audio signal out of the noise. With 2 at different frequencies you could demonstrate FSK.

--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Dana Whitlow
2017-12-08 11:57:56 UTC
Permalink
Use of a smaller opening would be the first strategy for getting higher Q.
Making sure that the walls of the vessel were solid reflectors would be
an important factor, too.

I noted that several of his sustained oscillators were basically either
relaxation or blocking oscillators, neither of which is noted for good
phase noise performance. Even the pendulum clock mechanism
was interfering severely with the pendulum's motion- if you look
closely you can see that the pendulum bob's position versus time
function was a severely clipped waveform.

The flex hose demonstration was interesting in that different regimes
of swinging speed resulted in oscillation in different modes. I wonder
why. But in hearing people learning to play different musical instruments,
mostly wind instruments but also including the violin, I was once
moved to say that playing these devices the art was in making an
oscillator run in a resonator mode other than the "natural" one.

For an interesting look at oscillating modes and a really oddball
sustained oscillator, view https://www.youtube.com/watch?v=m6631u7d4E0.
and/or google "mercury beating heart". The electrochemical effect
makes the blob oscillate between a hunched-up shape and a flattened
shape.

If given time, this hunching oscillation "pumps" a degenerate parametric
oscillation between the two triangular shapes at nominally half the rate
of the original oscillation. I first saw this demonstrated in high school,
but the demonstrator also could not get a sustained oscillation. I
thought about that for a while and decided to try a little external stimulus
in the form of low voltage DC from an external supply. After a little
optimization it worked beautifully and could run for hours on end with
little of no attention, giving me the luxury of trying a range of different
blob sizes. With different sizes I could get sustained parametric
oscillation in four different modes: 2-sided, 3-sided (as seen in the
You-Tube clip), 4-sided, and with difficulty even 5-sided.

Of course I didn't really understand what was going on at the time,
and didn't arrive at the parametric oscillation theory until years later.
BTW, I used a baking soda solution instead of a chromium-based
chemistry, and an electrode coming down from the top center, with
a large ring surrounding the mercury blob as the other electrical
connection.

Solution concentration, voltage, electrode tip height, and electrical
polarity were the parameters that had to be adjusted for best
performance. A mercury blob about one cm across in its resting
state seemed to be a good starting point. For polarity, use the
one that results in the blob's hunching up when the electrode tip is
gradually lowered into contact with the blob. If the other stuff
is not too far off, it will quickly take off oscillating at that point.

I've been wondering whether this could be made to work with
Galinstan obtained from modern-day clinical thermometers,
instead of that dreaded mercury. A worthwhile experiment
to try.

I never did get around to measuring the Alan Variance of one
of these oscillators; indeed, I didn't even hear of the concept
until years later. It's probably not up to Time-Nuts' standards.

Dana

On Fri, Dec 8, 2017 at 4:59 AM, Hal Murray <***@megapathdsl.net> wrote:

> What's the Q of a Helmholtz Resonator? What do I do to make a high(er) Q
> version?
>
> With a narrow band filter, it might make a neat demo/toy to pull an audio
> signal out of the noise. With 2 at different frequencies you could
> demonstrate FSK.
>
> --
> These are my opinions. I hate spam.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/
> mailman/listinfo/time-nuts
> and follow the instructions there.
>
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2017-12-09 07:39:42 UTC
Permalink
***@gmail.com said:
> The flex hose demonstration was interesting in that different regimes of
> swinging speed resulted in oscillation in different modes. I wonder why.

It depends on the speed of the air going through the tube.

https://www.youtube.com/watch?v=4aJ36-TlPD4

http://www.exo.net/~pauld/activities/AAAS/aaas2001.html
http://www.exo.net/~pauld/summer_institute/summer_day13music/Whirly.html


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bill Byrom
2017-12-10 00:30:52 UTC
Permalink
The Q of Helmholtz resonators is derived here:
https://en.wikibooks.org/wiki/Acoustics/Flow-induced_Oscillations_of_a_Helmholtz_Resonator

Some Q measurements of bottles are described here:
https://math.dartmouth.edu/archive/m5f10/public_html/proj/ArainGolvach.pdf

--
Bill Byrom N5BB

On Sat, Dec 9, 2017, at 01:39 AM, Hal Murray wrote:
>
> ***@gmail.com said:
> > The flex hose demonstration was interesting in that different regimes of
> > swinging speed resulted in oscillation in different modes. I wonder why.
>
> It depends on the speed of the air going through the tube.
>
> https://www.youtube.com/watch?v=4aJ36-TlPD4
>
> http://www.exo.net/~pauld/activities/AAAS/aaas2001.html
> http://www.exo.net/~pauld/summer_institute/summer_day13music/Whirly.html
>
>
> --
> These are my opinions. I hate spam.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Magnus Danielson
2017-12-10 11:37:46 UTC
Permalink
Hi,

Recommended reading relating to this is the acoustical and audio
engineering material.

"Acoustical Engineering" by Harry F. Olson is a really good book. I
covers many of the theory work on speakers etc. and uses electrical
models and ways to estimtate their effects. This book I have used to
again and again debunk "new" speaker concepts, as it turns out it has
already been tried before.

AES have released 4 collections of articles out of JAES on speakers. In
there is the moderload of Thiele-Small articles that extends the work of
Olson to more and more refined methods. Estimating the losses and
resonant properties is a key aspect, as they have huge impact on the on
the audio.

Also, there is a good book on acoustics by Beranek, of
Bolt-Beranek-Newman if that rings a bell to Internet old-timers, which
may be applicable.

I could do a more detailed dig in my library if needed, but there is
some good material out there.

Cheers,
Magnus - AES member and used to do professional audio PA system design

On 12/10/2017 01:30 AM, Bill Byrom wrote:
> The Q of Helmholtz resonators is derived here:
> https://en.wikibooks.org/wiki/Acoustics/Flow-induced_Oscillations_of_a_Helmholtz_Resonator
>
> Some Q measurements of bottles are described here:
> https://math.dartmouth.edu/archive/m5f10/public_html/proj/ArainGolvach.pdf
>
> --
> Bill Byrom N5BB
>
> On Sat, Dec 9, 2017, at 01:39 AM, Hal Murray wrote:
>>
>> ***@gmail.com said:
>>> The flex hose demonstration was interesting in that different regimes of
>>> swinging speed resulted in oscillation in different modes. I wonder why.
>>
>> It depends on the speed of the air going through the tube.
>>
>> https://www.youtube.com/watch?v=4aJ36-TlPD4
>>
>> http://www.exo.net/~pauld/activities/AAAS/aaas2001.html
>> http://www.exo.net/~pauld/summer_institute/summer_day13music/Whirly.html
>>
>>
>> --
>> These are my opinions. I hate spam.
>>
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-***@febo.com
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Ulrich Rohde via time-nuts
2017-12-10 01:57:38 UTC
Permalink
 
This is a very nice technical discussion ...
 
Ulrich
 
 
In a message dated 12/9/2017 7:31:15 PM Eastern Standard Time, ***@radio.sent.com writes:

 
The Q of Helmholtz resonators is derived here:
https://en.wikibooks.org/wiki/Acoustics/Flow-induced_Oscillations_of_a_Helmholtz_Resonator

Some Q measurements of bottles are described here:
https://math.dartmouth.edu/archive/m5f10/public_html/proj/ArainGolvach.pdf

--
Bill Byrom N5BB

On Sat, Dec 9, 2017, at 01:39 AM, Hal Murray wrote:
>
> ***@gmail.com said:
> > The flex hose demonstration was interesting in that different regimes of
> > swinging speed resulted in oscillation in different modes. I wonder why.
>
> It depends on the speed of the air going through the tube.
>
> https://www.youtube.com/watch?v=4aJ36-TlPD4
>
> http://www.exo.net/~pauld/activities/AAAS/aaas2001.html
> http://www.exo.net/~pauld/summer_institute/summer_day13music/Whirly.html
>
>
> --
> These are my opinions. I hate spam.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2018-04-23 04:28:27 UTC
Permalink
***@pacific.net said:
> Isotopes of an element differ in the number of neutrons.  The chemical
> reactions of an element are governed by the electrons, which are the same
> for all isotopes, so chemical means can not be used to separate the
> isotopes.

That needs a qualification, maybe inserting something like "easily".

Some chemical reactions depend slightly on mass. I was at a talk recently
where the speaker was using isotope ratios to investigate the source of lead
and mercury as pollutants. The key is that the technology for measuring
isotope ratios in now very good. 1E4 is common, at least in research labs
focusing on this area.



--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2018-06-15 04:25:28 UTC
Permalink
***@gmail.com said:
> I've been told that CD player type diodes can be successfully modulated up to
> about 600 MHz, but that going much further is either difficult or perhaps
> impossible.

Are DVD lasers faster?


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Dana Whitlow
2018-06-15 08:47:17 UTC
Permalink
I don't know of any particular reason why a DVD-player laser should be
faster,
since neither CD nor DVD players need to deliberately modulate the lasers
anyway. At least, that's the first blush answer.

However, these diode lasers are generally noisy, apparently due to the
inevitable presence of optical reflections back into the diodes. I once
read that deliberately FM'ing the laser by applying a high frequency RF
tone along with the DC bias current could be helpful in mitigating the
problem. But this was in connection with the *analog* Laser Disk video
recording format. I don't know if the laser noise was ever much of a
problem
with reading the digital formats of CDs and DVDs. If not, then surely the
makers of diodes for these mass market applications would not be investing
effort in making the lasers fast.

On additional factor is that CD players and DVD players use different light
wavelengths: ~780 nm for CDs, ~650 nm for DVDs. It might be that the
difference in semiconductor composition between the two types makes
a significant difference in the response speed. However, I doubt that this
would would be relevant for present purposes, unless Rb also has some
*useful* transitions in the 650 nm regime.

Dana


On Thu, Jun 14, 2018 at 11:25 PM, Hal Murray <***@megapathdsl.net>
wrote:

>
> ***@gmail.com said:
> > I've been told that CD player type diodes can be successfully modulated
> up to
> > about 600 MHz, but that going much further is either difficult or perhaps
> > impossible.
>
> Are DVD lasers faster?
>
>
> --
> These are my opinions. I hate spam.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/
> mailman/listinfo/time-nuts
> and follow the instructions there.
>
_______________________________________________
time-nuts mailing list -- time-***@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Henk Peek
2018-06-15 12:52:41 UTC
Permalink
Cheap (CW) lasers such as CD and short distance 850 and 1300nm telecommunication laser have an optical spectrum
of a few nm width. This optical spectrum width is filed width multiple optical frequencies.
Distributed feedback-lasers generate a single optical frequency. They much more expensive.
The most distributed-feedback lasers have peltier temperature control to set the optical frequency.
The laser output is proportional with the laser current. This laser current modulation is called direct modulation.
The simplicity of direct modulation has always offered most cost-effective transmitters compared to external
modulation techniques using continuous-wave (CW) laser diodes followed by electro-optic (EO) modulators,
or integrated external-modulated lasers (EML) with electro-absorption (EA) modulators. Direct modulation system performance, however, has been limited by the intrinsic chirp (FM) of directly modulated lasers (DMLs), and
induced spectrum broadening.

Henk

On Friday, June 15, 2018 10:47 CEST, Dana Whitlow <***@gmail.com> wrote:

> I don't know of any particular reason why a DVD-player laser should be
> faster,
> since neither CD nor DVD players need to deliberately modulate the lasers
> anyway. At least, that's the first blush answer.
>
> However, these diode lasers are generally noisy, apparently due to the
> inevitable presence of optical reflections back into the diodes. I once
> read that deliberately FM'ing the laser by applying a high frequency RF
> tone along with the DC bias current could be helpful in mitigating the
> problem. But this was in connection with the *analog* Laser Disk video
> recording format. I don't know if the laser noise was ever much of a
> problem
> with reading the digital formats of CDs and DVDs. If not, then surely the
> makers of diodes for these mass market applications would not be investing
> effort in making the lasers fast.
>
> On additional factor is that CD players and DVD players use different light
> wavelengths: ~780 nm for CDs, ~650 nm for DVDs. It might be that the
> difference in semiconductor composition between the two types makes
> a significant difference in the response speed. However, I doubt that this
> would would be relevant for present purposes, unless Rb also has some
> *useful* transitions in the 650 nm regime.
>
> Dana
>
>
> On Thu, Jun 14, 2018 at 11:25 PM, Hal Murray <***@megapathdsl.net>
> wrote:
>
> >
> > ***@gmail.com said:
> > > I've been told that CD player type diodes can be successfully modulated
> > up to
> > > about 600 MHz, but that going much further is either difficult or perhaps
> > > impossible.
> >
> > Are DVD lasers faster?
> >
> >
> > --
> > These are my opinions. I hate spam.
> >
> >
> >
> > _______________________________________________
> > time-nuts mailing list -- time-***@febo.com
> > To unsubscribe, go to https://www.febo.com/cgi-bin/
> > mailman/listinfo/time-nuts
> > and follow the instructions there.
> >
> _______________________________________________
> time-nuts mailing list -- time-***@febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.





_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2018-06-22 18:27:43 UTC
Permalink
***@pacific.net said:
> I'm experimenting with a WSPR beacon transmitter and part of how it works
> depends on pushing the start button at exactly 2 seconds past the minute.

What do you mean by "exactly"?


> SotaBeams, who make the WSPRlite transmitter recommends using http://time.is/
> which seems to be accurate on my cell phone but not on my desktop running
> the stock NTP. Can someone comment on time.is?

How does time on your cell phone compare with your desktop?

ntpq -p on your desktop will give you a lot of information on how good your
desktop time is.


time.is seems to be using javascript over http/TCP. That's not a good start,
but can be good enough, depending on your needs and the implementation
details. I'm generally not impressed with things that don't provide details.
(Maybe I just didn't look in the right place.)

In this type of discussion, a lot depends on your network connections. How
good is the connection to your desktop? How good is the connection to your
cell phone?


> The problem is the transmitter (and the start button) are out in the forest
> where the antenna is located.

What sort of connection do you have to the transmitter? I'm guessing there is
something like a Raspberry Pi out there. Why not run NTP on that system?


--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Van Horn, David
2018-06-22 18:29:02 UTC
Permalink
"Exactly" on the time nuts list... 😊


_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Brian Lloyd
2018-06-22 18:41:50 UTC
Permalink
On Fri, Jun 22, 2018 at 1:27 PM, Hal Murray <***@megapathdsl.net> wrote:

>
> ***@pacific.net said:
> > I'm experimenting with a WSPR beacon transmitter and part of how it works
> > depends on pushing the start button at exactly 2 seconds past the
> minute.
>
> What do you mean by "exactly"?
>

I think WSPR has a +/-2s window for the start of a transmission. It really
isn't all that "exact" and certainly nothing in the way of time-nuts
definition of accuracy.

The most popular WSPR implementation runs on a PC and uses a sound-card to
generate the baseband signal and then uses an SSB transceiver to upconvert
to the desired output frequency. Even worst-case the standard NTP
implementation on Windows is more than sufficient for synchronization.

--



Brian Lloyd
706 Flightline
Spring Branch, TX 78070
***@lloyd.aero
+1.210.802-8FLY (1.210.802-8359)
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bob Bownes
2018-06-22 19:14:23 UTC
Permalink
I can heartily recommend the QRP-Labs wspr beacon kit with GPS. It pull the
time from an attached GPS module and will beacon on 2m and below. Once
assembled and configured with your callsign and the bands you want to xmit
on it Just Plain Works. It can even provide 1pps out.

Pick the right GPS module and you get 10MHz out as well!

Bob


On Fri, Jun 22, 2018 at 2:41 PM, Brian Lloyd <***@lloyd.aero> wrote:

> On Fri, Jun 22, 2018 at 1:27 PM, Hal Murray <***@megapathdsl.net>
> wrote:
>
> >
> > ***@pacific.net said:
> > > I'm experimenting with a WSPR beacon transmitter and part of how it
> works
> > > depends on pushing the start button at exactly 2 seconds past the
> > minute.
> >
> > What do you mean by "exactly"?
> >
>
> I think WSPR has a +/-2s window for the start of a transmission. It really
> isn't all that "exact" and certainly nothing in the way of time-nuts
> definition of accuracy.
>
> The most popular WSPR implementation runs on a PC and uses a sound-card to
> generate the baseband signal and then uses an SSB transceiver to upconvert
> to the desired output frequency. Even worst-case the standard NTP
> implementation on Windows is more than sufficient for synchronization.
>
> --
>
>
>
> Brian Lloyd
> 706 Flightline
> Spring Branch, TX 78070
> ***@lloyd.aero
> +1.210.802-8FLY (1.210.802-8359)
> _______________________________________________
> time-nuts mailing list -- time-***@lists.febo.com
> To unsubscribe, go to https://lists.febo.com/cgi-
> bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2018-12-02 12:01:17 UTC
Permalink
***@gmail.com said:
> Can anyone point me to a readily-accessible reference that would address the
> following largely-intuitive notions that I carry around with me:

Google for P-wave and S-wave. The wikipedia pages are a good start.


--
These are my opinions. I hate spam.




_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Loading...