Discussion:
a newbie question: where can I purchase 794.7 nm VCSEL for building CPT rubidium clock?
(too old to reply)
mimitech mimitech
2018-06-04 13:31:56 UTC
Permalink
Hi, list member

I'm planning to build a CPT (coherent-population-trapping) rubidium clock
as my next hobby project. The main purpose is to learn the principles
behind CPT rubidium clock, and hopefully got similar or better performance
than commercial miniature rubidium clock such as FE-5680A.

To build this clock, I think for the Rb 87 vapor cell I could use a cell
removed from commercial rubidium clock. Another import component is 794.7
nm VCSEL laser diode. Does anyone know some good sources I could buy small
quantities (< 5pcs) of this special wavelength VCSEL (perfer a model
targeting for CPT rubidium clock applications), and what’s approximate
price? if necessary you could send relevant information to my email
<mimitech AT gmail DOT com>. Has anyone else already done or is doing
similar projects? I'd be very appreciative to know what your thoughts and
suggestions for this project. Thanks very much.
_______________________________________________
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.
mimitech mimitech
2018-06-09 07:01:54 UTC
Permalink
Thanks, Henk.

This is a very interesting paper. The 1560nm VCSEL is indeed very easy to
source thanks to widely application in Telecom, but the frequency double
and stabilization scheme in the paper is very complex. the purpose of this
paper to use 1560nm is because of its higher spectral purity than
customized and small-scale manufactured 780/795nm VCSEL, to achieve better
atomic clock performance (short-term stability).

If I could find some simple frequency multiplier schemes, it is worth
considering using a 1560 nm VCSEL. Anyway, I need to do further research on
it. Thanks.

mimitech.
Rb vapor-cell clock demonstration with a frequency-doubled telecom laser
Applied Optics Vol. 57, Issue 16, pp. 4707-4713 (2018) •
https://doi.org/10.1364/AO.57.004707
Abstract
We employ a recently developed laser system, based on a low-noise telecom
laser emitting around 1.56 μm, to evaluate its impact on the performance of
an Rb vapor-cell clock in a continuous-wave double-resonance scheme. The
achieved short-term clock instability below 2.5·10−13·𝜏−1/2 demonstrates,
for the first time, the suitability of a frequency-doubled telecom laser
for this specific application. We measure and study quantitatively the
impact of laser amplitude and frequency noises and of the ac Stark shift,
which limit the clock frequency stability on short timescales. We also
report on the detailed noise budgets and demonstrate experimentally that,
under certain conditions, the short-term stability of the clock operated
with the low-noise telecom laser is improved by a factor of three compared
to clock operation using the direct 780-nm laser.
© 2018 Optical Society of America under the terms of the OSA Open Access
Publishing Agreement
Henk Peek
On Mon, 4 Jun 2018 21:31:56 +0800
Post by mimitech mimitech
I'm planning to build a CPT (coherent-population-trapping) rubidium
clock
Post by mimitech mimitech
as my next hobby project. The main purpose is to learn the principles
behind CPT rubidium clock, and hopefully got similar or better
performance
Post by mimitech mimitech
than commercial miniature rubidium clock such as FE-5680A.
Building a CPT clock is slightly more involved than you might think
at first. The laser diode is only one part of it. You will most likely
be able to improve on the short-term stability of the FE-5680 (which
is rather poor). But I doubt you will be able to improve much on
the long term stability, which is where things actually become
interesting,
if you use a naive approach.
Nevertheless, I have not seen many 794/795nm diodes around. The only
one that I have the datasheet of is the one from Vixar.
You might want want to consider going for the D2 line instead of the
D1 line, as 780nm diodes are more commonly available than 795nm. You
will
also need to buy several of those and select the ones that come closest
to the wavelength at the desired opearating conditions (usuall spread
is +/-1nm to +/-10nm). Do not assume you can tune more than 0.1nm with
temperature and current (rule of thumb is that you get about 10GHz
per °C and mA). If you need more tuning range, you will need to add an
external cavity (can give you up to 5nm range), which then needs to be
tuned to the 3.45GHz (ie it's length needs to be approximately 8-9cm).
Alternatively, you can get two S1-0780-XXX from Sacher Laser
(cost IIRC 2500€ each) and keep them 6.9GHz apart (using an optical
PLL).
If you have enough money to spend, I'd go for two Cateye diode laser
CEL's
from Moglabs (cost AFAIK 5000€ each)
No matter what you choose, you will need some wavelength stabilization
scheme. You can either do that with the vapor cell itself or use
an additional cell and do a DVALL or a saturated absorption locking.
Note that this addtional cell will need to be without buffer gas.
An external cell will offer better stability and thus lower noise,
which directly translates into higher stability.
As polarisation scheme, I suggest using σ+/σ- as it seems to be more
robust than the lin/lin schemes.
Attila Kinali
--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson
_______________________________________________
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.
mimitech mimitech
2018-06-09 15:28:29 UTC
Permalink
Just do a little bit research on 1560nm telecom laser transceiver. The
closest wavelength is 1560.61nm (DWDM channel 21) as defined by ITU.
Usually telecom laser transceiver module uses DFB (distributed feedback)
type laser diode instead of VCSEL laser. However, the DFB LD should be
capable of direct modulation, so there maybe not much difference with VCSEL
in terms of usage. For the modulation bandwidth, I guess a 2.5Gbps module
should be OK for 1.71GHz analog modulation.
So, imagine I modulated 1.71GHz microwave signal onto 1560nm laser wave,
then used a KPT non-linear optical crystal to double the frequency,
hopefully I could get 780nm laser with +/-3.42GHz sideband spectrum, that
should be suitable to shine on a Rubidium vapor cell and trigger CPT
effect. I don't know if I understand this process correctly, many details
must be missed.
Another question is, after passing through the KPT doubler crystal, whether
the light remains single mode and linear polarization mode or not? If
someone could provide any relevant information and suggestion, thanks very
much.
Rb vapor-cell clock demonstration with a frequency-doubled telecom laser
Applied Optics Vol. 57, Issue 16, pp. 4707-4713 (2018) •
https://doi.org/10.1364/AO.57.004707
Abstract
We employ a recently developed laser system, based on a low-noise telecom
laser emitting around 1.56 μm, to evaluate its impact on the performance of
an Rb vapor-cell clock in a continuous-wave double-resonance scheme. The
achieved short-term clock instability below 2.5·10−13·𝜏−1/2 demonstrates,
for the first time, the suitability of a frequency-doubled telecom laser
for this specific application. We measure and study quantitatively the
impact of laser amplitude and frequency noises and of the ac Stark shift,
which limit the clock frequency stability on short timescales. We also
report on the detailed noise budgets and demonstrate experimentally that,
under certain conditions, the short-term stability of the clock operated
with the low-noise telecom laser is improved by a factor of three compared
to clock operation using the direct 780-nm laser.
© 2018 Optical Society of America under the terms of the OSA Open Access
Publishing Agreement
Henk Peek
On Mon, 4 Jun 2018 21:31:56 +0800
Post by mimitech mimitech
I'm planning to build a CPT (coherent-population-trapping) rubidium
clock
Post by mimitech mimitech
as my next hobby project. The main purpose is to learn the principles
behind CPT rubidium clock, and hopefully got similar or better
performance
Post by mimitech mimitech
than commercial miniature rubidium clock such as FE-5680A.
Building a CPT clock is slightly more involved than you might think
at first. The laser diode is only one part of it. You will most likely
be able to improve on the short-term stability of the FE-5680 (which
is rather poor). But I doubt you will be able to improve much on
the long term stability, which is where things actually become
interesting,
if you use a naive approach.
Nevertheless, I have not seen many 794/795nm diodes around. The only
one that I have the datasheet of is the one from Vixar.
You might want want to consider going for the D2 line instead of the
D1 line, as 780nm diodes are more commonly available than 795nm. You
will
also need to buy several of those and select the ones that come closest
to the wavelength at the desired opearating conditions (usuall spread
is +/-1nm to +/-10nm). Do not assume you can tune more than 0.1nm with
temperature and current (rule of thumb is that you get about 10GHz
per °C and mA). If you need more tuning range, you will need to add an
external cavity (can give you up to 5nm range), which then needs to be
tuned to the 3.45GHz (ie it's length needs to be approximately 8-9cm).
Alternatively, you can get two S1-0780-XXX from Sacher Laser
(cost IIRC 2500€ each) and keep them 6.9GHz apart (using an optical
PLL).
If you have enough money to spend, I'd go for two Cateye diode laser
CEL's
from Moglabs (cost AFAIK 5000€ each)
No matter what you choose, you will need some wavelength stabilization
scheme. You can either do that with the vapor cell itself or use
an additional cell and do a DVALL or a saturated absorption locking.
Note that this addtional cell will need to be without buffer gas.
An external cell will offer better stability and thus lower noise,
which directly translates into higher stability.
As polarisation scheme, I suggest using σ+/σ- as it seems to be more
robust than the lin/lin schemes.
Attila Kinali
--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson
_______________________________________________
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.
Attila Kinali
2018-06-11 07:52:02 UTC
Permalink
On Sat, 9 Jun 2018 23:28:29 +0800
Post by mimitech mimitech
Just do a little bit research on 1560nm telecom laser transceiver. The
closest wavelength is 1560.61nm (DWDM channel 21) as defined by ITU.
Usually telecom laser transceiver module uses DFB (distributed feedback)
type laser diode instead of VCSEL laser. However, the DFB LD should be
capable of direct modulation, so there maybe not much difference with VCSEL
in terms of usage. For the modulation bandwidth, I guess a 2.5Gbps module
should be OK for 1.71GHz analog modulation.
A 2.5Gbps module should be good even for 3.4GHz modulation. The 3dB frequency
of should be higher than the baud rate used. Besides, even if the 3dB frequency
is below 3.4GHz, that does not matter, it just means that your sidebands will
be damped (second order low pass, IIRC), but you don't need that much power
anyways.
Post by mimitech mimitech
So, imagine I modulated 1.71GHz microwave signal onto 1560nm laser wave,
then used a KPT non-linear optical crystal to double the frequency,
hopefully I could get 780nm laser with +/-3.42GHz sideband spectrum, that
should be suitable to shine on a Rubidium vapor cell and trigger CPT
effect. I don't know if I understand this process correctly, many details
must be missed.
Using a non-linear element will give you lots of intermodulation products.
Ie, your sidebands will be +/-1.71GHz, +/-3.42GHz, +/-6.84GHz,...
So you will have to be a bit carefull with the laser tuning in order to
get the right harmonics.

An alternative approach would be to use an EOM after the doubler
to get the sidebands. These have bandwidths in the order of 10-20GHz,
so applying 3.42GHz shouldn't be a problem.

BTW: a lot of the frequency doubler architectures for high stability
lasers use resonant cavities, to increase efficiency and decrease noise.
You will either have to tune the free spectral range of the cavity
to a divisor of 6.84GHz or use a non-resonant one.
Post by mimitech mimitech
Another question is, after passing through the KPT doubler crystal, whether
the light remains single mode and linear polarization mode or not? If
someone could provide any relevant information and suggestion, thanks very
much.
If I understood the optical process correctly (disclaimer: I'm not
a physicst and have never worked with optical systems), then the multiple
modes are a problem of the laser source, not of the doubler. As multi-modes
are a problem for high speed communication, I expect telecom lasers to
be quite clean. You will have to ensure that you are not operating it
close to a mode jump, though.

A side note on the doubler approach: The group at UniNE around Gaetano
Mileti and Christoph Affolderbach focuses only on double-resonant
Rb standards. That means they do not modulate the laser. As they
have gotten the whole standard to the SNR limit, they are now focusing
on getting the shifts due to laser (frequency and intensity shift)
and cavity (low Q cavity to avoid pulling) down. The current state
of affairs is quite nicely documented in their two papers at 8FSM[1,2,3].

Attila Kinali


[1] Proceedings of the 8th Symposium on Frequency Standards and Metrology
http://conferenceseries.iop.org/conferenceseries/issue/1742-6596/723/1

[2] "High performance vapour-cell frequency standards",
Gharavipour, Affolderbach, Kang, Bandi, Buret, Pellaton, Mileti, 2015
http://iopscience.iop.org/1742-6596/723/1/012006/pdf/1742-6596_723_1_012006.pdf

[3] "Pulsed Optically Pumped Rb clock",
Micalizio, Levi, Godone, Calosso, François, Boudot, Affolderbach, Kang,
Gharavipour, Gruet and Mileti, 2015
http://iopscience.iop.org/1742-6596/723/1/012015/pdf/1742-6596_723_1_012015.pdf
--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson
_______________________________________________
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
2018-06-11 22:52:44 UTC
Permalink
Post by Attila Kinali
On Sat, 9 Jun 2018 23:28:29 +0800
Post by mimitech mimitech
Just do a little bit research on 1560nm telecom laser transceiver. The
closest wavelength is 1560.61nm (DWDM channel 21) as defined by ITU.
Usually telecom laser transceiver module uses DFB (distributed feedback)
type laser diode instead of VCSEL laser. However, the DFB LD should be
capable of direct modulation, so there maybe not much difference with VCSEL
in terms of usage. For the modulation bandwidth, I guess a 2.5Gbps module
should be OK for 1.71GHz analog modulation.
A 2.5Gbps module should be good even for 3.4GHz modulation. The 3dB frequency
of should be higher than the baud rate used. Besides, even if the 3dB frequency
is below 3.4GHz, that does not matter, it just means that your sidebands will
be damped (second order low pass, IIRC), but you don't need that much power
anyways.
Considering that the bandwidth of the reference receiver for SDH/SONET
is 3/4 of the baudrate, and then using a 4-pole Bessel-Thompson filter,
I beg to differ regarding the expected bandwidth. The majority of the
modulated energy will be inside those 3/4 and you don't want much more
as it causes unnecessary problems.

But sure, sidebands will be damped and you can expect 6 dB slope there
from the bare diode.

This is stuff I can actually measure at home as I come to think of it.
Post by Attila Kinali
Post by mimitech mimitech
So, imagine I modulated 1.71GHz microwave signal onto 1560nm laser wave,
then used a KPT non-linear optical crystal to double the frequency,
hopefully I could get 780nm laser with +/-3.42GHz sideband spectrum, that
should be suitable to shine on a Rubidium vapor cell and trigger CPT
effect. I don't know if I understand this process correctly, many details
must be missed.
Using a non-linear element will give you lots of intermodulation products.
Ie, your sidebands will be +/-1.71GHz, +/-3.42GHz, +/-6.84GHz,...
So you will have to be a bit carefull with the laser tuning in order to
get the right harmonics.
Consider the roll-off of the diode/modulator.

Also, expect to servo the amplitude of the modulation for stabilization.
This is separate from the servo of the center of wavelength as well as
the width of the modulation. CPT has more of these loops than a
traditional rubidium.
Post by Attila Kinali
Post by mimitech mimitech
Another question is, after passing through the KPT doubler crystal, whether
the light remains single mode and linear polarization mode or not? If
someone could provide any relevant information and suggestion, thanks very
much.
If I understood the optical process correctly (disclaimer: I'm not
a physicst and have never worked with optical systems), then the multiple
modes are a problem of the laser source, not of the doubler. As multi-modes
are a problem for high speed communication, I expect telecom lasers to
be quite clean. You will have to ensure that you are not operating it
close to a mode jump, though.
The DWDM lasers is quite clean and their temperature-stabilization is
what you want to have to start with. However, I don't think any of the
ones I have will match the frequency needed, but I may be lucky.

Cheers,
Magnus
_______________________________________________
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.

Loading...