Laser power control - help me design a servo attenuator / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Laser power control - help me design a servo attenuator

previous 1 2 3 next

Move Thread

LAN_403

Steve Conner

Mon Feb 20 2012, 12:27PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I think speaking to a Coherent applications engineer might be of use. These lasers are so expensive, you have a right to expect some tech support.

They should be able to advise you on how stable the laser is supposed to be, and maybe give tips on improving it. It could simply be in need of a service, my feeling is that the output of a precision single-mode laser like the Verdi should be rock steady.

Shrad

Mon Feb 20 2012, 11:23PM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

the laser should be fine

I already used a verdi V10 and they are top notch lasers when regularly checked by coherent, so I can't think of anything

anyway, there is still something you might check, it's the dessicant cartridge which is in a box you can open, I think, without voiding warranty, and it must be of a blue colour if I recall. if it is red/orange, it needs to be replaced or reprocessed, and this could induce power fluctuations if you pulse humidity

if not, there must be something in your setup, but I can't figure out myself, sorry

maybe the pump diodes? your Ti:Sapphire suffering parasitic reflexions inside the mount? thermal instability inducing mechanical drift of your mounts?

anyway, you should correct the problem instead of driving or attenuating the verdi with feedback from your Ti:Sapphire, as it is not a good thing if you have an underlying problem

good luck finding the problem and please describe the solution if you got it, I'm interested!

hboy007

Tue Feb 21 2012, 06:37PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Shrad wrote ...

anyway, there is still something you might check, it's the dessicant cartridge which is in a box you can open, I think, without voiding warranty, and it must be of a blue colour if I recall.

Where would I have to look for the silica gel? The laser head itself is said to be hermetically sealed and I couldn't find any compartments at first glance.

Shrad

Wed Feb 22 2012, 07:56AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

it's a black box additional to the power supply which is connected to the interlock and air hoses

I don't remember if it is user serviceable, I had a look with a coherent guy, but there should be stickers

if I recall all you need is an oven at 60Â°C overnight and the cartridge should be revived at least for some time tille a replacement arrives

anyway, what I would do in order to attenuate the beam with precision and speed is use an AOM with a PID loop and get feedback from a beampick at the output of your Ti:Sapphire

be sure not to use a fast reverse biased PIN photodiode in that case as the control loop must keep slow enough not to oscillate

the pid controller could be an automation device, if not an analog error amplifier you should have in the lab could do the trick to get the 0-5V or 0-1V command for the AOM

hboy007

Thu Feb 23 2012, 09:51PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

OK, I've managed to hook up a serial cable to the Verdi and talk to him.

"Big Scary Laser" wrote ...

Verdi V-10> ?S 0
Verdi V-10> ?P 0.000
Verdi V-10> ?C 19.54
Verdi V-10> ?D1C 19.87
Verdi V-10> ?D2C 19.22
Verdi V-10> ?D1T 22.99
Verdi V-10> ?D2T 23.10
Verdi V-10> S=1
Verdi V-10> P=2.00
Verdi V-10> ?P 0.030
Verdi V-10> ?P 0.164
Verdi V-10> ?P 0.174
Verdi V-10> ?P 0.214
Verdi V-10> ?P 0.264
Verdi V-10> ?P 0.363
Verdi V-10> ?P 0.517
Verdi V-10>
Verdi V-10> ?P 1.953
Verdi V-10> ?P 1.953
Verdi V-10> ?C 25.09
Verdi V-10> ?D1C 25.46
Verdi V-10> ?D2C 24.70
Verdi V-10> ?D1T 23.07
Verdi V-10> ?D2T 23.17
Verdi V-10> ?P 1.953
Verdi V-10> S=0
Verdi V-10> ?P 0.000
Verdi V-10> P=0.01

\?P\r\n for example queries the current power level as monitored by the internal diode.
\?S\r\n queries the shutter status.
D1C : Diode Array 1 current (in steps of 0.01A)
D1T : Diode Array 1 temperature (in steps of 0.01K)
C : average? current
X=Y : set parameter X to Y (example: P=2.00\r\nS=1\r\n sets the output power to 2.00W and opens the shutter.
...

The CoherentGUI application also allows to log temperatures (D1T, D2T) to an SQL database for further analysis.
Unfortunately, the power reading is quite coarse. However, there is an option to connect to the external Fieldmaster power meter and log its values as well. Since the quantization is in steps of 10mW even for 200mW setpoint, I will have to further investigate whether the servo loops work with a higher resolution (judging by the raw drive values, the controller seems to work with 16bit signed integers internally).

hboy007

Fri Feb 24 2012, 11:36PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

did some steady state measurements today. I also found out that the laser head itself produces 60-80W of heat (100W max.). The riser / heat sink underneath the laser (as described in the whitepaper) is mainly a watercooled plate. However, with a throughput of 100ml / min., the base plate of the laser can exceed the water temperature by > 20K (temp.: 33-36Â°C). Even with the laser off, the base plate temperature stays 3-4K above the water temperature (presumably due to LBO heating).

The chiller is designed to pump >15L/min at 0mbar and still some 5-8L at 200mbar static pressure. With a piece of narrow hose attached, the throughput of the chiller alone dropped to ~1L/min.

I saw that the laser is aligned by two pins so removing the laser head clamps and the cover of the chiller won't imply too much readjustment. Is cleaning the heat sink / heat exchanger plate underneath the laser head worth the pain?

1330126585 1667 FT134515 Verdi V10 Normalized Power Over Time

...

Sat Feb 25 2012, 02:23AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

For compensating for slow variations, the linear attenuator with a stepper motor sounds like a good idea, although I would recommend getting one of the round attenuators, like

There was also a good comment that you will want to make sure that you filter the incoming beams well (a piece of scott glass would work well for the verdi, the ti:saph probably does not even need a filter as I would imagine the pump leakage should be pretty minimal), and keep outside light out of the feedback system.

You might want to do a study looking at the higher frequency noise components, if you have an oscilliscope with a FFT function you can use that in combination with a photodiode on a slow sweep rate to get an idea of what the mid frequency variations are like. The stepper motor system can probably lock it down to about 1Hz if you design the control loop well, but if your experiment has a higher bandwidth than that and the laser has significant noise about that you will be screwed.

You might also try the alignment on the ti:saph, I am not sure the specifics of your ti:saph, but those numbers do sound a little high for a properly aligned laser.

hboy007

Wed Feb 29 2012, 09:29AM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Yesterday, I tried a brute force experiment: I set the chiller to 5Â°C and waited 3 hours for the damn heat exchanger and laser to cool down. During the cool-down, I set up a fiber coupler facing a Si photodiode. Using a 10kOhm terminator, I connected it to the scope and watched the overall intensity of the exiting light.
The experiment was done with a SM450, the Verdi V10 was operated at 1.00W output.
No information about offset and amplitude of the signal is given due to dependence on optical adjustment, diode type and shunt resistor. Instead, the signal amplitude in units on the screen is given for comparison among the scope pictures.

Observations:

* cool-down is really slow (< 0.7 K / h for chiller temp set from 17Â°C to 5Â°C at the beginning of the experiment)
* for all times, no fluctuations can be observed when the beam itself is monitored
* at 33Â°C, no fluctuations can be observed in the modal structure when the beam is expanded using a short focal length lens and shone onto a piece of paper
* readjustment of the fiber insertion coupler was necessary beyond 27Â°C to reestablish a usable intensity at the end of the fiber
* the fluctuations of the insensity as seen by the fiber are temperature dependent.
* the suppressed polarisation component shows a doughnut-shaped profile at Tbase = 33 .. 35Â°C
* only minor and higher frequency fluctuations can be generated by moving the fiber, knocking on the table or mounted components
* experiment was done without disturbing the optical setup manually. The chiller was turned off for a minute to see the effects of its vibrations

Interpretation:

* the fluctuations are not caused by vibrations of the chiller
* fine channels inside the heat exchanger may be blocked or covered with dirt
* the total laser output may be well-regulated irrespective of the beam quality and etalon parameters
* the beam walks away over temperature, modal profile and/or wavefront characteristics change

Conclusion:

The light intensity coupled into a single mode fiber can be used as an indicator for laser performance. Our Coherent Verdi V10 exhibits instability when the base plate temperature exceeds 26.6Â°C (*). Thorough cleaning of the heat exchanger and an upgrade to a larger hose diameter is suggested.
Despite the manufacturer's advice not to use heat conducting compound, the application of silver-filled heat conducting grease is considered. Despite the manufacturer's advice, chiller water temperature will be held below 20Â°C to achieve <22Â°C base plate temperature while still avoiding condensation.

(*) max. operating temperature as described by Coherent goes up to 50Â°C, followed by thermal shutdown

PS: I also closed some apertures along the beam path a little bit more, discarding the outer paraxial rays. I'm running an experiment right now, min/max/average power used after the fiber is 11.7ÂµW/11.2ÂµW/11.4ÂµW. Good enough.

Carbon_Rod

Sun Mar 04 2012, 09:07AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

Glad your TEC unit finally started to behave hboy007, but note most fiber optic collimation rod lenses will warp out of alignment in abnormal temperature ranges.

Ignoring this phenomena could cause accelerated degradation given higher internal diffuse reflections are known to damage diode stacks.

Cheers,
Rod

Steve Conner

Sun Mar 04 2012, 09:25AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Hboy's last post wasn't too clear, but it sounds like the laser was unstable because it was overheating, and he fixed it just by turning the chiller down?

See, by taking a step back and looking at the real problem, we saved a lot of time and effort.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob