Closed loop operation

Go down

Closed loop operation

Post  Alessio on Wed Apr 08, 2009 12:00 pm

Yesterday (7.4.09), I started closing the intensity stabilization loop. Before doing so, I modified a little bit the set-up:

  1. set the AOM efficiency to approx 7%
  2. moved the lens in front of PD12595 closer to the PD to increase the spot size
  3. replaced the BS on the 1st order beam of the AOM with a total reflector

All the changes are summarized in the next figure.


After this I closed the loop. The integrator is a PAR5113, the input is set to A-AC_1s. Roll off high and low frequencies and the gain have been changed to see the effect on the noise.
Next figures show the results for the In-Loop and Out-Of-Loop photodiodes respectively.





From the In-Loop PD graphs it is evident the improvement of the noise performances. The Out-Of-Loop PD noise is clearly limited by the VIGO pre-amp noise at low frequencies, a small improvement at higher frequencies is visible. It is now mandatory to have a new low noise pre-amp!!!

One thing to notice: I tried to decrease the lower roll off frequency below 100Hz. This has been possible only after decreasing the loop gain from 100 to 25. With a gain of 100 and a low loop frequency of 30Hz the loop becomes unstable. Probably this is due the huge peak at 90Hz that drives the instability.

Next steps are:

  1. extend noise measurements to higher frequencies with the current set-up
  2. introduce a notch at 90Hz and see if the loop is stable with a low roll off frequency below 100Hz and a gain of 100
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

RIN requirements for Virgo+

Post  Alessio on Wed Apr 08, 2009 8:01 pm

Using the Cnum calculated a long time ago (maybe overestimated), I calculated the RIN requirements for two different Virgo+ set up, in the case of 8W of CO2 power and a safety factor of 2.



As it can be seen from the figure, the requirements are rather stringent. I must say that especially in the case of "Virgo+ F=150", this is actually a lower limit, because to have a Finesse of 150 the ITMs must be replaced. Hopefully, the new TMs would have lower coating absorptions, thus much less CO2 power would be needed.


Last edited by Alessio on Wed Apr 08, 2009 10:25 pm; edited 1 time in total
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

Low frequency noise

Post  Alessio on Wed Apr 08, 2009 8:17 pm

Today, we tried to decrease the Low roll-off to 10Hz. We realized we do not have a notch filter, we must get it in Frascati. So to keep the lock of the loop, we lowered the gain. We were able to keep the lock up to a gain of 30. The figure shows the results.



In the meantime, Maurizio is working on a program to acquire the analog output voltage of the power meter, to keep an eye on the DC power stability of the laser.
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

Daily report (9.04.09)

Post  Alessio on Fri Apr 10, 2009 12:18 pm

Supposing that the instabilities of the loop arise because of the huge peak at 90Hz, I tried to notch the peak. The notch is working (the peak in the noise spectrum becomes huge, meaning that the loop is not gaining anything at that frequency), but it does not solve the instability problem.

Today I also tried to optimize the integrator's parameters to minimize the noise, at least in the in-loop PD, both at high and low frequencies. See the figure for details.



It seem I found a good set of parameters for low and high frequencies.

Than I tried to work a little bit on the New Pre-amp, with no success at all.


Last edited by Alessio on Fri Apr 10, 2009 4:56 pm; edited 1 time in total
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

New PDs and new setup

Post  Alessio on Fri Apr 10, 2009 4:53 pm

We are currently blind on the Out-Of-Loop PD: the VIGO pre has gone nuts and the new one never gave a sign of good functioning. So I had the greatest idea of the millennium: let's use a TEC PD as the Out-Of-Loop PD, since in this case the pre is built in the detector cage!!!

I measured the dark of the TEC PDs, reported in the following picture.



The noise performances are very good as expected, both systems are working properly. When turned on, it takes a few seconds for the controller to lock the temperature. Both systems have a very high responsivity (circa 4500 V/W) and since the pre saturates at 10 V, this means that the maximum power we can shine on these PDs is of the order of 2 mW. So I changed the set up in the following way.



When the waveplate is rotated to the maximum transmission, then the power on the PD is approximately 3.7 mW, but turning the plate it is possible to reduce it to 1.5-2mW.

Next week, we can align the TEC PD onto the beam and finally start looking at the Out-Of-Loop PD noise!
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

TEC PD

Post  Admin on Wed Apr 15, 2009 4:25 pm

Today i have controld the setup of the bench and in particular i have aligned the TEC PD onto the beam using one traslation stage and one tester. (The traslation stage used for the VIGO PD (12595, 12599....) can't be used because his post is too high). Now the power on TEC PD is around 1 mW. Tha channel used for the DAQ is 2. The power supply of the TEC PD is into the laser lab.

Admin
Admin

Posts : 4
Join date : 2008-11-05

View user profile http://virgotcs.7forum.biz

Back to top Go down

TEC PD 2

Post  mdipaolo on Wed Apr 15, 2009 8:25 pm

Thanks to Roberto, I have used one mechanical part for the alignment of the TEC PD.

mdipaolo

Posts : 106
Join date : 2008-11-05

View user profile

Back to top Go down

TEC PD and mistery noise

Post  Alessio on Thu Apr 16, 2009 10:52 pm

I added another lens before the TEC PD to reduce the beam size. This is because the TEC PD is also optically immersed and its size is 2x2mm2. The new set up is shown in the next figure.



Then, I measured the open loop noise of the laser with both PDs (12599, the In-Loop, and 5343, the Out-Of-Loop). It should be the same. On the contrary, as shown in the next plot, PD5343 seems to measure a higher noise at lower frequencies. The origin of this noise is not known. I changed the cables, I also changed the socket of the PD power supply, but nothing changes.



Anyway, I decided to close the loop and see what happens on PD5343. Next figure shows the comparison of the noise when the loop is open or closed.



At high frequencies, the decrease of the noise is evident. At lower frequencies, instead, there is no gain at all. But there is a hint which tells us that the loop is working: the heigth of the peak at 90Hz decreases by about a factor of 10.

We must investigate the origin of this extra noise.
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

Mistery noise

Post  mdipaolo on Fri Apr 17, 2009 10:52 am

In the lab there is one lense that doesn't focusing, i dont' know the why...but if you have used this lense can be that the mistery noise dependet of bad lense..... !?

mdipaolo

Posts : 106
Join date : 2008-11-05

View user profile

Back to top Go down

Mistery noise reduction

Post  Alessio on Fri Apr 17, 2009 3:24 pm

I tried to investigate a bit the source of the mistery noise. I switched off the air conditioning in the laser lab, but no improvement. So I decided to change again the set up, moving the 2 PDs close to each other. The new non-definitive set up is shown in the next figure.



In this configuration, to avoid saturation of the PD5343, I had to rotate tha wave plate to the minimum transmission, this means that power on PD12599 is only 19mW.

Anyway, as seen as in the next figure, it has been possible to reduce the mistery noise.



But since on PD12599 is arriving such a small power, the noise measured by this PD has increased, as shown in the next plot.



This means the set up must be improved again, to do for the next week. All these measurements have been performed without closing the loop.
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

New RIN requirements

Post  Alessio on Wed Apr 22, 2009 1:49 pm

I realized I used the wrong Virgo+ sensitivity curve to calculate the RIN requirements. I am posting a new plot, with a safety factor of 5 and TCS power equal to 8W. For comparison, I put in the plot also the RIN measured in open loop and closed loop as seen by the In-Loop PD.

avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

info

Post  mdipaolo on Wed May 13, 2009 8:37 pm

You do the comparison between the signal "in loop" with virgo request..is it correct ?

In the last graphs you have used the TEC PD and it is out loop.

mdipaolo

Posts : 106
Join date : 2008-11-05

View user profile

Back to top Go down

Comparison

Post  Alessio on Thu May 14, 2009 10:52 am

mdipaolo wrote:You do the comparison between the signal "in loop" with virgo request..is it correct ?

In the last graphs you have used the TEC PD and it is out loop.

I used the data from the In-Loop PD, because these where the only data available at that time. There are inconsistent information: we know P. Willems does not take into account the In-Loop PD data, but if you look at this presentation by Fred ClevĂ , it seems that the In-Loop PD is the only one considered.
avatar
Alessio

Posts : 127
Join date : 2008-11-07
Age : 46
Location : Rome

View user profile

Back to top Go down

Re: Closed loop operation

Post  mdipaolo on Thu May 14, 2009 3:38 pm

If we must consider the in loop then it is ok!

mdipaolo

Posts : 106
Join date : 2008-11-05

View user profile

Back to top Go down

Re: Closed loop operation

Post  Sponsored content


Sponsored content


Back to top Go down

Back to top

- Similar topics

 
Permissions in this forum:
You cannot reply to topics in this forum