Zjs log

--Zjsimmons (talk) 18:33, 14 July 2014 (CDT)--Zjsimmons (talk) 18:33, 14 July 2014 (CDT) Made some changes today, see qualitative improvement in green lock. So we've been having issues with the locking performance, the transmitted IR and also generated green were intensity modulated at about 300Hz, a strange freq to see problems as it was not the cavity piezo resonance freq of about 420Hz. The SHG derived green looked at times almost like a pulse train as a result of the modulation. This was tracked to in part a suspected ground loop between driver and locking electronics. Unfortunately the fast feedback was needed to get good output power, but would also cause the modulation.

• Swapped the cavity piezo for the new shorter travel (6um) one from Noliac. Glued everything together with epoxy so as to cut down on load weight. Somehow, realigning the cavity didn't turn into a huge project. Note: when you see some green, put the spots on top of each other and you're most of the way toward good alignment.
• Performed the resonance test on the piezo/mirror system. We now see resonance at about 700Hz. This is higher but not dramatically so. Also, the resonance was qualitatively not as noticeable, so perhaps it is less strong as well which would be cool.

Upon locking with the new piezo, observed very robust stable lock, you can really rap on the table and it won't drop. Piezo/mirror system definitely less susceptible to noise, banging, disturbing the lock. However we still see the modulation present via the fast feedback. Turning the FFB on and tuning to the best gain the output looks pretty good. Would estimate fluctuation in about the top 25% of the transmitted IR or so. Turning the piezo gain all the way up saw no piezo resonance with FFB off. Green is relatively constant, i.e. not intermittent, although there is fluctuation in the top 1/2 or 1/3 so.

• Swapped R30 in the circuit from 1k to 100k, this should boost the piezo feedback by a factor of 100. Since we saw no resonance at max piezo gain with the new piezo, i decided to change the resistor to give us more gain to play with. Now at higher piezo gain (adjusted via pot) we can observe piezo resonance. FFB squelches this but we still see some wiggling on the transmitted IR and Green. Now the wiggles are more like 100-150Hz. There also seems to be a tad more fluctuation in general at low piezo gain now. I split the difference and change the 100k to 30k.
• Next priority: find some way to squelch the FFB noise problem. Looking online, and from Deniz's comments, it seems that the laser driver mode input is very susceptible to noise/ground loop issues.

--Zjsimmons (talk) 21:54, 11 July 2014 (CDT) Week in review: Frustratingly slow week; didn't seem to make much headway/i don't know anything. Re: Cowan Code: After trying about a bunch of different routes to get it working, i think i've more or less decided on the original prescription, the Alex Kramida pre-compiled version for PC. Despite the big incentive to pursue the un-compiled versions so as to be able to use the scripts, i haven't been able to get it to work easily. Here's what doesn't work, there's basically 2 routes that i worked on for a while(too long):

• Compiling the code in Ubuntu didn't work as I couldn't get the right compiler. Both compilers mentioned in the documentation are a bit of an issue. Unfortunately the Intel Fortran compiler doesn't work with the new version of Ubuntu. It is possible that it would work with an older version of Ubuntu or a different Linux distribution. g77 is not readily available/very old so also a problem. The gfortran compiler also doesn't work with the code, at least according to the documentation.
• Compiling the program in windows via cygwin also didn't seem to work either. The Fortran compiler (ifort) although available for free for windows is not available for windows. Besides, you can't run the make files in cygwin and have it utilize the windows ifort anyway. However, you can have it use the gfortran, the open source fortran compiler. Although it seemed to compile, the code didn't work. According to the documentation, apparently that compiler won't work.

Re: Laser linewidth issue: We finally made a little progress on this problem. So it appears there is a problem with the fast feedback, some sort of ground issue. Deniz showed pretty dramatically that just touching the BNC ground to the laser feedback terminal on the driver totally messed up the error signal. As much as i would like the electronics to not be responsible for problems, it appears that they are. I kind of figured.

• If we tune up the gains, piezo gain basically low as workable, fast feedback not too low but not too high(where it looks best when ramping), we get a reasonably stable lock. If we turn the fast feedback gain too low we see wiggles at the piezo resonance freq in the transmission. If the fast feedback gain is too high we see slow oscillations at around 300Hz, evidently the noise on the laser that was responsible for the more or less pulsed behavior of our green light.
• The Fabry-Perot apparently led us astray. Either we were looking at the peaks while the fast feedback was connected (i could swear it wasn't) or the cavity linewidth is not as good as we thought. I thought that the linewidth changed slightly when we tuned up the fast feedback by doing the threshold test, indicating not being limited by the cavity linewidth, but maybe that was wishful thinking and we were actually limited by the cavity linewidth.
• Then again, that's why we wanted to look with the interferometer, which worked very well. With the fast feedback disconnected, we observed a lindewidth of about .5MHz, very reasonable. However, with the fast feedback connected, even if it's off, the linewidth is more like 5-10MHz. Even just touching the ground of the feedback connection BNC connector ground totally disturbs it.
• One piece of information that doesn't jive is the frequency of the noise oscillations, ~300Hz? it's not a line freq or harmonic. One idea: perhaps there is some sort of feedback or cross-talk between the two feedback branches. What if the piezo feedback is modulating the cavity transmission a bit, i.e. gain at the SHG piezo resonance freq is not sufficiently low, there's some wiggle. If this modulation is being fed back some to the laser, maybe that's responsible to the increase in lindewidth. I don't know if something along those lines could be the mechanism, but if it were, swapping the piezo could really help. That's our next idea priority.

--Zjsimmons (talk) 16:45, 7 July 2014 (CDT)

Trying to nail down why the locked laser is misbehaving. The symptoms:

• The lock is stable, but the transmitted IR laser power is not constant but modulated about about 200Hz with a modulation depth of about 50%. The The error signal has a saw-tooth characteristic. This causes significant problems to the output green power as the SHG basically exaggerates the problem and we see pulsed green output instead of continuous green. We have observed the problem go away but we don't know what caused it to disappear.
• Additionally, when scanning the DC offset, the shape of the transmitted resonance changes it noticeably broadens depending on the DC offset. This may or may not be related to the modulation problem.
• When zooming in on a peak with the DC offset and ramp size, as you get very close, the peak looks crappy, there is a lot of structure present which at least part of is at the piezo resonance frequency.

• Tried replacing the lock-box ramp with an external ramp, the appearance of the peak when you zoom in is perhaps slightly cleaner, but about the same.
• Tried turning the waveplate on the cube that separates the polarizations for the two photodiode channels and then adjusting the error signal with the lock-box. This did not seem to help or change anything.
• Perhaps it's a cavity alignment issue? Could going to transmission mode help?

--Zjsimmons (talk) 20:01, 3 July 2014 (CDT) Touched some science this week!

• Spent a lot of time messing with the Cowan code. A couple issues to work out but i think it is workable, i'll write it up eventually, once i get it figured out. Here's the short story: There is a version of the code by Alex Kramida that is compiled for windows, it includes the 4 core programs: RCN, RCN2, RCG, RCE. This runs from DOS but the output is not very easy to interpret. There are scripts that automate running the above programs and parse the output into something a little more useful. However these only work with Unix/Linux, although there is a workaround for in windows using a sort of Linux terminal emulator called Cygwin. This should work although i haven't been able to get it to work so far. Alternatively you can compile and install the programs in Linux, but i haven't been able to get that to work either. The third option is to rather than use the pre-compiled versions, to compile the programs via Cygwin in Windows. I haven't been able to get this to work yet, but i haven't spent much time trying either. In any case, one of these options should work and allow the use of the scripts.
• The science: we put some green light through the Eu:YSO crystal at 5K. We're currently using the set-up with the tapered amplifier rather than the 1055 fiber laser as we're hunting around for the transitions and don't want to disrupt the fiber laser seed. We see some evidence of the transitions, not so much absorption but fluorescence; it looks pretty bright orange.

• Interestingly the orange appears at only narrow green wavelength, but we get a lot more fluorescence when the crystal is warm
• Transitions appear to be right where they are supposed to be from the literature. We expect it at 527.5381nm at see very bright fluorescence at 1055.0775/2=527.53875nm. Also see some orange in the vicinity of the adjacent site's peak at 527.6426nm.
• Wavemeter and spectrum analyzer give different peaks. 1055.0818 on the wavemeter is 1055.004 on the optical spectrum analyzer. OSA seems to be off by about 0.08nm.
• If you pull out the transfer line when the cryostat is cold, it warms up pretty slowly, might be able to use to take measurements as fn of temp
• There is also evidence of absorption but it's not very good at this point. The green transmission is more like a a pulse train than a constant output, this needs to be fixed/addressed. SHG seems to make variations in cavity transmission much worse; i suppose it sort of squares them. Hopefully we can get the output more constant.

--Zjsimmons (talk) 13:14, 23 June 2014 (CDT) Holy cow can't believe it's been a couple weeks since i've updated my log. A few things have happened:

• DAMOP. Great having it in Madison, got to see some old friends including Herman Batelaan and Jeff Sherman whom i hadn't seen since Seattle. Sort of surprising to me that i basically saw no other work using doped crystals, however made a good contact with Morgan Hedges. He worked under Matthew Sellers on some rare-earth doped crystal stuff. Been in touch with them and currently looking at some of their papers. I think this could be some very good folks to know int he field, folks who have some experience with these crystals that we're just starting to experiment with.
• received the lockbox PCB from advanced circuits and wired it up, this took much of last week. seems to be working ok, although we're planning to change the piezo and box the SHG cavity to hopefully get more stable performance.
• received the 1055 fiber amp from Nufern for the green laser, started setting it up.
• just today i finished my documentation for the circuit and for building circuit boards, will post on the wiki. Building circuit boards is under Procedures. The lock-box documentation is under Equipment List, perhaps documentation should have it's own page, but this will work fine for now.

--Zjsimmons (talk) 21:33, 30 May 2014 (CDT) Week in review. Was out of town early part of the week for Noah and Teresa's wedding. Did get a couple things done this week:

• ordered PCB for locking circuit, finally. Should be here in a week or so. The whole project was a good subject to learn about. I plan to document the method in a document and post to the wiki.
• Nick and I updated our poster for DAMOP
• Started work on the Eu:YSO crystal holder for the cryostat. Will finish machining next week.
• Plugged in and turned on the Menlo Orange laser. Measured power of 110mW. According to email with Menlo, Chris McRaven, he measured 129mW. I don't know why there is a discrepancy. Could be they saw lower power upon adjustment of pulse length.

• Note: last fri i finally submitted my paper review of the Optics Express paper on an Atomic Bragg Reflector. Really glad to have gotten that done. It really ate up more time than it should have but was a good experience to have done a paper review. Learned I would like to work on/understand the Stark effect better.

--Zjsimmons (talk) 16:11, 20 May 2014 (CDT)

Trying to get fast feedback to laser diode in place to see if it helps with cavity locking. So far it hasn't worked/i haven't been able to get good slow feedback locking. Difficult to untangle the different knobs we have control over. Issues/brainstorming to consider:

1. Does polarization instability screw us?
2. Can we trade a little bit of power (second waveplate, polarization incident on cavity) for better error signal performance?
3. Tune up cavity peaks (IR) with crystal temp de-tuned and then zero in on correct temp? My suspicion is that adjacent peaks from different orders are messing up the error signal which is why it maybe isn't working well. If we tune up the IR, maybe this would be better to look at than trying to look at the green.

--Zjsimmons (talk) 18:00, 19 May 2014 (CDT) We observed some decent locking of the SHG cavity today. At highest, we saw intermittent power of about 9mW w/ about 65mW of 1055 incident, which is a pretty decent conversion ratio of ~14 or so. Observations/issues

• Piezo is oscillating which is lowering output power. When the error signal gain is turned down, we see higher intermittent power but the lock is not very stable and drops. With higher gain, the oscillation is very pronounced.
• Also put in place the fast feedback circuit to the laser driver, does not seem to help, in fact turning up the gain kills the lock. There could be many reasons for this. I am wondering if the phase is not correct. I would think it would be the same as the slow feedback branch and two inverting amplifiers shouldn't mess it up, but i would like to check this. Perhaps there is an issue with using long BNC and/or i would think that the bread-boarded circuit would be more susceptible to crappy performance at fast feedback frequencies. I do want to get this straightened out soon so i can order the board and get the permanent circuit built.
• Also, it's becoming increasingly clear we should use a different piezo, something with higher resonant frequency would be very nice/easier to work with. I don't think we need the travel that the current piezo stack provides anyway.

--Zjsimmons (talk) 18:45, 16 May 2014 (CDT) Not much going on in the lab the last few days as i've been working on a report for my microscopy class. Tried looking at improving the resolution of HELM/SIM using a couple schemes. It appears they probably will not work, but it was interesting and i learned some stuff. Perhaps there's something there that could be developed further.

• Locking circuit PCB layout is more or less ready to go. Want to test the circuit some more and make sure it works with the crystal in the cavity before getting it made. Also want to test the fast feedback. Nick has got the crystal in the cavity and we can see some single-pass green but it is definitely going to be more challenging to lock it. There's a lot more structure on the error signal and the cavity modes seem to jump around quite a bit.

--Zjsimmons (talk) 14:14, 13 May 2014 (CDT) Have seen a fragile lock:) Despite Multisim showing gain should be -20 dB at 400 Hz, it still seems prone to 420Hz (resonance) oscillation. Issues:

• Switch to put bypass resistor to kill integrator when not locked seems important. Otherwise integrator will chug away and rail at some point. I'm also thinking having a low noise op amp for that stage should help prevent gain on op amp error for example. -added the switch.
• Should we use a different piezo with a higher resonance frequency?
• Piezo DC offset should only be positive- I changed this.
• Lock seems ok. Also added a 100nF cap on the output/input to the sum with DC offset as a passive low pass filter. I'm sure this could be improved but i think we should try it for now and see what happens.

--Zjsimmons (talk) 22:51, 12 May 2014 (CDT) Spent today trying different locking gain stage configurations, tried adding an additional stage to increase gain. Discovered error signal was not being properly added to DC offset, this seems to be corrected but two problems have cropped up:

1. Integral gain has a relatively large DC offset, i guess this is not to be unexpected, but sort of fights the goal of zeroing in on the resonance point by adjusting the DC offset of the ramp and then turning on the gain, i.e. killing the ramp and adding the gained error signal to ramp DC offset. This can be suppressed but comes at the expense of gain.
2. Piezo is resonating. Despite lowering the gain and lowering the roll-off point to very low frequency, the piezo is now oscillating. Can this be eliminated? Perhaps we need a lighter mirror set-up to raise the mechanical resonance frequency? It may be a good idea to get more piezos anyway, in case of failure.

--Zjsimmons (talk) 18:44, 9 May 2014 (CDT)

• Cannot seem to get the SHG cavity tuned up. Very frustrating, stuff in between my peaks is only getting worse, not smaller. Polarization from fiber is also drifting, annoying.

• Trying to see locking. Not working so far.

• Saw some strange behavior with TA, output power seems to have decreased somewhat since we first used it. With 5mW seed power:

Current (mA)	ASE (mW)	output (mW)
1000	16.6	120
1501	126	377
2001	274	632

Compared to before (4-14-2014), where:

Current (mA)	ASE (mW)	output (mW)
1005	18.9	165
1501	165	430
2002	-	838
2074	434	1W

This seems to be down a fair amount from initial output and spec, degradation? are we doing something wrong?

• Shout-out to Jared, need to have power meter at 1064 not 780 when looking at 1064:P

• Note: weird power meter behavior. Old power meter head seems to behave a little different than the newer one. The old one gives about a 20% higher reading between 1.06um and 1.05um. The newer one gives about the same reading for both wavelengths, something is fishy with that. Perhaps this explains some of the discrepancy between the TA performance now and before, if we measured before at 1.06 and now at 1.05, still a bit different. We should keep an eye on this.

--Zjsimmons (talk) 18:25, 8 May 2014 (CDT)

• examined Josh's lockbox. It seems to short the positive rail momentarily when first connected to power and then recover. Very strange behavior. Occurs w/ ICs removed as well, so it doesn't seem to be the chips. Checked voltage ref pots, doesn't seem to be for example shorted diodes. curious.

• Having trouble tuning up the SHG cavity, could not get very nice clean modes, always some garbage in between peaks. Subsequently spun the fiber launch and totally misaligned the whole shebang. I guess it will be good practice tuning it up. Still unable to get as clean of 'troughs' between peaks, may have to pick Nick's brain about it.

--Zjsimmons (talk) 17:30, 7 May 2014 (CDT)

• Tried repairing the fault finder today. Turns out there are some optics inside, two elements in small brass rings, one attached to the laser. Looks like they were glued together at one time and had since broken loose. This was why we could not get light through fibers with the fault finder, even though it was on, it was not correctly launching into fibers. I aligned the two elements as best i could using the transmission through a multi-mode fiber and glued it back together. It sort of works, coupling to single mode fiber is not that great, but with a little fussing(rotating the fault finder and pulling the fiber out slightly) you can get a decent amount of light through a single mode fiber. We should probably get a new fault finder as I don't know how reliable this repair will be.

• re: mysterious polarization noise, could it be we are blocking the beam with our fingers/fiber and that's why the power goes to zero?

• Nick showed me how to tune up the SHG cavity today. It didn't seem too bad to mess with; it'll be good to not be afraid to adjust it.

--Zjsimmons (talk) 23:23, 6 May 2014 (CDT)

• Figured out that the flat polished fibers were the cause of power fluctuations after the TA in 1055nm set-up. Using an angle polished fiber fixed the issue.
• Power fluctuations are also not the reason the circuit is not locking the SHG cavity. Perhaps the gain roll-off needs to be at lower freq? as we are seeing some wiggles around 420Hz (piezo resonance freq). Maybe this is involved in failure to lock. We could also try adding fast feedback branch and see what happens.

4-6-2014 How about a log? Maybe this would work better that the paper lab notebook. Lab notebook seems ill-suited to many functions and i find myself not looking forward to using it, perhaps this will work better. Should be a more natural medium for computer-generated material. It's nice that it would be searchable as well. It could also host PDFs and reports and other documents. I kind of wish we would have started a wiki a while ago! -z