Entered lab about Mon Nov 30 10:21:35 2020, after taking a COVID test through Caltech's new surveillance testing program.
I'll pick up where Shruti left off on the beat note. The comb of sidebands becomes a single line remains a comb when the PID is off; Koji suggests maybe the (PLL) PID is oscillating at 10Mhz.
exit Mon Nov 30 16:12:04 2020
Thanks, the photos are now on the shared drive.
For storing lab photos in W Bridge, you can use our shared google acct instead so that we all have access to it (see chat for secrets)
Isn't the PID oscilating at 10MHz?
Attachment 1: Video of spectrum analyzer with zoomed out beat after turning off the PID loop of west laser
Attachment 2: Another image of the zoomed in spectrum when the PID is on.
Anchal dropped off the Moku from CTN, along with its USBA->USBC cable, power cable, and ipad.
Entered lab around Tue Nov 24 13:24:57 2020 to finish photographing Zach's cantilevers.
some things about cameras, and in particular the FinePix F300 EXR
- Found a suitable power cable M-M for the New Focus 0901 power supply on the east table (I did not realize yesterday that these were the same cables). Then I checked the voltage on the pins and they were fine.
- Using the New Focus 1611 (1 GHz PD) powered by the New Focus 0901 +-15 V / 0.3 A max. power supply, I tried finding the beat note. I looked at the RF output on a HP 8560 E spectrum analyzer and the DC output on an oscilloscope.
The DC output ranged from 500 mV to over 1 V as I scanned the temperature of one or both lasers.
- When the east laser temperature read roughly 8.34 kOhm and west was 9.04 kOhm I saw a pattern as in Attachment 2.
Changing the temperature slightly did cause the peaks to shift about, and further when I changed the polarization of the east laser using the HWP the height of the peaks varied. They also disappeared when either of the beams were blocked.
The estimated peak power in the taller peaks is ~0.1 µW from the plot.
- I also tried scanning the temperature of both lasers again to possibly find a single peak. No luck yet.
Today, I didn't check the alignment very carefully and I probably have to tune it further after the changes that Aaron and I made over the past few days.
The next step is to do the phase-locking.
Entered lab Fri Nov 20 19:24:32 2020, usual sanitation.
[blue "Photodiode Power Supply] Looking for a DB9 to BNC adapter. I found this spider instead -- close enough. Use multimeter to measure 24V between pints 4 an 9... not promising. Confirm power is connected, no signal on the frontpanel BNCs either. Could remove this one and take a look on the benchtop, but above is...
[Newfocus +-15V current-limited power supply] Has 3 pole bananas and a power switch on the front. Found a power cable for the back. There's a bananas to 3-pin LEMO already there. Double check the voltage with a multimeter. Alas the connector doesn't fit the PD, but should be some cables in the EE shop or elsewhere...
Didn't find the right connector in EE. On the 'power cables' rack (NE corner Cryo), there was a M-F connector, but I need M-M. Could cannibalize the 12VDC supply? I think for now +-12V is working, so should look a bit more.
I moved the power cables for our preamps for better strain relief (attachment 2 is the before photo).
I also had left this ND filter sitting on the table (attachment 1). Yikes!
More photos here.
ExitFri Nov 20 20:39:43 2020
Entered lab Thu Nov 19 16:33:57 2020. Usual sanitation, personal reminder to report campus access with Caltech.
we want to coat some of Zach's cantilevers with a-Si so we can make a cold Q measurement. I've started to take some photos, but have become tired and will finish tomorrow. There are O(5) suitable cantilevers produced in January 2018, but I'll have to dig a bit more (or ask Zach) to determine what's what. We can measure the Q of the most promising few cantilevers to be sure they're acceptable.
I borrowed the digital camera from EE shop, but left its case (which is very dusty).
Sending W path beam to a Newfocus 1811 to measure free running laser intensity noise.
Following Shruti's recent diagram, I moved the Newfocus 1811 into position after OMTL1. I also moved PO1.1 back into the beam path, so I can use it to align into the 1811. Turn on the E laser and TEC, also had to move Ma for alignment. I still don't have a +-15V power supply, will ask around. Turn off the laser and TEC before exit at Thu Nov 19 20:41:22 2020
Attachment 1: An updated version of the diagram in elog 2577 where the path lengths to the beat beam-splitter are identical. The fiber launchers and some components have been moved around, but everything after PO1.1 along the beam has been retained as before.
Attachment 2: Retaining the same configuration to the beat BS, the cavity with Mach-Zehnder interferometer has been added. Also the path lengths to the MZ input BS along both laser beam paths have the same length. Except for the ring cavity, the Mach-Zehnder is also balanced.
Attachment 3: Updates pertaining to the current setup
This is the data using the Data Ray Beam'R2 profiler with the InGaAs window. Attachment 1 contains images of each of those profiles.
D: distance from fiber launcher in inches; The two values in each of the cells are [Clip 13.5%, 4 sigma] respectively, i.e., the method used to calculate the beam widths.
The previous measurement using a razor blade refers to 'sigma' which I believe explains why these values are 4 times larger.
These profiles were taken with temperature stabilized such that the powers were ~1 mW.
East laser set to 8.070 k Ohm, West laser set to 9.065 k Ohm. I don't understand why there is such a difference.
I had hooked up the ITC 502 combi controller to the west Rio laser and used only its temperature controller. (I believe both the thermistors that measure the diode temperatures are TH-20k Ohm.)
Both the PID controllers work satisfactorily: the TED 200 C with the east laser stabilizes to within few Ohms of the setpoint thermistor resistance within some seconds;
the ITC 502 stabilizes at a similar rate but at an offset of ~10 Ohms despite the integrator being set to maximum. I fiddled around with the P and I settings a little but realized that this configuration seemed optimal.
To measure these profiles at different distances I moved the fiber launcher head and then replaced it back to its original position, roughly.
I entered the lab somewhat before Tue Nov 17 14:56:22 2020. Exited Tue Nov 17 16:39:12 2020
Hand sanitizer on entry, also sanitized the bulky green laser goggles before and after my use (forgot contacts). Turned off the laser, sat at desk and considered turning on the laser. Took a break. On my walk I wrote this haiku
I've placed the following items outside the cryo lab:
1. Cryo liquid N2 dewar
4. Two pairs of cryo gloves
Optical layout: beam launch -> lambda/2 -> steering mirrors -> lens 1 -> ND 0.6 -> lens 2 -> PD 1611
There is only a 12 VDC power supply compatible with the 1611 power port, but the PD requires +- 15V. Surely there's one somewhere. Perhaps this is why I observe only -6 V on the DC mon with 1mW input power at 1550nm (checked against the Thorlabs S122C; I expected -10V). Maybe the beam is too large.
- Although when we talked about adjusting the MZ-phase, we decided that having the phase/path length control with fiber components might be better initially (Refer Attachment 2), for now I began doing everything in free-space.
- Attachment 1 shows the setup as it is now. Previously I'd placed polarizing beam splitters instead of 90/10 beam-splitters because I thought it would be easier to work with, but now changed my mind and decided to stick with what we planned.
(Once the beat is obtained on the spectrum analyzer)
- Measure the laser frequency noise
29 Oct 20:
I've added Attachment 3 -- which is the current free space version and some PLL electronics.
- It does not show the Mach-Zehnder part as that will be added only later
- This setup is asymmetric but in a future version we will change that
Today I modified the optical setup with the aim of obtaining the beat between the two diode lasers for phase-locking.
I added pick-off polarizing beamsplitters with HWPs in each path for now (to be able to adjust their power) and mixed them at a 50/50 non-polarizing beam-splitter to eventually reach a Newfocus 1811 low noise PD.
I will add pictures and more details later.
I was in the Cryo lab between 1215-1230 this afternoon. I removed two resonant RFPDs from what was Johannes' setup (encircled in Attachment #1). I also brought a SR554 preamplifier to the 40m.
I was wearing the usual PPE (gloves, face mask) while I was in the lab.
This is the spectrum coming off of the sample. there should be a peak at 1038Hz... but there isn't. And what is even weirder is that the spectrum analyzer that is built into Simulink shows a peak where I expect but when I do it here it doesn't show up.
Update: I think I have found why there is a discrepancy between the two versions of the power spectrum. The spectrum analyzer in the Simulink model requires non-continuous data so you have to use a block to make your data discrete. The sampling rate of that block affects where the peak of the mode is seen. So it seems that that the mode seen in the previous post was just caused by making the data non-continuous.
I got the script to run the simulation and added color to the diagram just to make it pretty :0
The next step is quantifying the error of the loop. My plan for this was to just calculate the Q from figure 3 of the moderinger paper. Then I can see if that value is consistent. Aaron suggested settling time and phase margin of the loop. So that is the next step (once I figure out how to do that)
Also I added the code to the Qryo github
Beautiful! Want to push this to the repo under git large file storage?
The step function represents the excitation of the sample. Ideally, it would excite to the setpoint and stay there but for some reason, it is jumping way past the point before returning. By messing with the gain and the frequency of the low pass filter I could get a variety of results the best is shown below at 50 gain and 40 rad/s.
The overshoot is interesting! To understand the loop shaping, I suggest checking out Gardner's Phaselock Techniques or Astrom and Murray's Feedback Systems. They both have sections on optimal PID controller design (at least Astrom and Murray do). You can make a pole-zero plot to help choose the location of poles and zeros in your loop shaping filter (the discrete zero-pole TF we added, after the gain).
Working with Aaron's suggestions (In the previous post) we got the mode ringer to converge. Previously the loop would continue to excite the sample to infinity but by fixing the following things we were able to get the step function to converge.
The power spectrum shown is taken from the spectrum analyzer shown in the loop. It shows what I would expect with a peak around our mode frequency of 1038 Hz.
Tue Sep 22 21:33:30 2020
I'm cleaning a bit, and gathering items not in use or in need of repair. They would make less mess in my office.
took an inventory of optics cleaning supplies, first aid kit, general cleaning supplies, wipes, etc. I found most were included in the first round inventory, but I took photos this time to convince my future self of object permanence. Will add to the wiki and update in a bit.
gloved up, shoe covers, went to QIL to check out the sprinklers and CTN to grab a GHz spectrum analyzer (HP8560E).
out: Wed Sep 23 00:20:03 2020
I suggest just using what's used for th QIL table.
Yesterday into today, I've been shopping for laminar flow HEPA fan filter modules for the PSOMA optical enclosure. I didn't find a lot of LIGO documentation listing specific filters, but here's what I've found online with some downselection on 'low vibration / high filter quality'. Please let me know If there's a company we often use, or if you can help direct this search at all.
Another consideration is flow rate relative to our volume. I can do this calculation, but what is the particle density (eg at 0.3 um) we want to achieve inside the enclosure? I realize that I never got the particle counter recording despite it being on my list, so I'll try to do that remotely today. We need to know the ambient particle count and the clean volume in the enclosure.
*all prices listed here are from publicly available pages
I updated the stage 1/2 optical layout to be more detailed after getting a sense of the sizes of things again last week. Even though this isn't how the table is currently set up, it might be good to accommodate future vacuum chambers in our earlier designs to minimize how much we need to move and realign optics.
I will update the PSOMA hardware inventory tomorrow to reflect the additional details in the new drawing. The updated diagram is available on git LFS, and the hardware inventory now reflects the diagram up through stage 1.
Notice of lab entry: 20 Sep 2020 evening
Fiber modulators on the table :
1. Intensity modulators (BW: up to 12 GHz) MXAN-LN-10
2. EOM phase modulators (BW: up to 150 MHz) MPX-LN-01
Dimensions of vacuum cans mentioned in attachments.
Here's the layout.
Some easy things that should be changed:
I measured the transmission of the Coastline 1m mirror at 180. ppm (S122C).
Alignment procedure while setting location of optics:
Alignment procedure subsequently:
Not sure how, but none of the drawers of the blue optomechanics cabinet are opening. I don't have a key. Here's what happened
Found someone who's had this problem before, might give it a try...
This worked, I used the metal meter stick to unlock the drawer.
--> note that link formatting breaks link for me, so here it is - https://www.practicalmachinist.com/vb/general/help-my-lista-locked-me-out-how-do-i-open-201606/
--> wrote up a similar experience with additional detail ENG_Labs/260
Can anyone tell me the specs / history of some of the custom optics in cryo? I'm mounting the 1m Coastline mirror and will start with that in the PSOMA cavity.
I took some photos of the existing layout. I'll just take apart the E beam path, and leave the W path unchanged for now as reference.
I moved the E fiber output coupler closer to the edge of the table, to make this path easier to reach.
Hopped around on the laser hysterisis curve for a minute. To optimize the temperature,
I started a cryo lab inventory that is separate from the PSOMA hardware inventory, and intended for stock items in the lab (optics, electronics, clamps, general safety and cleaning supplies, etc). It will be a work in progress. Both are accessible to anyone logged in to google drive with their ligo.org credentials.
Both are also linked on the PSOMA project wiki.
Did some mode matching, see the git.
Date of entry: 24 Aug 2020
The cameras were unfortunately lost in the mail, but we can use my laptop or other camera. Ended up leaving to do a couple comsol things that needed completing today.
Date of entry: 18 Aug 2020
The inventory is in Clickup, which is a new organizational tool I'm trying out. There' an easy csv export, so I can get it elsewhere if/when we want. We have a wide variety of lenses:
I found fewer curved mirrors, but there were a couple.
Steps I took for the temperature sensor:
1. Tried to see what's the temperature by reading the current temperature in ndscope or dataviewer. In dataviewer, go to 'Signal' tab, and enter the channel name or find it on the list of slow channels. For ndscope...
No data appears on either. I restart cymac1, which seemed frozen, but still nothing.
2. What's going on with these channels?
Looks like the channel is reading zero.
3. I traced the cable from the particle counter and found that it sends data to cominaux, the common auxiliary machine for the lab.
This is the database file that defines the channels on cominaux. I search for 'LAB_TEMP_F' and find the epics record for the temperature channels. The epics records are all "calc" records, and the temperature in Kelvin is taken from X1:AUX-ACROXT_AI_15. This corresponds to channel 15 of the acromag slow ADC.
That's starting to make sense, the cable from the particle counter didn't go to the acromag ADC. Starting from the ADC channel 15, I traced the cable back to what used to be the AD590 temperature transducer.
4. Where did the IC temperature sensor go? Searching the elog and my dusty memory... neither readily recalls where it went. Let's get another one, they are cheap and easy to use.
Purpose: Inventory lens supply, identify some combination of optics that will let us mode match to our ring cavity. Picked up some books on silicon from the library.
Measurements around table
Could probably move the table a few inches from the wall and make use of the space between the lights for the enclosure. There also isn't much room in the back corner in the NS direction, and we may want to shift in either direction. Orientation seems as good as it could be. The ceiling above PSOMA is lower than above cryo cavs.
Will mark up photos and post.
Entered the lab at 11:00 am. The lab is far too hot (78F) and humid (45%)!
Inventoried available photodiodes in the cryo lab, on the PSOMA wiki.
Pretty tired honestly, I submitted an order for a few things we want in the lab, such as:
The PSOMA optics table is from TMC vibration control (TMC 784-29701-01). I sent them an inquiry for the datasheet, and also asked whether they sell clean enclosures or have application notes that might help us put one together. Also talked with Arian Jadbabaie a bit about the Hutzler lab enclosures, and got some photos.
Exit ~430 pm
Purpose: Measure optical table dimensions, start to assess what we need for an enclosure. Set up EOM/EOAM to take some transfer functions. Set up realtime model. Figure out why particle counter isn't logging to frames. Turn on the HEPA for cryo cavs table.
Notice of lab entry: I will be at the cryo lab for a few hours today (04 Aug 20) afternoon
At a few locations along the beam I measured the vertical and horizontal beam radii by measuring the power at different positions of blade edge across the beam. The power measured at the photo-diode was fitted with where is the beam radius at the location along the beam and was either the horizontal or vertical position of the blade.
The standard deviation as estimated from the fits are lower than 0.002 mm for all estimated radii.
But with the crude nature of the setup and not having the laser temperature stabilized (resulting in power drifts while taking readings), I guess that the error in each measurement is higher. The data in Attachment 3 also has recorded the resistance of the temperature sensor at different points while taking the data.
- Attachment 1 has the updated setup with the clamped translation stage
- Attachment 2 shows the measured points with the error function fits. The offsets between the different curves along the x-axis are arbitrary.
- Attachment 3 shows a linear fit and an estimated divergence angle. I've assumed that all points I measured are outside the Rayleigh range, i.e., away the beam waist. The caption '5H' means that this set of data was taken 5 inches away from the output coupler (fibre-to-free space) moving the blade in the horizontal direction.
- Attachment 4 has all the data and jupyter notebook with analysis
- Attachment 5 shows the beam spot on the view card
Notice of lab entry: I will be at the cryo this (03 Aug 20) afternoon
- On entering the lab I noticed that the work with the light fixtures was completed. Unfortunately I could not avert it or get it all covered in time; I had assumed that I would be contacted beforehand but was not. But, by inspecting the table I do not think it looked any dustier than before. For any such activity in the future after we’ve cleaned the optics, I will remember to get it covered beforehand.
- The particle counter did not seem to be saving any data so I’m unsure what the effect of this activity was. The 1 micron particle count:
When I entered (1 pm): 80
(6 pm) : 70
This change might probably be entirely attributed to the air scrubber.
- When Aaron and I chatted with Calum today about COVID safety, Calum pointed out that turning on any HEPA filters in the lab would reduce the time between single-person occupancy to almost 0 for our room, as it would serve as an additional air scrubber. I tried locating a switch on the cryo cavity table for the HEPA fans but could not locate it today.
- After reading through bits of the laser operating manual and turning on the east cavity laser, I turned on the TEC (thermo-electric cooler) box [TED 200 C] with neither the TEC nor the servo on. I used this just to read off the resistance values of the sensor. With the current drive off it was 11.2 k-ohms [22 C].
- I waited until the sensor readings stabilized to changing at the level of a few ohms around 8.5 k-ohms [28 C] in a minute and began measuring the beam profile. (See Attachment 1 for setup)
I could not mount properly the only translation stage I found since the holes on it were smaller than 8-32, and moreover, I could not locate anything suitable that fits into the 1/4" holes on the table with this.
- Nonetheless, when I took some data and plotted it, it looked like an error function so I took more data. Will post plots later after fitting and analyzing. I may have to repeat this again.
- The sensor for the power measurement was S122C (Germanium) with a range of 700-1800 nm and 40 mW. After setting the power meter to 1550 nm, the measured power of the entire cross-section was ~2.3 mW.
- Also in the image in Attachment 1, the two optics (HWP and mirror) near the fiber-free space coupler, were previously in the path of the laser and now moved to either side of the beam. No other optics were moved from the previous set-up.
Notice of lab entry: I will be at the cryo lab today (31 Jul 20) 9 am - 2 pm.
Purpose: Look around for components [power meter charger, power connectors for PDs, flashlights, etc], laser operating manual; find out specifications of available mirrors, RF PDs
Also, there seems to be only one laser temperature controller in the lab. The west cavity laser TEC port is not hooked up to anything. For the time being this is okay since we'd be using only one laser.
Notice of lab entry: I will be at the cryo lab today (30 Jul 20) 9am-5pm.
Purpose: Beam profile the Rio planex laser
- Turned on the laser that goes into the east cavity on. Turned on the Tenma supply, then the laser current driver D1500207 labeled'E'. Saw a tiny, but bright, green spot on the detector card.
- Located razor blade and translation stage, for the beam profile measurement but not sure which power meter/ photodiode to use for 1550 nm. I plan to move the optics on the path of the east laser (to the east cavity) in order to do this measurement and later set up the initial ring cavity
- Aaron suggested I do another airflow measurement within the lab. Without the scrubber, there seems to be no measurable flow around either the Cryo-Q table or the cantilever table. I moved the scrubber into the lab near the workspace at the north side of the lab and let it run on medium speed. A few cm from the scrubber it the airflow rate is almost 5000 CFM, but it becomes <40 CFM even as close as the Cryo-Q table and again becomes negligible near the cantilever table.
- I set up the translation stage, moved relevant optics, but will continue to measurement later after I've researched and located a 1550 nm power meter
Two of the four overhead lights over the west (cantilever) table don't work anymore. Three other overhead lights are very dim/out.
Liz dropped off an air scrubber (Medify Airx MA-40) and an anemometer (Digi-sense 20250-15) today. I'm using the instructions on the DCC to measure the air flow and assess the occupancy limits of the cryo lab. I calculate an acceptable amount of time between lab uses, and for two people to occupy the cryo lab simultaneously. For these calculations, I use a conservative threshold of P<1% for the acceptable probability that a second person becomes infected, given one infected lab occupant who sheds virus at 10 nL/min. I measured the dimensions of the lab at about 30x20x10' (l, w, h), for a 6000 ft^3 volume. I expect that's a high estimate, as it doesn't account for things like the awkward geometry of the staircase, volume of lab equipment, or stagnant air inside cabinets.
These results are consistent with what I see for similar types of rooms in the LIGO spreadsheet.
Under an air flow model assuming perfect distribution of air from HVAC and HEPA scrubber throughout the lab, and very conservative requirements for probability of spreading COVID assuming a single lab user is infected,
Furthermore, the air in the lab Is not perfectly distributed. The air intake is several feet from the outflow, both near the door. The primary heat sources are the electronics racks along the N and NW sides of the room. The gradient from 71.1 F at the thermostat to 78 F at the cryocavs rack is uncomfortable to work in and bad for the electronics and optics. It's also too stagnant for the viral load conditions assumed in Evans P2000189 to apply.
We are not currently cleared for 2+ people to use the lab. This analysis suggests to me that before doing so, we should improve the air flow conditions in the lab. And, even under optimal conditions we may minimize the total time with multiple people in the lab simultaneously.
I disconnected the oscillator, PDH boxes, and laser drivers from the power strip, and powered the strip with the Tenma supply. I had the ground and negative pins switched at first, careful of this -- ground is black, negative is green on this cable.
I plugged in the electronics, but only turned on DC power to the laser driver. I just left them on long enough to confirm a bright green (on viewcard) beam spot for both lasers -- we're in business!
Turned off the lasers, disinfected the common surfaces and objects.
PS, the anti-fog wipes work wonders. My goggles went from fogging in seconds to no fog at all.
I located the materials for stage 1 PSOMA on the West optics table. I recorded what we have in the hardware inventory, and what we don't have is flagged for purchase. I start by cleaning up the electronics rack, removing anything I think is not in use.