J. Kissel, D. Barker
In the spirit of LHO aLOG 9204, we've shaken H1 SUS ETMY using H1 HPI ETMY in order to test the sanity of thresholds set for both the IOP Software WD (IOP SWWD) and the newly implemented Hardware Watchdog (HWWD). These values are currently set at 110 [mV] output of the RMS circuit (which is identical both in the HWWD and IOP SWWD), which via this test, is determined to be roughly 10 +/- 5 [um] or [urad] peak-to-peak (as reported by ISI ST1 CPS, SUS M0 OSEMs, and SUS L3 OPLEVs). See attached overall time-series of the entire test, whose timeline is below. This mount of motion has been blessed by my spider sense of "Wow, that's a *lot* of motion; it's giving me the willies. We should cut off SEI excitation after 10-20 minutes of this." Note that, gauging by the motion of speed-dials, this amount of motion is still *less* than what occurs if, say, Betsy and Travis are mechanically aligning the suspension, so it's still a conservative threshold (because, of course, when at vacuum, and the chamber is fully isolated no such motion should ever occur).
Things we've learned:
- The IOP SWWD and HWWD now have their threshold set to 110 [mV], equivalent to ~10 [um] or ~10 [urad] of motion, pk to pk.
- The RMS output of both the IOP SWWD and HWWD track each other quite well -- as long as the signal is above ~30 [mV] or ~3 [um] or [urad] -- only the IOP SWWD readout is more the accurate below these amplitudes.
- Driving a 0.1-to-10-[Hz]-band-limited, white noise excitation at he maximum amplitude of HEPI's DACs does not trip the IOP or HW WDs. One must focus the power at a given frequency to really ring up the RMS.
- The calibration from counts to mV for the HWWD RMS OUT channel -- for now, randomly stuck in H1:IOP-ISC_EY_MADC0_EPICS_CH30 -- is roughly 0.1 [mV/ct].
Why the HWWD user interface is icky, but it'll still work:
- (As mentioned above) the RMS output of the HWWD does not report any change unless there is a significant amount of motion already present.
- The RMS output of the HWWD is choppy -- performing a sample and hold on the RMS value, and jumps around by 10s of [ct], or a few [mV] under constant excitations.
- The three HWWD analog trip flag channels -- for now, stuck in H1:IOP-ISC_EY_MADC0_EPICS_CH27 through 29 -- are railed when not tripped, and low when tripped.
- The after entering value in the requested RMS voltage trip point, H1:SUS-ETMY_HWWD_RMS_REQ, the readback, H1:SUS-ETMY_HWWD_RMS_RD only returns a value *close* to what's been requested. Similarly the two readbacks, H1:SUS-ETMY_HWWD_TIME_RD and H1:SUS-ETMY_HWWD_TTF_SEI for the time until trip report values only *close* to the requested value, H1:SUS-ETMY_HWWD_TIME_REQ.
There's still a significant amount of hardware that needs installing / replacing before we can hook the HWWD up to actually kill ISI coil drivers and HEPI valve drivers, but we will now propogate these thresholds and this infrastructure to the remaining QUAD suspensions over the next week or so. We will also perform a similar test on HAM chambers once they become available, to set those SUS' IOP SWWD thresholds (or at least confirm that they should be the same.)
Detailed Time Line
GPS Time Amplitude Notes
1087849965 100000 WN; Y, RX, RZ
1087850060 200000 WN; Y, RX, RZ
1087850199 250000 WN; Y, RX, RZ
1087850287 400000 WN; Y, RX, RZ
1087850326 HPI USER WD Trip
1087850348 start over
1087850484 300000 WN; Y, RX, RZ
1087850742 300000 WN RZ RX; 100000 ct 0.5 [Hz] Sine Y
1087850995 done
1087854054 Full bandwidth HEPI TF
where WN = white noise excitation, between 0.1 and 10 [Hz], and 0.5 [Hz] Sine = single frequency sine wave at 0.5 [Hz].
As in the original test, I turned OFF all control on the ISI and SUS, and turned OFF isolation control on HEPI. When ramping up the test, I began driving HEPI without turning off the ISI -- the ISI isolation loops survived through 20000 [ct] of HEPI excitation.
For the latest round of ISS array improvements, pico motors were added to the beamsplitter AROM4 (Fig 7 in T13000327), however it seemed to have slipped through the net which feedthrough the additional in-vacuum cables need to be attached to.
Consulting with Eddie, S at CIT he gave this initial recommendation based on what knowledge he had about available feedthroughs on the HAM2 chamber: this leave us with D1-3C2 and D3-F10 as the only viable options. We'll need 156" or a 180" cable to get from CB-6 to either D1 or D3,
With regards to the new Picomotor being installed in HAM2, the cabling is the following:
picomotor--->D1101515 (quad mighty mouse)--->CB6?--->D1101659 (seis resp cable)
However the D1101659's are only 108", and it didnt appear we had a longer cable on hand.
Thus I did a survey of the feedthroughs on HAM2 and I have marked up where there are unused ports in the attached pdf (hopefully you can understand my chicken scratch). I have not labeled which is D1, D2, etc.
There are a number of spare ones that it appears we could use, the most convenient (for a number of reasons) is the feedthrough in what I believe is the SW corner which has 2 unused ones. These are actually allocated for IO for the adaptive optics.
I requested to use the bottom one of these two unused ports (marked with a star). I have checked verbally with Guido and he was happy for us to use it, as is Hugh and Calum.
It was also suggested initially to swap out the CB6 "L" cable bracket from a 2-tier to a 3-tier. I however recommended adding an additional single cable high "L" cable bracket near the edge of the table on the western side near IM1. It is out of the way of any beams (and would be below beams anyway). This will help us use the short (108") cable we have available to go from feedthrough to this cable bracket and also trying to swap out CB6 with a 3 high cable bracket….though possible has the potential for me to bang/hit/damage something. My suggestion seems a simple and safe solution. Calum and I checked with Hugh about rebalancing and he doesn't see it as a big issue. Volker and I don't think it will be in the way of anything.
After the lenses were put into position and the AROM2 mirror placed back in position I completed the in-vacuum cabling from the feedthrough, to stage zero of the ISI, from stage zero to stage 1, to the new "L" cable bracket. Corey had already connected up the picomotor to the D1101515 cable (he told me he used the longest of the 4 cables which would be the 60" long one cable, "cable #4", connector "J5").
With the 3 spare unused cables I dressed them near the edge of the table so that the connectors were floating in free space. The location of the added "L" cable bracket and the routing/dressing of the in-vacuum cables can be seen in the photos
Now that we have the additional pico motor in chamber we now also need to think about how to hook it up/control it from the in-air side. I posed this Q to Rich A and here was his initial suggestion.
1. We can (if we are careful) use 28 AWG wire inside the vacuum system, which will free us up as far as finding the correct length (156 inches)
2. The only way we can use 28 AWG wire inside the vacuum system will be if we use less than ~50 feet of 22 AWG wire outside the vacuum system.
3. Assuming we use the one spare axis available on top of IOT2L (as provided by Picomotor Driver #3), we can stay within the 50 foot limit
4. A special cable will have to be made that goes between the RJ-9 style connector (see connector.jpg) on the one available output of the Picomotor driver and the 25 pin vacuum feedthrough.
5. The Picomotor driver front panel looks like this: (see picomotor.png)
As you can see, there are two 25 pin connectors that are normally used to drive each picomotor. In parallel with these 25 pin connectors are the individual motor connectors shown in a phone jack style connector. This is how we will pick up the one unused axis for HAM2 ISS steering.
6. The interface between the phone jack and a D-sub can be done by making a short transition cable to D-9
7. Now we just make a 9 pin to 25 pin cable. Easy.
What Rich will need is:
1. The length for the in-air cable from Picomotor controller number 3 to the proposed vacuum feedthrough
2. Account number
3. Permission and agreement from Systems and PSL
I have forwarded this request onto SYS and PSL representativies
Apologies in advance. This will contain quite a lot of detail and be quite verbose but want to get everything down whilst still fresh in my mind
We were tasked with installing new picomotors in the setup for the ISS array and also getting the mode matching correct for the ISS array.
Moving ISS Array
We were informed that the ISS array was initially installed in the incorrect location as the cookie cutter was used upside down. Thus the first thing we did was to put the cookie cutter back on (correct way up) and move the ISS array into what should be a more correct position. This moved the ISS array more towards the center of the table.
PicoMotor on mounts Swap
All that had to occur here was to swap out the fixed mount ROM RH4 (see Figure 7 of T13000327) that the beamsplitter optic site in for one that is controllable by picomotors (note AROM RH2..the last mirror before the ISS array) is already controlled by picomotors.
Before anything was swapped out, AROM RH2 was removed and the beam direction off of ROM RH4 was marked with irises (this is so that we could get the same alignment as currently there back. The base was kept in the same position and the current fixed mirror mount swapped out with one that is pico controlled. The same beam splitter optic was used. The mount was adjusted until the beam went back on the same alignment as before the swap. (One thing to note....and should be able to see on one of the pics in this log, is that the beamsplitter isnt located exactly as per its supposed location in D0901083 v12. Its about 1/2" further from ROM RH3 than design. This is no big deal other than when trying to pre plan on where clamps, etc can go). We tested that the beam from the beamsplitter heading past the ISS array has no clipping by the lid on the ISS array, and its not close to clipping.
See above in the comment section about the in-vacuum cabling of the picomotors. With the in-vacuum cable connected to the feedthrough, a temporary in-air setup was used to test that the picomotors worked. They did successfully.
Note: The Kapton "washers" have not been put in as we have only just got the Kapton sheet into clean and bake
ISS array mode matching
T1400176 gave various recommendations on how the mode matching to the ISS array could be performed. The decision made by SYS, et. al. was to use the two telescope lens solution. In particular Case 1 in Table 3, which is for the first lens after the beamsplitter to be a 2" lens of FL 343.6mm and the second lens to be a 1" lens of FL -171.9mm.
The initial position of where these lenses should go is indicated in D0901083 v12 Sheet 2. However with AROM RH2 still out (from above work) and before we put the lenses in, an iris was positioned in the beam path, and a beam scan on a rail (wiped down and wrapped in foil where appropriate) was bolted to the table (see pic beamscan) in an orientation so that the beam was centered on the beamscan as slid the beamscan back and forth along the rail. These will also act as our targets to check that we have the lenses positioned so that the beam goes through the center of the lenses.
The edge of the rail was positioned roughly in the same location as the optic in AROM RH2. This means that the focus should be roughly 16" (40cm) from the edge of this rail. So where do I get the 16" from. Well it is approximately 10" from AROM RH2 to the entrance of the ISS array, and then Ollie informed Volker that the beam travels 6" inside the ISS array (I dont know where he got the 6" from, but thats what we are going with). Also in Table 3 in T1400176 the waist is called out to be 304um, however Rick S informed Volker and I that we should be shooting for something more like 250um if we can.
The lenses were positioned roughly as per the positions indicated in D0901083 v12 Sheet 2 and it was very quickly seen that the focus position was not even close to the location wanted. After discussion with Guido M. Rodica M, it was decided that we could just slide the lens positions down stream a bit and shouldnt have to much affect. So we moved them both approximately 4" downstream (as indicated by the blue "A"s (seen in New Location for L1L2 pdf). We did notice that the beam onto the beam scan was hitting lower on the beamscan than without the lenses in but we made no attempt to fix that at the moment (more on that later).
Moving the lenses moved the waist position like we expected and so we had an initial go at taking some measurements with the beam scan. Below is the raw data:
Couple other notes:
As Volker and I were dressed up and in chamber, Rick S kindly volunteered to plot the data. The results can be seen in 1st scan horizontal.pdf and 1st scan vertical.pdf
As can be seen in these results, the waist is to small and still not in the correct location. However now that we have some results, using the jamMT program Rick had on the computer, we could use the results we have above and the lens positions to work backwards on what the incoming beam profile is actually like so that we could then alter the lens positions to get the beam size and focal position where we want. The results can be seen in jammt final solution.pdf (note the origin in the horizontal axis is where the edge of the rail is (ie the position roughly of the AROM RH2 mirror)).
The lenses were positioned in the positions as indicated by the jamMT solution and we did another beam scan measurement. Below is the results:
Couple other notes:
Again because Volker and I were in chamber, Rick S again plotted the results for us. These can be seen in 2nd scan horizontal.pdf and 2nd scan vertical.pdf
The results are pretty much bang on what we wanted. Great.
Now that the lenses positions are known we wanted to look at the beam being directed down vertically some (1-2mm) with the lenses in compared to not being in. This means that the beam is not going through the center of the lenses. However with the spacers we had available, we could not find a nice solution, short of kludging together something with washers. However Rick S suggested (and also wanted us to look at anyway) just how sensitive the telescope was if alter the picos on the beamsplitter ie do we need to tilt the mount a lot to get any change). Turns out we dont and its very sensitive so only a small amount of movement of the mirror mounts vertical position moved the beam back to the position we needed (sure this still means that the beam isnt going through the center of the lenses, but we now have the beam height as we want it). Again not a perfect solution but works with what have on hand.
We aso put in the back glass beam dumps. We had to position these by eye/guesstimation as we couldnt see the back reflected light. But we did our best to make sure they clip no beams on the table
Directing beam onto ISS array
Now that the mode matching is right, we can now direct it onto the ISS array. First off we need to swap out the curved optic that was initially in AROM RH2 for a flat mirror. The details of the mirror swapped in can be seen in optic used in AROM2 pic.
This optic had first contact painted on both sides of the mirror, but it also unfortunately in marker had the details of the optic written on the barrel (see pic writing on optic). Decided to try cleaning it off with fresh acetone and the swabs (see pic cleaning optic). The marker came off after a few swabs, and then we used 3-4 swabs on the barrel after we were confident it was clean just to be sure.
Before putting the optic into the mount the back first contact was removed (not using top gun (decision made in consultation with Calum)) and then the optic placed in the mount (see pic optic with first contact). Once mirror installed in position indicated in D0901083 v12 sheet 2 we removed the other layer of first contact.
Rick S gave us circular inserts that go into the aperture of the ISS array (I dont have a pic or a DCC number sorry) and by putting the mid size one into the aperture we were able to direct the beam (using AROM2 to adjust the beam direction and a handheld IR viewer to see whats going on) until it was centered on the hole in the insert. This should have the beam initially aligned to the ISS array as was our task :-)
We then put the SiCarbide baffle which goes in front of the ISS array (with the two holes) back into position. To enable the beam to go through the two holes we had to move the Brewster angled beam dump that is near the side of the ISS array and dumps a beam from HAM1 (I dont know its designation), closer to the edge of the table to allow the baffle we are trying to put in, in.
Everything was then dog clamped down and in-vacuum cables dressed
Pics:
Pic 7-13...are pics of the various components altered/installed so as the as built drawings can be updated
Pic 14 shows the one tiny bit of free space that is still on the table
Pic 15 shows roughly the "beams" view as going from IM1 to IM2 showing should be no clipping of this beam by black glass installed
For bookkeeping's sake, will go with the following labeling/naming of new/changed optics: