40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  PSL  Not logged in ELOG logo
Entry  Fri Nov 4 23:34:28 2016, awade, DailyProgress, Electronics Equipment, Trouble shooting in the north path 100_0258.JPG100_0261.JPG
    Reply  Tue Nov 8 14:01:11 2016, awade, DailyProgress, Electronics Equipment, Trouble shooting in the north path 2016-11-08_11.35.34.jpg
       Reply  Wed Nov 16 12:46:59 2016, awade, DailyProgress, Electronics Equipment, Trouble shooting in the north path and Murphy's law 
          Reply  Wed Dec 7 18:59:46 2016, awade, DailyProgress, Electronics Equipment, FSS monitor daughter board damage and Murphy's law D040424-B.pdf
             Reply  Wed Dec 7 21:14:07 2016, Koji, DailyProgress, Electronics Equipment, FSS monitor daughter board damage and Murphy's law 
                Reply  Thu Dec 8 17:37:32 2016, awade, DailyProgress, Electronics Equipment, FSS monitor daughter board fixed 
       Reply  Thu Jan 5 15:50:55 2017, awade, DailyProgress, Electronics Equipment, Trouble shooting in the north path 2017-01-05_15.17.55.jpg
Message ID: 1770     Entry time: Wed Nov 16 12:46:59 2016     In reply to: 1764     Reply to this: 1784
Author: awade 
Type: DailyProgress 
Category: Electronics Equipment 
Subject: Trouble shooting in the north path and Murphy's law 

I went back to these loop oscillation issues yesterday.  After excluding some possible I have tracked down the hypersensitivity of the offset to the fact that the 'boost' was engaged on the north path.  As detailed in PSL:893 and later in PSL:1140 a capacitor was added in series to a feedback resistor on the U7 op amp on the fast stage of the servo board DCC-D040105. This has the effect of adding effective integration in the common path, bumping up the low frequency gain. The reported choice of caps was different between the two 4.7µF and 6.8µF, I didn't check to verify that this is still the case. Turning off boost seems to fix these oscillation issues and the railing of the feedback signal.  This is because any small DC offset will be integrated until it rails.

Possible damage:

In opening the TTFSS boxes to see what was going on inside I may have caused some damage to the EOM monitor buffer. I pulled the TTFSS daughter board (DCC-D040424) off to have a look. It turns out that the plug of this board allows the pins to be shifted along one column.   When I put it back I did it wrong. When I powered back up the EOM buffer op amp got read hot, there is some visual damage to the outside (see picture) but its not clear if its affected its operation. Seems to be a classic case of Murphy's law: its possible to insert the plug into socket such that the +15 V pin goes into the -15 V pin and the -15 V pin can go into an opamp input and visa versa.  I've learned a lesson about plug and socket design choices. Blerg.

 

Lower left Op Amp (U7) has some burn marks on the outside now.

 

Correctly aligned
Possible to insert socket on column of pins to the left or right (incorrectly aligned).

 

 

Quote:

Now with a useable error signal for the north path, I've been trying to track down the source of ringing in the north FFS control loop. Initially I though it was too much gain, but it turns out to be related to offset.  With the current settings the fast monitor output is showing ringing at ~1 MHz.  It turns out that this is because the offset is set to low, as it rails it set of this oscillation (pictured below) that persists until I reset the lock. Vpp of this oscillation is 1.08 V (min=-1.78 V, max=-0.680V).  The upper rail (at 1.52 V) didn't have this oscillation.

I blocked the lasers, turned off the Marconi output and also added terminators to the test input on the RF stage (I don't know why this wasn't done before).  Watching the voltage at the fast monitor output I adjusted the north offset until the fast voltage was zeroed.  It was extremely sensitive at the zero crossing point. Even when I left the knob untouched it was flickering up and down.  I tried the same thing on the south path with identical gain settings.  The offset smoothly tuned from one rail to the other (only going into oscillations at the lower rail) and was stable. It didn't have any of the flickering level of the north path. To clarify the north path pretty much flipped from one rail to the other with very little adjustment, the south path was much less sensitive.  I'm not sure what is going on here.  Possibly a issue with grounding? Or something about the stages of gain has been altered in the north TTFFS box? Something is broken?

 

Maybe when I relock the loop the OLG will suppress this residual noise. But the sensitivity to offset knob tuning my mean that it will drift back to generating this 1 MHz oscillation, which is above the UGF of the loop.  I think this is something that needs to be solved before a sensible beat note can be detected.

--

 

In another unrelated problem I noticed that the ±24 V power supplies were flickering on the edge of current limit mode.  We are pulling almost 1 amp @ 24V to run all the electronics in the rack.  Seem surprisingly high.  I turned up the current limit, the +24 V line is now giving full voltage as it should.

 

ELOG V3.1.3-