Koki, Shin, Yuta, Kiwamu,
We had a close look at the ETMY glitch issue and found some features. The investigation is still ongoing.
Here is a summary.
[Coincidence with the local sensors i.e., LVDTs]
As shown in this figure, the optical lever frequently detected a number of kicks in the payload. They are the same type of glitches that Osamu reported the other day (4043). We looked at all the vertical gas LVDTs, but didn't find anything coincided with the optical lever kicks. Then we found that the BF LVDTs had the glitches which coincide with the optical lever kicks. So we determined that the kick was somehow associated with the BF LVDT and their damping system.
First of all, we believe that the glitches are not from the mechanical things based on the time scale of glitches. The glitches look like a lossy damped sinusoid with a frequency of something like 250 Hz. Usually they decay within a few cycle.
Importantly, we found that the glitches happened in an LVDT only when the LVDT had a static DAC offset (on the order of 1000 counts). And there are no crosstalks, meaning the glitch occurs only in the LVDT that has the offset and others don't show the glitches. This was true for all the LVDTs except for V2 which usually doesn't show glitches. Also, when offsets are applied to multiple LVDTs, they see the glitches simultaneously.
[Things to check or do]
- IOP DAC counts
- Raw voltage signals at around the coil drivers.
- Unplugging or shutting down the LVDT sensors.