Reports 1-1 of 1 Clear search Modify search
MIF (General)
satoru.takano - 23:20 Monday 14 October 2019 (11014) Print this report
ALS DARM loop problem was solved
[N/T akano]
 
We continued REFL PD issue (10958, 10959, 10987).
 
Finally we found that a Dsub cable was problematic. Now the DARM signal looks as quiet even on resonance as off resonance.
 
Here are what we did today. 
 
Condition0
  • We restored the complete setup (with all cables connected properly)
  • We kept DARM locked and the laser injected to REFL PDA1 (a PD at issue)
  • We remove cables one by one and checked the feedback signal
As a result, nothing changed the situation...
 
After that we checked the connection btw REFL PDA1 and REFL table, then found that
  • REFL PDA1 was still connected to REFL table
  • At that time REFL QPDA2 (only this case is connected the table) was totally isolated from asc0 rack
  • Connection situation was changed. The sound from a multimeter was a kind of pulse (beep beep beep ...). If QPDA2 is connected, the sound is continuous.
Also we noticed that REFL table connected to the vacuum chamber, IMC table, and asc0 rack. it might be that all tables and racks are probably connected via the chamber(?)
 
Next, we tested independent power supplies.
 
Condition1
  • The interface was on the rack and turned off
  • The power supply was TEXIO PW18-AD
  • The socket we use was SC-12 (on the wall)
    • DARM error signal was quiet even IR was on resonance. Good.
Condition2
  • The socket we use was SC-3 (in the clean booth)
  • Other conditions were the same
    • The power supply made a strange noise (pssss) and stopped.
    • After that the supply didn't output ±18V...
    • We checked the interface board work properly.
Condition3
  • We tried to connect the power cable from another rack (ioo1)
  • The socket : S-14 of ioo1 rack
  • The interface was still on the rack and turned off
    • We connected the cable to the interface and... guess what?
    • Spark!
    • After that we checked that it seemed not to happen anything wrong.
Condition4
  • We brought another power supply PAN 35-10A from SRM rack
  • The interface was the same situation
    • DC power was not supplied properly
    • GND and -18V might be connected (sometimes connected, sometimes not)
Condition5
  • We changed the interface board for REFL PDs
  • Dsub cables were also replaced
  • PD cables were connected to another interface on ioo rack
    • The DARM signal looked quiet even IR was on resonance
Condition5
  • We changed the interface again. In this case we used the one on asc0 again.
  • We swapped the Dsub cables which connect PDs and the interface.
  • Channel assignment was the same as the condition 1-4. (ch1: PDA1, ch2: PDA2)
    • The DARM signal got noisy again
Condition6
  • The laser injected to PDA1 was blocked by a beam damper.
  • The other conditions were the same as before.
    • Nothing changed
Condition7
  • in this case the laser injected to PDA2 was blocked.
  • The other conditions were the same as before.
    • The DARM signal got quiet!
From these investigations, we concluded that the source of this problem is a Dsub cable that connects PDA1 and the interface. We should have replaced this cable to a new one, but due to lack of time and our energy we temporary used an interface on ioo rack instead.
 
By the way...
 
Ultimately this huge GND loop (table <-> chamber)  may cause another problem. We should consider the way how to avoid or solve this loop.
Comments to this report:
shinji.miyoki - 7:02 Tuesday 15 October 2019 (11015) Print this report
Quite tough works to be respected.

In order to control and understand the ground condition, we need to collect the ground connection and AC power supplying method information of individual DGS system, VIS, VAC, MIF, IOO, AOS, CRY-suspension, PEM, CAL and interlock.

Among of them, is there any systems that cannot disconnect the ground of DGS system with vacuum cambers ? VIS and CRY ?, because of their structure.
While, all components on the optical tables can be disconnected with them by using proper spacers.

One of most annoying selection is whether we should disconnect the ground of instruments with AC line outlets in the booths and so on, using 2p-3p converters.
From the point of view of safety, this is not allowed, however, this connection definitely make huge ground because the AC ground line of the corner and both stations are connected with each other with some amount of resistance that is generated by the long ground cables.

Note that this is based on the "single signal transfer", not on "differential signal transfer" that is realized in KAGRA.
Search Help
×

Warning

×