Reports 1-1 of 1 Clear search Modify search
takahiro.yamamoto - 15:38 Friday 18 March 2022 (20155) Print this report
Mystery of H1 coil gain of TM was solved

Tanaka (Aoyama Gakuin univ.), YamaT

Since O3GK, coil balance gain for H1 at PRM_TM is twice larger than other coils.
We found that this is the problem on the SCSI cable between DAC and AI chassis.
Finally, we can fix it by replacing the SCSI cable.
After coil balance check, COILOUTF_H1_GAIN should be changed form 2.000 to 1.000.

Coil balance gains for PRM_TM (K1:VIS-PRM_TM_COILOUTF_*_GAIN) had been set as follows since O3GK.
According to the original values, H1 gain is twice larger than other's

We wasn't able to believe this was coming from coil itself for a long time and Takahashi-san confirmed that this problem occurred at upstream of AI chassis output yesterday.
So we investigated around outputs of digital system today.

At first, we put the DAC output for each coil from COILOUTF_OFFSET as 3000ct and read voltage at AI chassis output pins as shown in Fig.1.
Then, measured voltage of H1 output was surely twice smaller than the expected values. (Detail values are shown in table below.)

For replacing cables, AI chassis, or DAC, cables around AI chassis was unplugged and then we found one pin of SCSI cable was broken (See Fig.2).
So we replaced SCSI cable and did same test.
Finally, output voltage of H1 coil became consistent with the expected value by replacing SCSI cable.

channel ct at OFFSET coil balance gain ct at OUT16 expected Volts measured (before) measured (after)
H1 (1-6pins) +3000 2.000 +6000 +1.831V +0.943V +1.825V
H2 (2-7pins) +3000 -0.994 -2982 -0.910V -0.921V -0.921V
H3 (3-8pins) +3000 0.729 +2187 +0.667V +0.651V +0.651V
H4 (4-9pins) +3000 -0.788 -2364 -0.721V -0.725V -0.724V
Images attached to this report
Comments to this report:
takafumi.ushiba - 18:30 Friday 18 March 2022 (20160) Print this report

I changed the H1 COILOUTF gain of PRM from 2 to 1.

Search Help