This work is related to klog#20639 and klog#20642.
And also, similar work was done on ITMY reported as klog#20658.
 

Abstract

- I re-installed original ADC board S1402718 as ADC0.
- Timing synchronization and F3 GAS glitches came back.
- Glitches and/or loud noise seem to be caused by the combination of ADC and AA not only ADC board.
 
 

Details

Because recovering timing synchronization is the first priority, I decided to replaced ADC0 for ITMX again.
I tried to 3 ADCs which are new product lots (rev C). But all of them cannot be synchronized in properly.
(ADC DT showed 20, which means 1-sample (~15us) delayed.)

So I used old product lot (rev B) which is not same as a original installed one.
There is no problem on timing synchronization if we use "rev B" ADC.

By using this ADC, F3 GAS channel (CH2 of ADC0 shown in Fig.1) became quiet.
But BF H1 channel (CH16 of ADC0 shown in Fig.2) became noisy.
So we need to find ADC whose all 32 channels are healthy...

All measurements including also noise of ADC1 and ADC2 are stored as /users/DGS/measurements/ADC/K1IX1/*.xml
After measuring all channels, I noticed all 32 channel on ADC1 seems to be healthy.
So I tried to use ADC1 as "ADC0".

But strange spectrum structure can be seen on some channels.
Fig.3 shows the one example. This figure shows the noise level of ADC1 which was used as "ADC0".
CH0 has strange peak around 30Hz.
ADC1 seems to be healthy if it's used as "ADC1". And also, before swap trial, CH0 on "ADC0" seems to be healthy (see again Fig.1).
So this means AA chassis, ADC adapter board, ADC cable etc. should be also health.
Though I used only healthy instruments, measured ADC noise didn't seem to be health.

It may be a combination problem of some instruments and cabling.
I wasn't able to decide which ADC should be used as "ADC0" at that time.
Today, I decided to re-install a original ADC board as "ADC0".