Virgo Logbook

AdV-DAQ (Data Acquisition and Global Control)

rolland, pacaud, letendre, masserot, genin - 13:05 Wednesday 21 February 2018 (40648)

Replace EEroom DaqBox used for EIB-SAS control and PSDs (dbox 35 replaced by 54)

The DaqBox 35 used for the EIB-SAS PSDs in the EEroom (2 ADC mezzanines and one DAC1955 mezzanine) was not working since some weeks, with flat output signals.

I have replaced it by the DaqBox 54 this morning (all mezzanines were also replaced).

Connecting the ADC signals one by one, it appeared that there must be an issue with the cabling of the PSD or cable called PSD_B_laserlab. The corresponding ADC channels output was completely flat when connected alone, and all the 8 channels of the ADC mezzanine became completely flat in some cases when also other channels were connected in parallel with these ones. --> the cabling must be checked.

SInce about 12h30 LT, I have reconnected 8 ADC channels that look fine and let the 4 faulty signals not connected. I have also reconnected the 8 DAC channels (EIB-SAS and phase camera channels).

The DaqBox 35 will be brought back to LAPP for checks.

Images attached to this report

Comments to this report:

nocera - 16:28 Monday 26 February 2018 (40679)

B. Montanari, F. Nocera

Today we had a chance to work a bit on the small bench PSDs which acquisition was affected by some issues lately since the ADC replacement (40648).

We tested both the sensors' front-end electronics installed close to the small bench in the laser lab and the cabling going from there to the ADCs in EERoom.
Unfortunately, the data of interest are stored only as trend and therefore our test sinewaves (~1Hz) are not visible. No plot is therefore attached.

It turns out that there is a clear cross coupling between channels associated with the ADC36 (DAQ Box 52) which it is used to acquire PSD1, PSD2, and PSD3 (x and y): even when only one sensor is connected to the front-end electronics and only its asymmetries are acquired (everything else is left disconnected at both ends) the signal associated with them shows up in all other channels.
This does not happen with signals associated with PSD4, whose asymmetries are acquired with a different ADC (#70).

All tests done seem to rule out the problem was either in the front-end electronics or in the cabling and pointed towards a poor behavior of the ADC #36.

To make sure, after lunch we injected a test signal (2 Hz sinewave, 1 Vp) first in ADC36 ch0 and then in ADC70 ch0.
We got the exact same results we had got in the morning.
The cross-coupling is around -60 dB (rough estimate, we did not go through the pain of a systematic check in all possible cases) for ADC36 while is not measurable for ADC70.

Conclusion: the ADC36 must be checked.

Letendre, Genin, Rolland, Pillant, Masserot - 19:39 Wednesday 28 February 2018 (40699)

We found that the problem came from input signals that exceed ADC mezzanine input voltage. Standard mezzanines accept +/-10V differential.
This was solved by changing these mezzanines with +/-20V ADC mezzanines (means -20V < VINp-VINm < +20V).

Now the DaqBox 54 contains ADC mezzanine 34 and 09.

Letendre, Genin, Rolland, Pillant, Masserot, Paoletti - 19:58 Wednesday 28 February 2018 (40700)