[Dsg-hallb_magnets] Solenoid QDs

[Patrizia Rossi]

Thank you very much Brian for the detailed explanation.
So far so good. Taking data without any issue related to the solenoid.

-------------------------------------------
# Patrizia Rossi
# Jefferson Lab
# 12000 Jefferson Avenue
# Suite 5
# Newport News VA 23606 (USA)
# e-mail: rossi at jlab.org<mailto:rossi at jlab.org>
# phone: +1 (757) 269 7740
# fax: + 1 (757) 269 5800


From: Brian Eng <beng at jlab.org>
Date: Saturday, November 5, 2022 at 3:53 PM
To: dsg-hallb magnets <dsg-hallb_magnets at jlab.org>
Cc: Patrizia Rossi <rossi at jlab.org>
Subject: Solenoid QDs
First a quick bit of background ... There are 3 levels of quench detection (QDs) for the solenoid:

1. slow dump aka controlled ramp (current set to 0 A) done by the PLC
2. fast dump done by the PLC
3. fast dump done by the hardware QD boards

The thresholds on each of these is set such that 1 < 2 < 3. For the purposes of the recent events we'll ignore levels 2 and 3.

There are 21 voltage taps on the solenoid are these are combined in various ways to make up 10 comparators that are what the thresholds are evaluated against, all use the absolute value. See images VT.png for the physical locations and COMP.png for the math involved.

These voltage taps are sampled at 10kHz by the cRIO and every 200 ms the average of 2000 samples is sent to the PLC. These averages are what the PLC uses to combine into the actual values used to generate a slow dump. There is also an additional 1.25 second delay in the PLC before a trip is generated, which means the value would need to be bad for around 6 consecutive readings for the PLC to generate a fault.

I looked at the fast daq data from the cRIO (i.e. the raw 10 kHz data) at a three different times: full negative current, full positive current, and zero current, see attached Excel file mainly the Statistics sheet. Do note that these numbers were done by eye and should all be thought of as approximations.

What Probir and I had thought was a potential issue (VT17 being significantly noisier than VT3) probably isn't really one as that is always noisy in either polarity and the PLC is only using the mean. Really the only thing that seems so stand out is that COMP3 is extremely close to the previous trip point of 0.08 V but that really depends on the size of the window that the data is examined in.

Hopefully the increase in threshold earlier today will help. Maybe the lead/splice QD thresholds can be raised even more?

There is also a change so that we can monitor the status of the PLC comparators in EPICS that will be deployed the next time the magnets are ramped down to zero amps.

If anyone has any questions or comments let me know.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/dsg-hallb_magnets/attachments/20221105/ddfd9d46/attachment.html>