sPHENIX discussion of the superconducting solenoid

[Kin Yip <kinyip AT bnl.gov>]

Hi Paul Giannotti, John Haggerty and others,

1.  This morning, Paul came to 912 and looked at the strain gauges for more than an hour.  

The problem since yesterday (soon after the fast discharge) was that except the first strain gauge and the last few, the middle ones all seemed to show resistance ~0 <=> strain ~ 0.48xxx.
[  strain = (R - Rref)/Rref/g_factor  (ignoring temperature correction), where g_factor is ~2.05.   When R ~ 0, strain ~ 1/2.05 ~ 0.48xxx...  ]

To avoid dealing with the "short" etc., Paul talked about picking up "an offer" that Winston Pekrul had told us a few times before. 
That's, instead of using the current source and asking KeySight to measure the voltage, we can just directly measure the resistance on the strain gauge, thermo-sensors and linear-pot.

As an after-thought that I learnt later (as described below), we measured resistances close to ~360 Ohm and ~330 Ohms.  The latter were those due to the "short".

Winston quickly implemented it for us.


2.  Yesterday evening, Peter Oddo was nice enough to take a look at the strain gauge setup for some time.  He apparently thought about it.
I talked to him again after I said good-bye to Paul and came back to 911.    Peter pointed out two important things.  

2.1   Firstly, measuring resistance doesn't solve the problem completely.   Because the "shorted" (low impedance) path would screw up our measurement.   Instead of measuring resistance R for a strain gauge,  we are actually measuring that R in parallel with a bunch of 10+ resistance,
  R-prime =  R ||  (R1--R2--R3......in series)

So, ~360 Ohm in reality would become ~330 Ohm when trying to measure.  

To prove it, when we just opened/unplugged one jumper (which used to be the current path for everything in the entire series of strain gauges, thermo-sensors and linear-pot), the ~330 Ohm measured before suddenly became ~360 Ohm or so.

Just to repeat what I said in the Elog : "Peter has helped determine that there was a low impedance (~short) around 14 Ohms between channels 3 and 82 in the phenix connector block.   We opened jumper wires on 5 and 81 --- one is enough but this is just to show/remind us where the shorts are."

2.2  I also learnt from Peter, this disadvantage of measuring resistance without using a constant-current source is that by measuring resistance of the strain gauge, we unavoidably also measure the cable resistance (which 7 or 8 ohm each). 

But it's probably OK as we are mostly interested in the change during a ramp.

This effect showed up immediately on the large strain gauge values (as they are all very small anyway).  To make strain gauge values to be not so much bigger than before and slightly more compatible with the old ones, I've asked Winston to universally subtract 7.5  Ohms for all resistance measurements.


This is what we've had for the strain gauge, temperatures, linear-pot etc.   They're good for measuring the difference during a ramp as we only care about
their differences.


Kin