Air core conduction test (Jan 2014)

Talk from Ben

Magnet tests 05/21/2014

Setup

• Single hybrid used with FEB
• L1-3 half-module 16 is used for all tests
• FEB is oriented horizontally w.r.t. to field (i.e. 90deg offset compared to real running)
• Flange board will be approximately along x=y=0 line from the magnet center
• Flange board has two orientation in the fringe field 
  • 2: y-z plane
    
  • 1: x-y plane
     
  • Note that fringe field is non-uniform which makes it hard to emulate completely 

Run plan

Run statistics: 100Hz, 10k events
Config: MagnetTest.xml 
Data located at: tracker: /u1/data/hps_magnet_test_05212014

Magnetic field

Peak expected field is 1.5T for FEBs in HPS.

Fringe field shown below. The signal flange boards are located at approximately z=-417mm, y=0 and |x|=475mm with our nominal target position at (0,0,0). 

The expected fringe field at that position is not known but at x=y=0 the field is less than 5% of full field which is a good upper limit. We could scan to about 0.1*1.5T=0.15T=1500Gauss.

Summary of tests 05/21/2014 (Pelle)

Hi all,

I thought I’d write a summary of today’s testing of the FEB and flange board in the magnetic field. All our notes and some more info can be found on:

https://confluence.slac.stanford.edu/display/hpsg/SVT+DAQ+Magnetic+Field+Testing

We took pedestal data with a test run module, 20V bias, and in parallel ran high-speed (4+ Gbit) data taking. 

* We scanned the signal flange board from 0 to 1200Gauss (0.12T) which is about 8% of the full field (1.5T) and also larger (by at least x2) then we expect in real running.

-> No effect on noise or high-speed signal transmission. 

-> We conclude that the signal flange board is ok, even though it has some fairly magnetic inductors on it. 

* We placed the FEB horizontally in the magnet and scanned the field from 0 up to 1.48T a few times up and down. 

-> The system was a little flaky: it would start up in a state where AVDD and V125 are stable (<10mA or so) but then get into another state where these currents fluctuated by 100mA or so. The power monitoring data is not yet in the data stream (hopefully next week) so the way this was observed was by reading the register status manually from the GUI 1-2 times per second. 

->  There are 1 channel  (maybe a few, hard to tell from online display) at each extreme of the hybrid (based on online display) with noise that is between 100 and 200ADC. We are not sure what that is. In the state with stable currents the current of those edge channels are above 200ADC and in the fluctuating step they are around 100ADC.

-> The noise level across the hybrid was stable between 55-65 across most chips across most runs. Switching the field on/off didn’t change the noise RMS beyond that based on the online display. See next point.

-> We need Omar to look at these files quickly and look at the details comparing these runs (sample to sample, particular channels) because the online display auto-scaled things which means it’s sometimes hard to see things to within a few ADC’s of noise. Nevertheless there are a couple of observations:

General

—> As the system go into the fluctuating state we can see that apv0 has in general a few ADC higher noise. It also has a more prominent shape across the chip than the others. 

—> When the currents were fluctuating the edge channel noise was ~100ADC and when they were stable they were about 200ADC. The noise level across the chip was very similar as far as we could tell between the states but we need more detail here. It looks like the shape of apv0 is not in the runs where the currents were stable.

—> The fluctuating currents started once sometime between when we moved the board into the magnet and then ramped to 0.3T. Since we also saw bit errors from the high speed link which was stable w.r.t. to field during this time we think this happened during moving the system (loose connector?). The other time we saw the fluctuating currents start was during ramping the magnet from 0 to 1.5T. 

—> We were able to get the system into the non-fluctuating case once when the field was 0T. We tried when the field was 1.5T and we couldn’t.

—> Based on this we think at this point we can’t conclude that the field induces this change in state. Ben has also observed fluctuating currents in his office. 

—> We also think it’s important to figure out the edge channels (header being interpreted as data somehow?)

B-field effects:

—> First time we went from 0.6T to 1.0T the noise jumped up by about 5-8ADC across the hybrid. 

—> Then when going from 1.0T to 1.5T the noise dropped back down with almost the same amount. Strange.

—> When we ramped down to 0.6T the noise dropped by about 4ADC (hard to tell)

—> Then ramping down to 0T then noise increased by about 4ADC and up to 8ADC for apv0.

—> Ramping to 1.5T after that lowered the noise back to the level observed previously. 

—> During the last run with field at 1.5T the magnet tripped at about 90% of the run. We should look carefully if there is a transition in the data Omar?

Summary

Signal flange board is ok.

FEB is not conclusive. We need detailed plots from Omar before we conclude anything but based on the online display I personally think that the changes seen are from the DAQ having issues and not from the B-field. But I want to see the the detailed plots before I say for sure. We also scheduled a day to come back next week. Ben would try to exercise the DAQ a little (understand edge channels, add power monitoring to data stream for correlations, noisy channel really noisy or a daq issue) to have a better idea what is normal and what is not. If we can get the detailed plots soon we can maybe be in a better position for the next test to know what to look for. 

Sorry for the long email. Since I’m away a few days I thought I would summarize things. Ben can answer questions about the test (probably better than I).

/Pelle