[Halld-controls] Update for FCAL CAN System talk

[Daniel Bennett]

Hi Everyone,

I received some feedback from IU on some of the issues that were brought 
up in yesterday's control meeting.

The first issue was the instability of the ADC Cathode reading, how base 
responses could vary in the hundreds of volts some percentage of the 
time.  Paul tells me that this instability is due to the switching noise 
of the capacitors in the Cockcroft Walton chain, and is not a physical 
effect on the PMT Cathode.  So this is something that I will just need 
to work around.  I have seen this noise on the scope when monitoring the 
test pulser, but the capacitors can apparently affect ADC measurements 
as well.

The other issue was setting a Trip Current on the firmware.  The ADC 
Current only monitors Base current, while the PMT will only pull in the 
microAmp range.  I had shown a distribution of ADC currents over a long 
period of time for many bases, and there was a large peak from 2 to 4 
mA, and a long tail extending out to 15mA, which the last 15mA bin 
peaking due to the ADC maxing out.  Paul says that this tail is likely 
caused by extra current drawn by the ADC to take the actual ADC Current 
measurement, and is also unavoidable.

I'll try to focus on solutions to both of these issues on the Firmware 
side.  One possibility for ADC Cathode would be to take several 
measurements on the firmware during an ADC Cathode request, and if the 
RMS is over a certain amount, throw that measurement out and re-measure. 
  The switching noise is in the nanosecond range, so multiple 
measurements would be much slower than that scale.  This would only 
allow real effects on the ADC Cathode to show up back in Epics, but you 
would risk a base not responding indefinitely if the base had an error 
which caused every set of ADC Cathode measurements to produce an RMS 
which was not below that threshold I just mentioned.

For ADC Current, this same solution would not be as ideal, since the 
effect is a consistent high-side tail in the distribution.  I will still 
look into averaging these Firmware side, but Paul had suggested that 
using the Power Supply current monitor, which will average the current 
drawn by 100 bases, would be a better indicator for Base current draw. 
  Either way, the Bases are individually fused, and the entire PSU is 
fused, so I do not know how critical a fast threshold current would be.

I'll keep working on all of this, and keep everyone updated.

-Dan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/halld-controls/attachments/20120921/27285a38/attachment.html