[Vetroc_daq_sbs] Simple occupancy test results

[Robert Michaels]

This is pretty good work, and it seems consistent with my expectations that you start getting deadtime in the few-hundred kHz region, and things break down at 1 MHz.  For presentations, it may be good to generate a few plots, like maybe the deadtime as a function of rate, and perhaps a TDC plot showing the resolution and how it changes between low rate and high rate (a self-timing peak or something).  Basically, how does the deadtime vary with rate and how does the performance of the VETROC vary, for the parameters you are trying to measure.  

Bob

----- Original Message -----
From: "R. Evan McClellan" <randallm at jlab.org>
To: "Vetroc_daq_sbs" <vetroc_daq_sbs at jlab.org>
Sent: Friday, January 13, 2017 5:03:23 PM
Subject: [Vetroc_daq_sbs] Simple occupancy test results

Hello All,

Today, Carlos and I finished a simple occupancy rate test of the VETROC TDC capabilities.

Setup:
-Use GRINCH prototype to send real signals to ~64 VETROC signal input channels
-use a pulser to fire PMTs with an LED
-use the same pulser signal to directly feed one NINO input (we then watch to see if any hits on this channel aren't recorded)
-use a separate pulser to generate signals for the VETROC reference channel and trigger the TI
    -this pulser was at 10kHz for all tests

We varied the rate on the PMTs, increasing it until the "test channel" started missing hits. There were no missed hits until the rate got above 250 kHz. Above 250 kHz, the VETROC started missing ~50% of the hits (and 50% of events!?!). When the rate got to 1 MHz, the format of the raw data was corrupted, so we stopped the test there.

We then verified that this rate limitation is due to occupancy by turning off the PMTs. Indeed, the data and test channel look perfect when the PMTs are off, even at 1 MHz.

Conclusion:
The VETROC cannot handle a *periodic* rate above 250 kHz, simultaneously, on 64 input channels. Presumably this limit would be reduced to ~125 kHz if using all 128 input channels.

Caveat:
I understand that we expect the individual PMT rates to be ~1MHz. With an independent 1 MHz rate on each PMT, the occupancy will be much lower than in this test. (Assuming reasonable internal time windows...)

Next Steps:
1- Get the prototype setup ready for cosmic detection, then use a source to induce a truly random background on the PMTs.
2- I will discuss with Ben/Bob and try to learn the details of the VETROC firmware.

Cheers,
Evan 



R. Evan McClellan, PhD
Hall A Postdoctoral Fellow
Jefferson Lab
_______________________________________________
Vetroc_daq_sbs mailing list
Vetroc_daq_sbs at jlab.org
https://mailman.jlab.org/mailman/listinfo/vetroc_daq_sbs