TOA Alignment

[taylorjcolaizzi]

Developing an algorithm to determine the best TOA_VREF and 'TRIM_TOA' values for each half/channel that will result in aligning the TOA measurements for incoming signals. Currently developing a procedure for this at UVA; will post updates over time.

[taylorjcolaizzi]

TOA_VREF has a range of 10 b. When doing a pedestal run, I'm measuring non-zero TOA for only a small range of TOA_VREF. Specifically, I'm only measuring non-zero TOA for TOA_VREF values between 133 and 166. By comparing past pedestal runs with the TOA data, I think that the higher values of TOA_VREF correspond to a higher TOA trigger. This would mean that TOA is triggered to fire at a higher ADC if TOA_VREF is higher.

[andpet1324]

I made a function that sets the toa to a threshold above the pedestal. See #177 (comment). I'm hoping it can work as a base for a more full-fledged TOA calibration down the line.

[taylorjcolaizzi]

Great! I just took a look at your function set_toa and I like your method of calculating the efficiency. I'll try to incorporate a similar method in my analysis scripts for the scans in my draft pull request.

[taylorjcolaizzi]

Drafted a quick python notebook to look at the efficiency of the toa triggers called do_toa_vref_analysis.ipynb (will push a commit later). Specifically, after doing a TOA_VREF scan, you can calculate the efficiency of the toa as the number of triggers divided by the total number of events. You can also do this for each toa_vref, and for each channel! Here's a plot to show what I mean.

In this graph, I've done 1000 pedestals per integer value of toa_vref. The vertical axis is the number of non-zero toa triggers divided by the number of events (1000), per channel. Essentially, each channel is a different color.

It's overwhelmingly full of 0 values, but that's probably because one of my channels is dead and the fact that toa almost never fires on a pedestal.

We can use the efficiency to help figure out where to select the best toa_vref for each half of the HGCROC.

[andpet1324]

Very nice! Do you know if there's a pattern to which TOA_VREFs that aren't triggering (i.e. only have 0's)? While looking at TOT, I saw a similar occurence where there are intervals of TOT_VREFs that don't trigger at all (#160 (comment)). I removed the TOT_VREFs that didn't trigger and used a legend to see which ones remained. Would be interesting to see if TOA_VREF has the same pattern.

[taylorjcolaizzi]

This is such a great question, but I'm not sure exactly why so many values of TOA_VREF don't trigger at all. My gut instinct is saying that the ROC automatically rejects triggering TOA off of incoming signals that don't have the correct "shape" to them, but this is purely me speculating. I'm not sure if this is part of the zero suppression or the shaper circuits, both of which I don't yet understand, or if this is due to something else entirely.

I'm going to look into making a plot that removes the values that don't trigger and make a legend for the ones that remained!

[taylorjcolaizzi]

We can calculate and plot the efficiency of TOA as a function of calib. This sample graph shows the fraction of times toa triggers (out of 100 events) for a given calib value. It's uncertain if I calibrated the pedestals correctly, but this is a good start toward visualizing the turnaround points for TOA for each channel.

[taylorjcolaizzi]

Merged pull request into main. Thanks, everybody!