sPHENIX discussion of electronics

[eric mannel <mannel AT bnl.gov>]

Craig, Nathan-

I spent a few minutes looking more closely at the inspection data provided by Hamamatsu.  Here are the captions for the gain measurements in each of the 7 deliveries to date.  If you look closely , for deliveries 1, 2, 6 and 7, the quoted bias voltage (Viased) is in the middle of the range that they claimed to use.  However, for deliveries 3, 4 and 5, the quoted bias voltage is 1V lower with delivery 3 being a slightly off.  Not sure how this impacts our studies.

Cut and pasted from the inspection sheets posted in DocDB (https://docdb.sphenix.bnl.gov/cgi-bin/private/ShowDocument?docid=159).

Delivery 1:

For gain sampling, we used devices whose Vop range is 69.42V to 69.46V. Viased voltage was 69.44V (average of Vop range).

Delivery 2:

For gain sampling, we used devices whose Vop range is 69.32V to 69.36V.
Viased voltage was 69.34V (average of Vop range).

Delivery 3:

For gain sampling, we used devices whose Vop range is 69.32V to 69.36V. Viased voltage was 68.38V (average of Vop range).

Delivery 4:

For gain sampling, we used devices whose Vop range is 69.31V to 69.35V. Viased voltage was 68.33V (average of Vop range).

Delivery 5:

For gain sampling, we used devices whose Vop range is 69.46V to 69.60V. Viased voltage was 68.48V (average of Vop range).

Delivery 6:

For gain sampling, we used devices whose Vop range is 68.06V to 68.10V. Viased voltage was 68.08V (average of Vop range).

Delivery 7:

For gain sampling, we used devices whose Vop range is 68.51V to 68.55V. Viased voltage was 68.53V (average of Vop range).

On 9/27/19 12:48 PM, Nathan Grau wrote:

Hi Craig,

Thanks for the information. I don't recall the conversation where I was told that Hamamatsu wouldn't/couldn't divulge the details of their gain measurement -- maybe Eric knows. In principle, if we had more information on a device-by-device basis, e.g. breakdown voltage or slopes, that would make everything much easier. But we do not. I've been trying to monitor some QA histograms to see if there are any systematic issues in the testing. But there are none obvious to me.

I am also not certain what the gain values that Hamamatsu reports exactly means. If Vop is determined at a fixed gain, why don't all devices I receive from Hamamatsu have a fixed gain? Why don't the 30 SiPMs per shipment for which they report a gain value aren't distributed around 2.3x10^5? One outstanding worry is that I don't understand the data to which I am trying to compare.

Thanks for restating where I see things: that the effect on the constant term in the energy resolution is small. But I'm hedging that a bit based on what Gabor is independently looking at.

Thanks,

Nathan

On Thu, Sep 26, 2019 at 4:44 PM woody <woody AT bnl.gov> wrote:

Hi Nathan,
  Thanks for the slides. This was indeed a very interesting study. I can't say that I understand all that's going on, but it's somewhat disturbing that we can't reproduce the 40 mV spread in the Vops that Hamamatsu is supposed to be giving us. We know from previous measurements that our SPS method of determining the gain was very nicely correlated with Hamamatu's method of measuring the gain.
    I'm not sure why they consider their method of determining the gain a trade secret since we've discussed this with them many times. Maybe they've changed their method, but in the past, they have simply used a DC light source that they calibrate with a PIN diode for which they know the absolute QE (which gives them photons/sec), and then use that to measure the current from the SiPM for which they (presumably) know the QE.  However, maybe the problem is that they don't know the absolute QE of every device and just have to rely on the average, or maybe they use some other trick. I think we can probably find this out from them with some coaxing, but I'm pretty sure, whatever method the use, it probably involves measuring a DC current rather than the SPE spectrum, and we do know that there can be some difference between the two. However, I would think that may result in some overall systematic difference between our method and theirs, but it should not give any batch to batch variations (unless perhaps there is some variation in the QE from batch to batch).
  Nevertheless, the good news in all this is that even if the spread is as large as you measured, it doesn't seem to have a huge effect on our energy resolution. It would certainly be good if that were true. However, as we discussed in the meeting yesterday, things will only get worse after radiation damage, so let's hope the overall effect remains small.

Cheers,
Craig

On 9/26/2019 9:43 AM, Nathan Grau wrote:

Hi Craig,

Here they are. I'll try to get them posted today. I thought that Eric might move some things to the Sept 25 date. Your input would be much appreciated.

Thanks,

Nathan

On Thu, Sep 26, 2019 at 2:24 AM Craig Woody <woody AT bnl.gov> wrote:

Hi Nathan,
   Could you please send me (or post) a copy of the slides you showed at
yesterday's Electronics Meeting ?  I'd like to go back and have a closer
looks at some of the results you showed.

Many thanks,
Craig

--

Nathan Grau

Associate Professor of Physics

Augustana University

2001 S Summit Ave

Sioux Falls, SD 57197
(605) 274-5012

--

Nathan Grau

Associate Professor of Physics

Augustana University

2001 S Summit Ave

Sioux Falls, SD 57197
(605) 274-5012

_______________________________________________
sPHENIX-electronics-l mailing list
sPHENIX-electronics-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-electronics-l

-- 
Eric Mannel, Ph.D.
PHENIX Group 
Dept of Physics
Brookhaven National Lab.
631/344-7626 (Office)
914/659-3235 (Mobile)