[Primexd] [New Comment] FCAL LED pulser

[shepherd at jlab.org]

Comments:
--
 - by shepherd on Mon, 11/19/2018 - 21:15.

Sasha said: - 
Sasha said:

“You mentioned that HVs for the inner part of the FCAL had been equalized based on LEDs on 10/25.
FCAL channels with the large LED spread don't match completely with channels where the pi0 peak
is shifted, I am confused”

In the absence of any other information we attempted to adjust HV using the LED knowing that the issues I noted are problematic.  Now that we have pi0 peaks we should use that for HV adjustment.  The thing the LED is good for us mapping gain change to HV change.  I think we have all this coded so it can be easily done now.



 - by somov on Mon, 11/19/2018 - 20:40.

Ok, these plots are more - Ok, these plots are more consistent with that, which Iliya presented during the meeting on Friday
(in the sense that pi0s can at least be seen now). As I said, LEDs cannot be used for the calibration. 

You mentioned that HVs for the inner part of the FCAL had been equalized based on LEDs on 10/25.
FCAL channels with the large LED spread don't match completely with channels where the pi0 peak 
is shifted, I am confused. Anyway, we have calibration from Iliya, we start from there.


 - by shepherd on Mon, 11/19/2018 - 19:45.

I would be very cautious - I would be very cautious about interpreting anything about relative channel-to-channel gain variations by looking at this plot.  The system was not designed to be a uniform light source for every block.  

In layer 1, there is significant radiation damage, as Colin noted.  This varies quite a bit as one moves around the beam hole.  The LED system is the most sensitive to radiation damage, much more so than measured derived from true showers.  It is the canary in the coal mine.  The downside is that it is hard to assess the impact on physics -- one needs to study variation of pi0 mass with run for some energy of photon.

The biggest problem with trying to do any channel to channel comparison using the LED is the variation of injected light.  Recall that it is surface imperfections in the acrylic that cause light to get into the bar from the LEDs.  There are a lot of edges around the beam hole as the corners of all four sheets meet there.  Our intent was to have a light source that was time stable for each block but not one that was the same for every block.  It is much different than a fiber injection system.

With a small change to the skim thanks to Sean and no modification of the clusterizing, Colin was able to regenerate pi0 plots and we now see pi0 peaks clearly in every block.  I believe this is with production HV settings and software gains of unity for the inner three layers.  So, the deviation from the pi0 mass is a measure of how unbalanced the gains are.  I attach as a PDF -- you can easily see the variation isn't quite as dramatic as Sasha sees in the LED.  I suspect it is because the LED source is pretty nonuniform in this area of the detector.



 - by somov on Mon, 11/19/2018 - 18:52.

This plot is for the nominal - This plot is for the nominal voltages.

I would not expect rad damage to be very different for adjacent cells 
(where amplitudes differ by a lot for some modules, see layer 3).

We need to look on gains. The LED is not the best tool to look on rad 
damage, only a small layer of the glass at the entrance can be damaged 
(before shower is developed).

 - by gleasonc on Mon, 11/19/2018 - 18:18.

Sasha, is this for the - Sasha, is this for the standard production HV settings? Or are all channels set to the same HV?

One thing you will need to keep in mind around the beam hole when using LEDs is the amount of radiation damage in each block (see this image https://halldweb.jlab.org/wiki/images/e/ea/4x4_array.pdf for an image of radiation damage for each channel around the beam hole). The lower amplitude of the bottom row could be due to radiation damage and not a "wrong" HV set point (we know it's wrong in the sense that it is uncalibrated, but there may only be a small correction needed). For this row (27_30,27_29, and 27_28), the HV is set to the max EPICS value of 1800V with the production settings. 

 It would be interesting to check this with the green pulser, at say 25V, as that is not affected as much by the radiation damage. Nonetheless, I look forward to seeing the result with properly tuned HV.

--

Logentry Text:
--
FCAL FADC waveforms for 3 inner FCAL layers induced by a violet LED pulser operated at 15 V
(data is taken in a raw fadc mode). Each histogram represents the FCAL module. Histograms are 
normalized to 500 fadc counts. FADC amplitudes have different peak values even in the 3rd FCAL
layer, which can be accounted for either different light collection or PMT gains 
(we'll correlate amplitudes with the calibration coefficient obtained by Iliya).

---

This is a plain text email for clients that cannot display HTML.  The full logentry can be found online at https://logbooks.jlab.org/entry/3628625
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/primexd/attachments/20181119/bdb34e8c/attachment-0002.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: application-pdf.png
Type: image/png
Size: 346 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/primexd/attachments/20181119/bdb34e8c/attachment-0002.png>