peculiar waveforms

[northern-nuke]

With the benefit of a friend's digital scope I have some waveforms at last. Both units are showing what I take to be rapid pulses from the detector and high levels from the peak circuit -- leading to some counts when the interval is long enough for the peak to decay. It is deceiving because when first powered up unit 1 seemed to be counting. Now that I have seen these waveforms I hope my suspicions that light leakage is the problem will turn out to be correct. I will try to use more black tape and see if there is improvement. Photos (inverted from black background): Unit1_TP1_TP2, Unit1_TP1_TP3; Unit2_TP1_TP2 Note Unit2_TP1_TP3 no file .. behavior similar to Unit 1. Note that the height of the SIPM pulse with "Unit2" is less than 100 mV whereas "Unit1" is ~ 180 mV.

[spenceraxani]

Hi,

The problem is coming from upstream of TP1, which means that it has to be with either the biasing voltage (DC-DC booster circuit), or SiPM/SiPM PCB. TP1 and the BNC receptacle on the back of the detector are connected to the same line (pretty much — there’s just a 1k resistor between the two). Since TP1 looks odd, you don’t have to think about the other test points, you need this one working first. Counting is not an indication that the detector is working. The signal on TP1 should look like a ~10ns rise time, followed by an exponential decay of roughly 1us. The amplitude will probably be ~30mV on average. From the size of our scintillator, these pulses should come in at the single Hz scale.

Your images look like they might be set to scale the voltage by a factor of 10:1. My interpretation of the images you sent are that you are seeing a high rate of square wave pulses that appear to be on the order of 200us from TP1, is this correct? I’m looking through my emails, and can’t find pictures of your PCBs. Have you sent me them already? If not, you might want to take a picture (particularly of your SiPM PCB), so that I can have a look. These waveforms suggest a new problem for me.

The image below gives an idea of what the pulses should look like at each of the test point connections (TP1 =A, TP2 =B, TP3 = C).

Thanks,

Spencer

On Dec 3, 2018, at 5:27 PM, northern-nuke notifications@github.com wrote:

Unit2_TP1_TP3

[northern-nuke]

Thanks Spencer

I attached the "images" of the digital scope traces in my comment on github.

I added more black tape to the "bottom" side of the PCB and the edges on both detectors. I haven't visited my friend with the scope again yet.

With this mod the count rates became much more reasonable. Unit 1 is systematically counting fewer than Unit 2 -- there was more abuse to the Unit 1 sensor.

Prior to attacking the light leakage, the Unit 1 sensor waveform was ~ 100 mV at TP1 while the Unit 2 sensor was ~180 mV.

With 6 hours of counting side by side Unit 1 counted 10124 and Unit 2 16500. (61%) With Unit 1 on top of Unit 2 (slaved) for 1 h 15 min the counts were 2377 and 482. With Unit 1 upside down on top of Unit 2 (slaved) for 16 h 36 min counts were 26884 and 13085.

I am hoping beyond all reason this means they are working. (A health physicist friend suggested inverting the upper unit to minimize the spacing of the scintillators to reduce cosine effects.)

They have calmed down markedly with the extra tape. Silly me -- I thought the ground plane on the pcb would have been sufficient.

The smaller amplitude with Unit 1 lead me to wonder about reducing the discrimination threshold so the side by side results would be more similar (assuming I did some damage to the SIPM)?

regards Bryan

[spenceraxani]

Hi Bryan,

I just checked your numbers and they seem reasonable. The count rate on unit1 is rather low, meaning that the average event in your detector is seeing fewer photons. This could be due to many things, such as:

1. the coupling from SiPM to the Scintillator isn’t sufficient
2. the surface of the scintillator where the SiPM couple to isn’t transparent.
3. perhaps the reflective foil isn’t as good as the other unit
4. damage to the face of the SiPM.

You can lower the threshold on Unit1 (or raise the threshold on Unit2) if you want them to match. The cosmic ray muons deposit quite a bit of light on every interaction, and it’s likely that you are still triggering on the vast majority of them with Unit1.

The count rates that you provided are approximately what I expected, and the behavior of the slave detector is appropriate. This is a good indication that your detectors are working. However, I’m now confused by the oscilloscope traces you sent in your last email.

Thanks,

Spencer

On Dec 6, 2018, at 7:06 PM, northern-nuke notifications@github.com wrote:

Thanks Spencer

I attached the "images" of the digital scope traces in my comment on github.

I added more black tape to the "bottom" side of the PCB and the edges on both detectors. I haven't visited my friend with the scope again yet.

With this mod the count rates became much more reasonable. Unit 1 is systematically counting fewer than Unit 2 -- there was more abuse to the Unit 1 sensor.

Prior to attacking the light leakage, the Unit 1 sensor waveform was ~ 100 mV at TP1 while the Unit 2 sensor was ~180 mV.

With 6 hours of counting side by side Unit 1 counted 10124 and Unit 2 16500. (61%) With Unit 1 on top of Unit 2 (slaved) for 1 h 15 min the counts were 2377 and 482. With Unit 1 upside down on top of Unit 2 (slaved) for 16 h 36 min counts were 26884 and 13085.

I am hoping beyond all reason this means they are working. (A health physicist friend suggested inverting the upper unit to minimize the spacing of the scintillators to reduce cosine effects.)

They have calmed down markedly with the extra tape. Silly me -- I thought the ground plane on the pcb would have been sufficient.

The smaller amplitude with Unit 1 lead me to wonder about reducing the discrimination threshold so the side by side results would be more similar (assuming I did some damage to the SIPM)?

regards Bryan

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/spenceraxani/CosmicWatch-Desktop-Muon-Detector-v2/issues/28#issuecomment-445086873, or mute the thread https://github.com/notifications/unsubscribe-auth/ABpeu3K4BjrcFYTk-PzinpnTF0vmAlAyks5u2b8pgaJpZM4Y_vHm.

[northern-nuke]

I have some new scope data - will post tomorrow. The TP1-3 waveforms look much better with the added black tape - but the amplitudes are still smaller than in your Instruction Manual. Apart from the possibility of damage to the SIPM, I wondered about that thins smear of silicone grease I'm using as a coupler? It looked "clear" when it was squeezed between 2 pieces of PMMA -- but it might not have the necessary spectral properties? I haven't thought of any other common mode sins.

[spenceraxani]

It could be the silicone grease. You want something with a high transmission percentage and an index of refraction of about 1.5-1.6.

On Dec 10, 2018, at 7:55 PM, northern-nuke notifications@github.com wrote:

I have some new scope data - will post tomorrow. The TP1-3 waveforms look much better with the added black tape - but the amplitudes are still smaller than in your Instruction Manual. Apart from the possibility of damage to the SIPM, I wondered about that thins smear of silicone grease I'm using as a coupler? It looked "clear" when it was squeezed between 2 pieces of PMMA -- but it might not have the necessary spectral properties? I haven't thought of any other common mode sins.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/spenceraxani/CosmicWatch-Desktop-Muon-Detector-v2/issues/28#issuecomment-446043471, or mute the thread https://github.com/notifications/unsubscribe-auth/ABpeuxpd55nanzvZDwAPZMrkl28LL3h2ks5u3xCTgaJpZM4Y_vHm.