spenceraxani / cosmicwatch-desktop-muon-detector-v2 Goto Github PK

When connected in coincidence mode the count on the slave device is running about 0.15 signals per sampling time when connected and viewed on the Cosmic Watch site. Is this a reasonable value to expect? It is far lower than the signal rate when each device is monitored independently.

Hello, I'm having an issue where part U2 has been overheating. I have been able to create a full circuit that gets the correct voltage but after a couple minutes of running this part will overheat to the point where it burnt to the touch. Have you had this issue come up before? If so is there a way to fix it? I replaced the part itself and the issue continued to happen.

Thank you,
Abbey Siegel

I've built two boards with the same issue, there is 0.4 V coming out of the DC-DC booster for the SiPM instead of 29.5 V. I've checked each component in the booster circuit, including the connections on the LT-3461, and everything seems right. Has anyone seen this problem before? Any help diagnosing this problem is appreciated. Thanks!

The CosmicWatch paper has a really nice plot of count rates during a flight. I'm wondering how many people have tried this measurement and if they'e experienced any problems getting the detector through security. Any notes about your experiences would be greatly appreciated!

Spencer,

I just noticed that my count fluxuates with the power supply I use. For example, when I plug the detector into my professor’s Mac the counts sink down to .250. When I plug it into a wall socket or my Mac the counts rise to .422. I will be building a second Detector as you suggested but thought this might be relevant to my low counts (beyond the likelihood of a less effective scintillator).

Jim

Hello,

I'm trying to determine which DC-DC booster to use from here: http://www.linear.com/samples/LT3461A

The purchasing list says to use the LT3461, but the instructions PDF says to use the LT3461A. These are two separate models with one slight difference: the LT3461 switches at 1.3 MHz while the LT3461A switches at 3 MHz.

Is there a specific one I should pick over the other, or would either work fine?

Thanks!

Hello,
Would it be possible to put a note on the website (http://www.cosmicwatch.lns.mit.edu/detector#steps) about this version because I (and maybe a few other people) keep getting the versions confused as I cannot find any reference to this version other than through the facebook page and pdf (https://arxiv.org/pdf/1801.03029.pdf). Maybe I am just not looking properly.
Thanks.

After a week of operating my Detector reads .2847 with +/- .001. I didn’t see count guidelines in the repository materials (I easily could have missed it given my track record in building the Detector). Are these readings about right? Haven’t had the chance to download the raw data to my Mac.

Just a suggestion for V3 hardware: As the quality of power delivered via USB can vary substantially between different sources (5.25v to 4.4v allowable), I would recommend using the Nano's 5v regulator to source 5v to the rest of the board. The on-board regulator is capable of sourcing 800ma @ 5v, so this should be plenty for the other equipment. By doing this it should help cut down on any noise coupled into circuit from the USB power supply as well as give you more consistant 5v rail making the results between different people's equipment more comparable.

This would involve bringing back the power jack and connecting it to the Nano's Vin and ground pins. You just need to supply the jack with 6-12vdc there after. Given the simplicity of this mod, I'm planning on employing it on my V2 equipment once I get my parts together.

Hi Spencer,
I don't see a listing for the face plate screws. Looking for a recommendation please.

Thanks,
Michael

Hey Spencer,

I'll try and upload pictures in two days if I can as the project is currently in the lab, but I will explain the problem as best as I can.

Currently, the muon detector does not detect any muons. Well, technically, it usually detects exactly one when it is reset or receives voltage. When we try and give the main PCB an artificial wave pulse, we do "detect" muons, but when the SiPM is put in place, it results in only the one detection.

When we used a multimeter on the VH going to the SiPM, it results in 29.66 V, which is in line with what it should. and so we thought it was a problem with the SiPM PCB. We redid the SiPM to make sure it wasn't a problem with any of the components but it still results in the SiPM component heating up, which is probably not a good thing, and no detentions.

There was a point at which C6 was not populated correctly, but we didn't think that that was a major problem.

Another note is that we are currently lacking the Temp sensor as we were trying to keep the budget small but will order one if absolutely needed.

We have 2 detectors which seem to work by themselves. But when used in coincidence mode, the slave detects even more than the master. We're not sure what would cause this issue, but were wondering if it's been seen before.

I’m sure this is a rock dumb question but does it matter functionally that I installed the Arduino

Spencer,

Thanks to you and others I now have a fantastic waveform through the BNC, looking just right. A huge sense of accomplishment. I also have a functioning OLED and LED. However,

1. I get nothing through TP1 (seems impossible since TP1 is a direct connection)

2. I get a wave through TP2 which pulses in tandem with the BNC wave but which is much smaller than the BNC wave (seemingly the opposite of being amplified); this may be some glitch in how I have set up Channel 2 in the scope.

3. I get nothing through TP3

4. The OLED shows no counts and the LED doesn't flash

I had no glitches in loading the Arduino software and the splash screen works fine. I have quadruple checked every solder joint.

If I understand the debugging notes in the manual this points to a problem with the Arduino software? If so, I suppose I could cut and paste the source code from Github, compile and reload it?

Happy New Year down there,

Jim Gallatin

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.

I’m sure this is a rock dumb question but does it matter functionally that I installed the Arduino

Hi,

Maybe that's a dumb question, but I think it would help more persons to have a better understanding of the data we collect from the detector.

I'm doing some collections and I'm having no success to analyze that data. I think I'm doing some misunderstands.

For example, below, this is the data I collection after 24h.

And that was my understanding of which row:

Is that correct? How or where can I find the routines of the analyses I see at the website, like the periodogram and the histogram? Should I consider which line a new count?

Thank very much!
Rogerio

Hi Spencer,

I've just finished building two detectors and I noticed that there was no white-dot (which I assumed is to signify a used connection of that pin) on the ARef pin. This makes soldering the pins of the nano to the mainboard a little ambiguous and can result on a floating reference voltage. In my case, I had to debug why I was always measuring 1023 ADC units on A0.

I also wanted to say, thanks for making a fun little project!

-- Kurtis

Hi,

I'm having problems to find the C6 capacitor (20nF - 50V) to buy at my country. Could I have some problem in the funcionallity of the detector if I use a 22nF capacitor instead the nominal 20nF?
Thanks!

Hello Spencer,

We are attempting to put this detector on a rocket at Spaceport America Cup, and see the change based on altitude as our scientific payload. However, we are having quite a few issues with the detector.

1. There are time stamps for only some measurements. We are not sure why this occurs. What are ways we can ensure there is a timestamp for all measurements
   

2. The measurements are not consistent and do not seem to make sense. The only way we get measurements is if we press down on the detector. Even then, it is very inconsistent. I replaced the 6 pin header, however, the problem persisted. The ADC values also seem to be off, they are only ranging from 0~10.
   
   
   

Thanks,
PSP Solids Team

Spencer,

I have a good looking pair of PCBs, SiPM and Main. I have the appropriate current throughout. But no counts. I went through the Main PCB components again touching up the solder connections and made the weird discovery that when I touch the soldering iron to the capacitors between the peak detector circuit and the Arduino, or a capacitor in the peak detector circuit, the OLED shows a batch of counts.

I plan to replace the components where this happens under the theory that when the soldering iron comes into contact with the circuit that it closes an otherwise open circuit. Does that make sense?

Once I have gotten through Christmas I will build a third, and hopefully perfect set of PCBs and send them to you for review.

Happy Holidays to you and the team in Antarctica,

Jim Gallatin

Sensl was taken over by ON Semiconductors. Their website for purchasing a SiPM is no longer functional. I have so far not been able to reach ON Semiconductor to see if they well support selling small numbers to individuals. Does anyone have an extra SiPM that they might be willing to sell?

Hello,

Could you please add a parts legend? E.g which resistor values on the board go with which value.

I am having issues having the single detector reading out data consistently. The Oscope data for the most part seemed good, but TP2 did seem noisy when I lowered the trigger threshold. The issue I am having is nowhere lists the resistor values for it. The picture of the button shows a 300 ohm resistor but we bought a 249 ohm resistor for R3 as the parts list suggested. I did try retouching the legs between the peak and amplitude detector.

It seems after running it several times on average it seems that the count increments every 6.66 seconds after letting it run for several minutes. However it is by no means consistent sometimes it will go minutes without blinking then it will rapidly start blinking/counting and catch up.

Here is my recorded Oscope data, when I looked at TP2 for the second time it seemed that once the peak was detected it would go crazy up and down, but when it was slightly higher it seemed normal.

When the arduino is connected to the cosmic board I cannot connect the arduino to my computer to change the code. There is no USB recognized. When I disconnect the Nano from the board, it connects fine. Is this normal? I have uploaded the software with the Nano off the board and then added it to the board but I get no counts. All of the voltage tests are good.

When the power is connected to the arduino the oled shows the Cosmic Watch logo and the clock advances every second but the count does not change after counting once. I changed the const_threshold from 20 to 15 but no change.

Not sure what to do.

Thank you,
Robert Baker

Ok weird problem here:

My arduino keeps connecting and disconnecting from my computer as I try to do anything with it, especially if the SiPM is connected via the 6-pin connector.

From what I can gather from arduino webpages, this means that the current draw may be too high, but they were powering a servo, and this circuit shouldn't have too much draw.

https://forum.arduino.cc/index.php?topic=496581.0

I'm assuming from a lack of GitHub issues on the topic that people generally have not been having problems with this. I am using a genuine arduino nano from arduino.cc because I am basically 0-7 with the cheaper arduinos (terrible luck but I guess that's just how $5 arduinos work).

Any suggestions on how to fix this issue or what may be causing it? I am running Windows 10 if that helps.

Hi Spencer,

I'm having a little trouble with my Main PCB. I fully populated the PCB and plugged the Arduino into my power bank. (I tried with a phone charger as well, same result.) At first, I measured 4.6 V between the HV and GND pins, after which I resoldered the legs of the booster, which gave me about 33 V output. So I've resoldered a few legs again and now I'm stuck with something between 23.5 and 24.5 V, varying when I put some new solder on one of the legs - or take some away - and apparently also when room temperature changes (I just closed my window and voltage dropped from 24.5 to 23.5 V). Could I power the SiPM with that voltage, and do you have an idea where my trouble might be originating from?

Plus, I noticed that the inductor gets really hot while plugged in, should I worry about that? The booster heats up as well but not as intensely as the L1 inductor.

This is my circuit, if that's of any help.

Thanks a thousand,
Theo

Spencer, I am switching from your email to this account to declutter your life.

I have built two sets of SiPM and Main PCBs, and think I have them working. The voltages are correct, and there is current throughout. (I only have 29.3V on one PCB’s SiPM connectionbut I don’t see that as a problem.) Soldering in good shape after some redos and replacements. However, I am not getting a signal through the BNC, and no counts on the screen or through the Python data reads.

Since my first two scintillators were less than perfect, I am assuming that my core issue is the Scintillator or Scintillator-SiPM connection. I plan to assemble a third combo but want to get some adult supervision before my third try. (I used to work with NVG technology but this is way different.)

Please excuse any and all ignorance here.

1. I am confident I have wrapped the Scintillator well in one layer each of tinfoil and tape. Should I move to two layers of each as some have suggested in these comments?

2. My Scintillators are clear but were very slightly warped during heat treatment. Think concacity. Issue?

3. It is conceivable that the SiPM is not super flush against the Scintillator, but I did use a good amount of optical gel as you recommended. How flush does it need to be? I have been assuming the muons have nowhere else to go with all the wrapping.
   .

Using Windows 10

error message:
arduino_fail.txt

Also having difficulty sourcing LT3461A -- trying direct from China.

thanks
Bryan

Hi Spencer,

When I connect to a scope, I am reading negative pulses. I cannot identify where I went wrong. Is there a way I can change the code to correct this? I tried making my threshold and resets negative, as well as flipping the inequalities, but it did not count. What do you suggest?

My detector was not functioning properly, and I decided the issue was likely the scintillator or SiPM or both. While unwrapping the scintillator, the SiPM became loose, and I was concerned an edge was crushed. Having the SiPM pressed up against the scintillator concerned me, so to cushion them I bought neoprene washers (90133A005) from McMaster Carr. 0.12" ID, 0.25" OD.

I had originally hand drilled the 4 holes before hand wet-sanding the scintillator using 600 to 2500 grit wet sandpaper and regular white ink-jet paper, but the second time around I touched the scintillator up with cotton flannel and the Novus system per these instructions: https://eljentechnology.com/images/technical_library/Plastic_Polishing.pdf. This really contributed to the clarity. Spencer states elsewhere that all machining must be done before heat polishing. There are any number of materials that will fracture if drilled into at high-speed, so slow speed drilling while keeping the material cooled with water is what I typically do.

After a careful rewrap - wrapping the tinfoil like a miniature present and adding a second layer, I put optical gel on the SiPM and carefully screwed the SiPM PCB to the scintillator. I sighted the gap between the surface of the scintillator and the SiPM PCB to make sure I wasn't crushing the SiPM.

I rewrapped the scintillator with electrical tape.

I also added the neoprene washers to the short standoffs also.

This time everything works great. Photos added in case anyone else needs to use this approach.

Hi, Spencer!

I just assembly and welded the project during that last week. With the arduino routine compile in the arduino nano, I started to make the tests in the main PCB. Unfortunatelly, at the pins 4 and 5 of the U1 component a smell of burning appears, like if something was in short, at even with a light in the pin.

Did you have some problem like this or do you know what could be happening? I decided to pick off the component from the PCB and analyse the main PCB weldings again.

Anyway, in the first seconds of test, the OLED worked perfectly, showing the initial message and the name I gave to the detector.

Sincerelly thanks,
Rogerio.

Hi, Spencer, do you use lead-free or tin-lead solder? Thanks!

Hi, Spencer, I am seeking to test the SiPM prior to wrapping it up with the scintillator. The basic tests with the voltage meter work fine. I placed the SiPM in a black box in a dark room and ran jumpers to the 6 pin connector then allowed light (ambient) to enter the box through a hole. The OLED/LED did not register any additional event. Is it possible to register an event without the scintillator - possibly by adjusting the threshold? And could you let me know what the threshold of 20 is equivalent to?

Also, since we are only using anode and cathode on the SiPM, I'm not able to grasp how the current flows through the female 6 pin on the board. Per the circuit diagram, the 6 look like they are mirrored 3 pin columns. I also bought one of SensL's SMTs and am unable to figure out how to connect it to the 6 pin.

Thank you as always!

Hello,
A quick question on the performance difference between polished and unpolished scintillators.

I have built two detectors. I was not quite game enough to begin with to heat polish the scintillators, so I left the surfaces as they were (one side glass like, the others slightly rougher and opaque, with the SiPM attached to the glass like side). Both registered around 1.5 counts per second.

Later, I had to take one of the sensors apart to fix a dry solder joint on the SiPM. While I was at it, I had a go at heat polishing the scintillator using a gas soldering iron with the tip removed, practicing on a bit of old acrylic first and being very careful not to burn the scintillator. This sensor now reads at around 0.6 counts per second. This seems to be the opposite as I expected, although now more in line with what the website suggests as a standard count rate.

The scintillators are made from Bicron 408 plastic according to the listing that I got them from and the instructions that came with them seemed to suggest that performance was better with them unpolished, although the instructions mainly seemed to focus on round scintillators attached to photo multiplier tubes, as opposed to square scintillators attached to solid state sensors with screws.

My question is: Is this difference in performance expected or have I made another dry solder joint or some other error somewhere and if there is, why?
Thankyou very much.

Hi, Spencer, what is the reason for selecting a blue-sensitive SiPM instead of a red-sensitive SiPM? (What nm range is the flash in the scintillator?) Thanks!

Hi Spencer,

I have built several devices and they all report slightly different rates which isn't surprising. I have noticed that when running in coincidence mode, the rate reported when connected to the MIT server is slightly lower than that displayed on the OLED or saved to the MicroSD card. This occurs when running the systems with a windows PC or when running a Linux machine. Any thoughts?
Thank you, Paul Edelen

Hi Spencer

I'm currently building my CosmicWatch and I wish to get some accurate voltage measures. Considering I'm using some cheap copies of Arduino Nano, should I carry out the calibration test again and make an ADC to voltage conversion function for my device? Looking forward to your advice.

Cheers,
Tien

Hi, I'm ordering from Elecrow and I have these issues:

1: The spreadsheet with part quantities does not match the quantities on the populating list. So basically from the quantity list they think they have more soldering to do than the populating list and pictures of the board show.
2: Elecrow is asking for a partlist generated from Autodesk Eagle - which I don't know how to do with the Gerber files. Do you have a DXF? Can you generate that partlist? Thank you.

Spencer,

I have finished a new Main PCB (I’ll send pictures next). I am pretty confident that the soldering is adequate. I have a solid scintillator and SiPM combination. I have 29.4V at the SiPM HV/GND points. I thought I was fresh out of reasons for why the resulting detector won’t work when assembled and plugged in.

And yet. When I assembled and plugged in, everything lit up, screen came up, etc. However, within seconds after plugging it in, the LT 1807 became too hot to touch and faint smoke was emitted.

To be sure it wasn’t that particular 1807, I replaced it. Same result (didn’t wait for smoke). I assume I have a short!? Any ideas?

The Part List specifies 3/8" stand-offs for SiPM PCB. The interfacing screws are 1/4" in length. Using two of these screws results in a clash as both screws come in contact inside the stand-off. The SiPM also droops. This can be corrected by using a slightly longer standoff.

It's possible to free up and additional 1.5k Flash by loading the latest Uno Bootloader onto the Nano. This may provide enough space to allow both the OLED and SDCard to operate together.

Reference: https://arduino.stackexchange.com/questions/51866/arduino-nano-atmega328p-bootloader-difference

With an update to the Nano bootloader, the good folks at Arduino neglected to reduce bootloader partition from 2Kb to 500b, as was done for the Uno. Loading the Uno bootloader onto the Nano provides this benefit. You just have to remember to identify the board as an Uno in the IDE.

If you have a sketch with both the OLED and SDCard functions active, you could see if it properly compiles for the Uno. If so, it would work for the modified Nano.

Hello Sir,

I wanted to know about the following lines in code..

"const long double cal[] = {-9.085681659276021e-27, 4.6790804314609205e-23, -1.0317125207013292e-19,
1.2741066484319192e-16, -9.684460759517656e-14, 4.6937937442284284e-11, -1.4553498837275352e-08,
2.8216624998078298e-06, -0.000323032620672037, 0.019538631135788468, -0.3774384056850066, 12.324891083404246};"

How you decided these values & how to calculate them..
I am confuse as to how to explain these values...

Thanks.

Hi, I'm just getting into this project, and I have a quick question about the SD card sockets.

What are the requirements for the SD card sockets? You link to a $12 purchase on Amazon, but since we are only making one, it's more economical to purchase an individual model from Digikey.

Here is a selection of SD card sockets from digikey. As you see, there's quite a few.
https://www.digikey.com/en/supplier-centers/m/molex/microsd-memory-card-connectors

Will any of these work, or is there a significant difference? In the video of how to make the main and SD PCB boards, you said you ordered the wrong ones, but didn't explain what was wrong with them.

Hey,
As the title states, your older code seems to work with count rates around 0.09Hz. I know this is rather low, but I don't recall adding light paste. The annoying thing is switching to V2, it claims there's one count whenever I reboot, but never catches anymore.
Oscilloscope readings show numerous peaks ranging from 30mV (trigger) to >100mV. Was there some change in the circuitry between the two iterations? Do I need to update the modules in the code? (not 1.0.2, 1.0.1, etc.)

Last time I checked, all the test points are reporting as intended.

With new code.

I see that a number of elements on the SMT_reference are simply listed as "NS." What does this mean?

Also, C14 is listed as a 0 Ohm jumper. Is this correct?

Hi,
Last week I finished the construction of my first detector.
Everything looks fine, but the counts are very high. By using a THRESHOULD of 50 I'm finding something about 200 counts per minute. I had to use 220 of THRESHOULD to find the expected counting of 0,5 per minute

I already polish heated the scintillator, changed the screws for something with a smaller head and recovered the package with a better quality black tape. Anyway, the counts still high. I'm working with +29.7V at the HV pin.

Is that value of THRESHOULD trustable to work?
I'm grateful for any help with that issue!

Att.
Rogério.

Hello,

My detector is trigering at about a 0.095 cps which I know is pretty low. I am having problems determining the problem. Now I know that my scintillator isn't exactly to size with yours, being 3cm * 10cm * 1 cm so maybe that's why ( Here is a short video of the 3 test connections. The initial change is me having trouble attaching the leads of the osciloscope, not the voltage jumping arround. While I'm seeing count rates of about 0.095 cps I am unable to see the peaks in the osciloscope. It's a cheap usb one, so maybe that's why.

I did all the testing provided in the FAQ, my SiPM is reciving 29.5 Volts, and it does have the correct polarity. I am pretty sure it isn't random noise as it is pretty consistent, giving 0.095 cps in a period of 5 hours.

I don't know, if you think this numers are correct factoring the change in scintilator I'm fine with it, it just seems on the very low end. Also I'm pretty sure the SiPM works as I can trigger it by letting light in or not.

SIGNAL_THRESHOLD = 10

RESET_THRESHOLD = 15

LED_BRIGHTNESS = 255

Thank's for all,

L

Hello Spencer,

In another issue post, you commented that the 1cm thinkness for the scintillator was chossen so that it would be easier to calibrate the detector to muons and not other background radiation.

I happen to have however a 2.75cm * 2.75cm * 10cm plastic scintillator with both ends crystal clear (it's either a BC408 or BC412 not really sure).

Shoud I just stick the SiPM in one end and wrap the whole thing to have a gamma + muon detector? Or should I cut a 2.75cm * 2.75cm *1 cm slab? What would you recommend?

Thank you making this awesome project,

L