Red Pitaya MCA

[GigaBecquerel]

Hello!

I am still looking for a good, high resolution MCA for portable fun and have stumbeled across the Red Pitaya for this.
There is an MCA application for it, with two different interfaces:
https://github.com/sefffal/redpitaya-mu ... elanalyser
http://pavel-demin.github.io/red-pitaya-notes/mcpha/

A 14 Bit ADC at this sample rate looks very nice for what I'm looking for, my only gripe is that there seems to be no way to set the channel number.
16k channels don't make much sense, as the LSB will be noise anyways, and more than 8k channels are just too much, even for most HPGes.
Has anyone got experience with the software, maybe I just missed this setting, as the hardware is still in the mail.
I'd love to set 4k or below to get my spectra decently fast, more doesn't make sense for Scintillators anyways.

Best Regards,
Lukas

[sgt_bear]

This looks like a fantastic project which all of the hardware design already there!
One of my future projects will be to build a mca my self with an fpga and mcu.

For sample rates and ADC-bits, it can be tricky, its not the ADC bits that count, its more the ENOB (effective number of bits). 8k Channels are actually 13bits, you need an ADC better than 13 bits. 16k would be a perfect 14bit ADC, which usually do not exist. One of the biggest issue of spectroscopy is that most of the signal is not of value, only the peak is of interest. Thats very hard to measure

[GigaBecquerel]

That's exactly what I meant, 16k channels with a 14 bit ADC doesn't make much sense, 16k doesn't even make much sense in most cases.
I'm not good enough in EE to make my own MCA and all the sound card solutions always felt a bit bodged, so the Pitaya looks like a good place to start.
If you get y working product we can start talking production! :D

[Conor Whyte]

Both programs have source codes on Github and are 100% modifiable. The author of the last program -MCPHA doesn't seem to understand wanting scalable bins and argued with me about not adding such a function in. He is under the impression that 16K bins is better than 4096 or 2048....

Anyways, maybe we should make our own (home-rolled) application?

[GigaBecquerel]

Making our own application would be awesome, but sadly I have zero programming skills.
Maybe I should get into contact with the author, too and try to convince him to add scalable bins... ;-)

[GigaBecquerel]

I got the board yesterday and it all works so much better than expected!
You can select between 1k and 16k channels, I am fine with that "compromize".
The input range of either 0-1V or 0-00V is a bit annoying, as nim signals are usually 0-10V, but that's a "non problem" and can easily be solved by a divider / amplifier.
The software is very rudimentary, but it can export .csv so you can do all your energy calibration, peak detection and efficiency calculation magic in your favorite spreadsheet calc program.

The nonlinearity in the screenshot below is because I used a non compensated divider, I just forgot that we are already in a timescale where that matters.

Screenshot from 2020-09-13 09-21-58.png (87.99 KiB) Viewed 28402 times

All in all it feels like a powerful tool for portable spectroscopy and I'll be sure to update you on what I do with it!

Lukas

Edit:
With the attached circuit it accepts 0-10V signals and uses the whole dynamic range without compression.

Screenshot from 2020-09-13 12-42-51.png (82.75 KiB) Viewed 28393 times

[Conor Whyte]

Using the Gamma spectacular as a HV source and an iRad labs splitter the incoming voltage is right within the 0-1V range and works (as you've suggested)
The first program ( the one you are showing here, MCA) is written by a local astrophysicist, Will Thompson who lives/lived in Victoria B.C.

btw, how did you get the program to display like this? what were your settings?
Mine works only in a logarithmic view otherwise nothing comes up.

[GigaBecquerel]

This is the regular way the software looks to me, I have not changed any specific settings etc.
if you only see stuff in log maybe you've got some noise peak that drowns out everything else?

Also, as I'm planning to build a portable spectrometer with the pitaya, so I'll have to do all custom hardware anyways, so I can fit the (pre)amt to my needs.

[Conor Whyte]

You are using a positive input? What about your baseline?
Mine was set to automatic baseline.... I think there is some quirkiness to this software due to Java.

What will you use for the HV supply?

[GigaBecquerel]

Yep, I am using positive input, coming from my Silena 7613 Nim Amplifier with 500 ns shaping time, of course divided down with the circuit above.
Here are my settings:

Screenshot from 2020-09-17 11-29-36.png (44.61 KiB) Viewed 28326 times

At the moment I am using some NIM HVPS, but for portable use I will go with a regulated CCFL inverter, in my experience they work great as HV sources, because they are pretty linear in output / input voltage and can be filtered for a very clean voltage.