Stretching and shaping analog pulse

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
Hello,
I am currently trying to design a PIN diode radioactivity detector. Starting from the simulations to avoid any assembly problems I've come across a problem concerning length and shape of the signal. For my little research i would need to measure those pulses with an 1MSPS ADC. (STM32 will be used) The most important is the peak value of the pulse as it determines the energy of a particle or a quantum.
The pulse from TIA lasts for about 1us and it absolutely needs some kind of stretching and maybe shaping.
Q:
-How much should the pulse be stretched if i want it to sample it with that 1MSPS ADC (or 2,4 that does not change much)
In other words how many sample points do You find sufficient to measure peak value of a pulse?
-Do you know any circuits schemes that could help me?
I've found some materials like this, but yet implementation to spice and my circuit has not been too successful.
Ideas:


or this http://www.simplecircuitdiagram.com/2840-averager-and-peak-holdextender-signal-conditioner/ ?

Inkedcirc_LI.jpg Inkedpulse_LI.jpg
 

Attachments

ericgibbs

Joined Jan 29, 2010
18,766
hi Z,
Can you post more details of the actual input pulse to the circuit, from the detector.?
Rate, Period, Source Impedance etc, is the pulse up from 0V level.?

E
 

MrChips

Joined Oct 2, 2009
30,706
You only need one sample to measure the pulse height.

Unfortunately, the STM32 ADC will not work for what you are trying to do. Your results will be poor because of differential non-linearity of the ADC.

You need to tell us more about the radiation detector being used and what energy resolution you are aiming for.
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
The source of the current pulses is a simple BPW34 photodiode, biased and boostraped (actually 10 of them in parallel). The particles hitting the semiconductor plate are supposed to create short current pulses measured by the TIA. The problem with the detection is that it is all chaotic and the only thing I can determine is the maximum detection frequency. I would like it to be as precise and wideband as possible. So that i am able to do some amateur measurements of gamma quanta and beta particles.

The only estimation i have seen is from circuitsalad: https://circuitsalad.com/2012/11/19/a-solid-state-photodiode-gamma-radiation-detector/
"For example: Americium-241 generates a 59Kev gamma pulse, therefore 59Kev/3.62eV = number of electrons = 16.3K electrons per gamma strike. This assumes all of the possible electrons and holes are formed , which is an ideal assumption and not realized.
Now to find Q in coulombs, we take 16.3k * 1.6×10^-19 = 2.61 ^-15 coulombs"

I am afraid that anything lower than 10^-15 would be impossible to measure due to the noise.
So resolution starting form this 10^-15 and ending at the maximum adc or amp operational point voltage*(after multiplying the current by TIA and stages, let's say the max is 1,8V )
Still undefined is the source pulse length, which might be assumed as much as a few hundred nanoseconds max? like 300ns?
 

MrChips

Joined Oct 2, 2009
30,706
Next question, what radio-isotope are you trying to detect?
What is your energy range and resolution you expect to measure in your spectrum?
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
Next question, what radio-isotope are you trying to detect?
What is your energy range and resolution you expect to measure in your spectrum?
Everything that is possible starting from the environmental radiation to some artificial sources like thorium rods or working Roentgen lamps. That does not sound professional, but this is only an experimental circuit.

The idea with sampling (OPA615) would be great if i knew exactly when to sample. Additional rising/falling slope trigger would be necessary for reliable performance.
 

MrChips

Joined Oct 2, 2009
30,706
Everything that is possible starting from the environmental radiation to some artificial sources like thorium rods or working Roentgen lamps. That does not sound professional, but this is only an experimental circuit.

The idea with sampling (OPA615) would be great if i knew exactly when to sample. Additional rising/falling slope trigger would be necessary for reliable performance.
I think you are going to be terribly disappointed.
There is very little background radiation. The largest source of background radiation will come from potassium-40 in your body.
That small PIN does not have sufficient volume to make it an effective sensor for nuclear radiation.
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
No worry I'm used to that :V
I have some backup sensors devices like glass photomultiplier, chunk of scintilator and maybe i will acquire some semiconductors dedicated to radiation detection.
 

MrChips

Joined Oct 2, 2009
30,706
No worry I'm used to that :V
I have some backup sensors devices like glass photomultiplier, chunk of scintilator and maybe i will acquire some semiconductors dedicated to radiation detection.
In that case, why don't you get your system working with a scintillator and PMT first before trying the PIN diode?
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
I want do try something yet not so widely described (but promising ? :V)
I will take pains to try and adjust a CR-RC shaper to my needs
 

ericgibbs

Joined Jan 29, 2010
18,766
hi,
As per MrC.??

I think you are making it over complicated, all that is required are OPA's configured as Amplification [ for Calibration] and Sample/Hold.
The LTC6269 you are using is a suitable 500MHz OPA.

E
 

MrChips

Joined Oct 2, 2009
30,706
Correct.

What you need is a peak detector, sample and hold. Do ADC and reset the peak detector.
You will also need an analog comparator to detect when a pulse has been received.
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
The idea was to shape the pulse - gaussian way then sample it and get a few points to fit a gaussian curve with matlab or something like that. With that gaussian curve I would measure the peak of it and analyze stacked pulses of each peak.
Yeah that would be it...

Your idea seems legit and better so that I will start again from the circuit simulations but this time using Altium Designer and Texas Instruments components as they have Pspice libraries. I'm disappointed with the Linear spice models being limited to their own software so let's manage with more universal ones.

So I plan it like that:
1)TIA stage with LMP7721 (17MHz, bias:3fA, low noise)
2) 2X Amplification stage with OPA2301 (150MHz low noise...)
3)Peak detector and holder with OPA615 and TL3116 comparator (10ns)
 

BobaMosfet

Joined Jul 1, 2009
2,110
Well, let's make this more simple. If you want to register a voltage, then you need (just like an oscilloscope) you need to read a voltage at a frequency at least 5 times faster than the frequency you are monitoring. Next, in any given MCU with an ADC, when you configure the ADC, one of the steps is to set the ADC frequency. So, this is where you establish your scan frequency, so that when you fire off the ADC once to read that voltage, it does it n times and averages that value before turning the averaged value around to you.
 

Thread Starter

Zuavirr

Joined Dec 28, 2017
25
Hmmm I still haven't got my STM32F4 dev board started and so on...
Is there any replacement for OPA615, because it's a little bit too fast for this application and maybe a slower one would be cheaper?
I intend to do a few of them ( a matrix to get more counts ) so reducing the price is always nice ;)

Although, i don't need a sample and hold amplifier if i can do it by a synchronized comparator and peak detector . In that way i would get a peak voltage clamp which is exactly what I need. There will be some issues with timing but let's hope I can optimize it all in the simulation.
 
Last edited:
Top