Charge sensitive amplifier for photomultiplier

[StillTrying]

To preserve the shape of 200ns wide 20uA to 1mA current pulses a single transistor is very hard to beat, even the physical layout isn't very critical, a small piece of stripboard is fine if you're sensible with the strips.

[jmelson]

Yeah making a base with two connectors might be the nicest solution. I don't have a PMT yet, but I'm eyeing one of this guy's kits on the Ebay:
eBay auction: #https://www.ebay.com/itm/3-Scintillation-Detector-BASIC-KIT-Includes-PMT-Scintillator-Electronics-BNC/263250251252

It's pretty cheap, it's a nice, large area PMT and I get a Scintillation crystal in the same diameter.

Note that those plastic scintillators give a VERY short pulse, just 10-15 ns, compared to crystalline scintillators like NaI.
So, any circuitry after the PMT needs to be able to properly handle those extremely short pulses.  Unless your preamp
has a 5ns rise time or so, you will end up with no signal on the output.

Jon

[ChristofferB]

I would skip in the test input, just to keep the input capacitance small. If a test signal is need, create it optically with a LED.
Similar I would avoid the extra connectors / longer cable to the detector - here every pF counts. The FB capacitor would be more like a gimmick wire close to the input. AFAIR I even skiped the coupling cap and had the detector directly at the TIA. For a final version I would consider a low noise OP here (e.g. OPA172, OPA141).

AFAIK there may be one Alpha source sometimes still available: some old ionizing type smoke detectors.  They are phased out some time ago so it could become difficult, possibly even illegal to get. A tiny piece of uranium mineral could be an option too. No need to have much as only the upper µm counts. Radon gas is also a Alpha source, but hard to handle in vacuum  . Thorium in some form (e.g. welding electrode, gas lamp) may work too.

That sounds reasonable, I'll try putting the diode directly on the bias circuit and ommit the longer lead.

I was under the impression that gimmick caps were only useful at fairly high frequencies. I put in a 0.5-5 pf trimmer cap for now.

I'll figure the source thing out, even without any source I should  still see some background pulses.

Yeah making a base with two connectors might be the nicest solution. I don't have a PMT yet, but I'm eyeing one of this guy's kits on the Ebay:
eBay auction: #https://www.ebay.com/itm/3-Scintillation-Detector-BASIC-KIT-Includes-PMT-Scintillator-Electronics-BNC/263250251252

It's pretty cheap, it's a nice, large area PMT and I get a Scintillation crystal in the same diameter.

Note that those plastic scintillators give a VERY short pulse, just 10-15 ns, compared to crystalline scintillators like NaI.
So, any circuitry after the PMT needs to be able to properly handle those extremely short pulses.  Unless your preamp
has a 5ns rise time or so, you will end up with no signal on the output.

Jon

That's good to know. Yeah it looks as if there's a reason why the NaI(Tl) and CsI(Tl) with/without thallium are very popular. They're also less robust, and more expensive. Getting a 3" one would likely cost an arm and a leg!

But it should be possible, right? Using a high speed op amp of course. Actually the datasheet for Bicron BC412 lists pulse width as around 5 ns. Still, they seem generally applicable, even on field scintillation counters.

[sdouble]

If all what you want is counting, you do not need an amplifier I guess. The pmts have a typical max gain of 10^6 to 10^7 which is usually enough to get pulses of 100s of mV at least.
for the detector you were mentioning, you will get pulses of 10s to 100s of ns duration which is long enough for triggering a scope for instance.
Typically people use a discriminator to deliver a logic pulse out of a pmt signal. the good ol' NIM logic is still in use. negative (-800mV/50ohms) high.
wheb people are after highly accurate timing they do use a constant fraction discriminator to limit the so-called walk effect.

[ChristofferB]

What I really want to do is pulse height analysis and stuff like coincidence/anticoincidence experiments. So after going through a shaping amplifier to widen the pulse, and a gate to generate a logic indication of a detection, I want to feed an ADC and then a multichannel analyzer.

[ChristofferB]

Actually, the OPA354 cmos opamp they used in the scintillator schematic I posted seems to be plenty fast enough for a plastic scintillator preamp. Maybe I should just use that then.

The reason I included the 'test' input is because every single preamp I've seen for setups like this, Ortec, Canberra etc. brand has it, and it is used to put a marker peak in your alpha/gamma energy histogram to calibrate with.

[Kleinstein]

For coincidence experiments at high count rate one may want to preserve a high speed. For just getting the pulse amplitude, there is no need for a super fast amplifier. The pulse shaping is essentially some well behaved band-pass filter. So the is not real need to have the input amplifier faster than the pulse shaper and amplitude analyzer. If the pulse is much shorter than the response time of the amplifier, one would still see a pulse, but a wider one and with reduced peak amplitude, but the same area under the pulse. So a fast amplifier is not absolutely needed. A slow amplifier could be seen as part of the pulse shaping.

For the amplitude analysis there are several options and this also effects the preferred pulse shape. The 3 main methods I know of are:
a) analog peak detector circuit and "slow" ADC  (probably the easiest method - I have build this type).
b) fast ADC and numerical amplitude calculation (usually more from area than peak)
c) analog integration with some baseline restoring and slower but higher resolution ADC and looking for "steps" instead of peaks.
 The last method does not need an extra peak shaper.

The amplifier should normally be stable enough, so that one does not need an extra marker. The weak point is more the stability of the high voltage that sets the PMT gain. So if at all a test pulse would need to be optical. The alternative is known isotope from time to time.

[r6502]

Hello all,

this thread is a little older, but I'd like to show a reverse engineered schematic of a preamp for an older Harshaw PM with an NaI (TI) crystal installed. Shematic is very simple 2 transistor version and they used older germanium types. Mine is the type 6S6/E detector.

I have attached some photos of the detector and its preamp and the schematic of the preamp.

it is still working fin. I use also some NIM modules for pulse sharpening and using a MCA card located in an older PC made by EG&G.

Guido

[coppercone2]

I have one of those scillnators with a crystal but I took it apart since it was not working, it had a broken amplifier tube circuit of some kind. I have been meaning to do something with it but the whole gamma thing is a little unsettling, the only project I made was a cast containment jar (like a heavy lip balm tin) to house the radioactive sample that it came with, very weak but I decided to keep it in a lead box, it was bugging me every time I thought about it, but the rating on the test sample is safe by goverment levels.

[r6502]

Hello all,

I had too look a bit, but I found a catalogue for Harshaw radiation detectors - it is hard to find on the web, so decided to share it here. I'm not sure about Harshaw, but when you search fot the company you can find the website of Saint-Gobain. In ther history of they point on Harshaw.
https://www.crystals.saint-gobain.com/about-us

A few years ago, when I got my 2 Detectors from Harshaw, I contacted Saint-Gobain, but they did not have any literature in the archive. For the 2nd detector, it is a 12SW12/3EX that has a larger crystal, I'm not sure, what photomultiplier Harshaw used. from the outside you are not able to see the mfg of the photomultiplier tube. If someone has an idea, what's inside - please let me knew.

Guido

[doktor pyta]

Thank You r6502 !

[penfold]

I'm speculating a bit, but it may be an ADIT brand tube... especially if Harshaw had any relationship with Bicron (I'm a bit unsure whether the dates check out there, not knowing the age of your tube and any history of ADIT before the early 80's)

Are there any clues about the diameter of the PMT itself? It looks like a 2" base but 1.5" crystal. I would guess its maybe something similar to an ADIT B51B03, I couldn't tell you if that was the historically correct model number used when the detector built, but I'd be pretty confident its the same design of tube

[ChristofferB]

Thanks for the info! It's good to have a reference design!

I've actually had so much success not using a preamp at all - just an R to bias and a C to a spectroscopy amplifier - with perhaps the highest resolution I've so far seen with my small 1.15x1.15x3 square scintillation probe. Attached below is a couple of hours of 232Th mantle counting time, background subtracted.

Obviously gain varies hugely with cable lenght this way!