Needing more than a spark test?

[homebrewed]

There are high speed op-amp circuits that can ride up a pulse, and stay stuck at it's peak until discharged. I will agree the scheme has the merit of low cost simplicity. It has been a while since I used LTC6244. They are now part of Analog Devices.
--> PEAK DETECTION

I think the nature of the pulse - how sharp, or broad is it, or is it two pulses partly overlapping, is better to know than to just seek it's peak. We now have low cost high bandwidth A/D conversion chips available that can capture a pulse with enough samples through it's duration to get it's shape, including it's peak. Should we consider a pulse trigger circuit which starts a sampling, capturing (say) a gated 1second's worth at some sample rate like 250KHz?

If the pulse is faster, we increase the sample rate, and maybe reduce the sample interval.

We won't know until we have some hardware and time to play with it. I doubt the Theremino/MCA implementation uses a track/hold, possibly it's doing something similar to what you're suggesting. If so a sample rate of ~40KHz might be sufficient.

The case of two overlapping but different-energy photons could be a difficult nut to crack. If you have a quantity C = A + B but you don't know A and B, the possible solutions are infinite. This is worst-case though, because you know, in general, what set of elements could be present. The algorithm would have to be informed by the set of possible photon energies. This is not an unusual approach to take -- the EDX S/W I've used gives you the option of selecting an element and telling the S/W that it HAS to be present; or tell it that its guess at an element is totally wrong. I've seen the S/W think that something like Europium or Hafnium is present and that just wasn't possible.

 

[RJSakowski]

You need to consider the decay times for the scintillators. This will determine how much acquisition time is required. The CsI(Tl) has a decay time of 900 nsec. CsI has a decay time of 16 nsec., NaI(Tl), 230 nsec.

 

[graham-xrf]

It occurs to me that if there were several (say 3) repeats of a single pulse that were similar height, one could subtract the sample values from a more smeared muddle of more overlapping pulses, and expose some other peaks.
Maybe not realistic - just thinking out loud.

I have ordered a cheap eBay "new other" PMT, and I am giving consideration to a new Hamamatsu part. The PMTs do not need to have a responsive bandwidth beyond what the scintillator material can provide.

The inorganic scintillator materials produce a much slower pulse , I read microseconds to tens of microseconds and more. The plastics produce very fast pulse flashes, in tens of nanoseconds. Easy enough to count, but not so easy to capture and measure. The information from @RJSakowski suggests the measurement is a much more high speed thing - unless the photomultiplier provides a lag of it's own.

At first, I plan to use my scope. It is a digital storage type good to 500MHz. I want to see these things.

The eBay CsI Scintillator in post #56 seems kind of tall.
https://www.ebay.com/itm/Scintillat...636979?hash=item2620f44673:g:QS8AAOSw8b1Z56ok

Does the cylinder of scintillator have to be that thick? It looks as if one could cut several discs out of it.

I haven't yet bought any scintillator. Still thinking on that.

 

[homebrewed]

You need to consider the decay times for the scintillators. This will determine how much acquisition time is required. The CsI(Tl) has a decay time of 900 nsec. CsI has a decay time of 16 nsec., NaI(Tl), 230 nsec.

Ouch! Sounds like a peak-hold circuit would be a more cost-effective approach. Thanks for the reality check!

 

[homebrewed]

It occurs to me that if there were several (say 3) repeats of a single pulse that were similar height, one could subtract the sample values from a more smeared muddle of more overlapping pulses, and expose some other peaks.
Maybe not realistic - just thinking out loud.

I have ordered a cheap eBay "new other" PMT, and I am giving consideration to a new Hamamatsu part. The PMTs do not need to have a responsive bandwidth beyond what the scintillator material can provide.

The inorganic scintillator materials produce a much slower pulse , I read microseconds to tens of microseconds and more. The plastics produce very fast pulse flashes, in tens of nanoseconds. Easy enough to count, but not so easy to capture and measure. The information from @RJSakowski suggests the measurement is a much more high speed thing - unless the photomultiplier provides a lag of it's own.

At first, I plan to use my scope. It is a digital storage type good to 500MHz. I want to see these things.

The eBay CsI Scintillator in post #56 seems kind of tall.
https://www.ebay.com/itm/Scintillat...636979?hash=item2620f44673:g:QS8AAOSw8b1Z56ok

Does the cylinder of scintillator have to be that thick? It looks as if one could cut several discs out of it.

I haven't yet bought any scintillator. Still thinking on that.

I found that particular scintillator on eBay, too. I was put off by the claim that it's "pure" CsI, no indication that it's thallium-doped. You could contact the seller to see if they know any more, but often eBay vendors are pretty ignorant about what they are selling. I found some CsI scintillators that were removed from a PET scanner for less money and bought some to try. I'll need to buy an eval board for a SiPM though. That will be a little more money compared to an NOS PMT, but, since I already have some lab power supplies I can save some money there, compared to getting a PMT + HV supply.

I had a sudden thought that the scintillator material itself would exhibit an XRF spectrum that might cause problems w/regard to analyzing steel alloys, but a quick consultation with my Thermo-Scientific mousepad showed that cesium, iodine and thallium all have vastly different X-ray lines. Their K-alpha lines are right up there (Thallium's is 72.176Kev), and their L-alpha lines are significantly lower in energy compared to vanadium through nickel. Dodged a bullet there....

 

[rwm]

I agree that eBay scintillator crystal seems a little too thick. I wonder what the optimal thickness for x-ray interaction would be? Thinner than a gamma camera or PET detector I suppose since the energies are lower. Although there may not be a disadvantage to a thick crystal? What to encase this in to prevent light leak?
Robert

 

[RJSakowski]

Ouch! Sounds like a peak-hold circuit would be a more cost-effective approach. Thanks for the reality check!

Since you have no control over the timing of the next event I would think you would want to use a fixed window for scintillator photon collection. The timing for the open window would start at the first photons to hit the detector, as seen by a rise in the output voltage from the detector. Photons would be collected raising the output voltage until the window closed. The captured voltage would then be analyzed and a count added to the appropriate register, and reset and the window opened again for the next event. A simple circuit to capture the peak voltage would be a diode-capacitor with provision for discharging. An operational amplifier circuit would be more complicated but probably faster.

It seems to me that this is very similar to the way a solid state camera works.

 

[graham-xrf]

I agree that eBay scintillator crystal seems a little too thick. I wonder what the optimal thickness for x-ray interaction would be? Thinner than a gamma camera or PET detector I suppose since the energies are lower. Although there may not be a disadvantage to a thick crystal? What to encase this in to prevent light leak?
Robert

We need to read up on NaI(Tl) (I think) and CsI (also doped) and others. I have to agree with @homebrewed that in the case of the eBay crystal, the stress on "pure CsI" in the description is perhaps disingenuous. We need a rather impure CsI, in the right way. I think if the correct scintillation doping is not present, the eBay crystal ends up as a fifty bucks desk ornament that will absorb water. Natively, it will scintillate, and is apparently "efficient" for extreme UV wavelengths.
--> Applied Optic Abstract CsI

There should not be a problem shielding the measurement from light. The end of the PMT is put hard up against the CsI, with a drop of optical immersion oil between, and then sealed up around the periphery. Then a mu-metal sleeve is slid over. All sorts of sources exist for that, but about $14 from eBay buys a sheet that can be (carefully) rolled around something circular to shape it, and secure with a couple of cable ties. A piece of black or grey PVC bathroom drain pipe fitting might do, with the mu-metal either inside it or over it.

Probably something soft between the glass PMT and the tube is a good idea. Over that lot, (or inside with the mu-metal) goes a sheet of lead, I guess 2mm or 3mm thick, that extends forward far enough to pick up the scatter from the alloy under test, but not so far as to block the illuminating radiation from the Americium buttons. (best ask @RJSakowski about that).

The turned fitting around it can be plastic or metal. It too can be conveniently fashioned from plastic pipe, but the fitting that mounts the Americium buttons will need a lead disc behind the sources, and a lead ring shroud. Some sort of soft, compliant, cleanable ring on the end to allow a light-tight seal when put up against a alloy. Alternatively, a box that mounts the probe, and a sample of the alloy put in it - light tight.

Photomultiplier tubes are probably not OK to have the high voltage turned on unless already dark. I think there needs to be a light sensor diode in there to use as an interlock, and perhaps a fast-acting inherent protection limit against overload currents. We have to pick through the existing radiation detector examples. They must have got to the point of switch-on without busting the thing!

I am committed. I have purchased a cheap "new other" photomultiplier to mess with, develop circuits, etc. but I think I will end up with a new Hamamatsu PMT, or try out a silicon avalanche photodiode array, or both. If each element of an 4 x 4 array has it's own connection, we get to capture some separate events that might otherwise have just shared the same PMT photocathode.

I know - I get it. That last thought comes with 2 x 8-channel Analog Devices A/D converters. Don't worry. It only gets to play if it is large enough area, very affordable, and widely available.

 

[graham-xrf]

I like RJ's suggestion.
Even though it picks out a single scintillation, locking out others until the the window content is grabbed, it does limit the probability of a separate photon contribution being so coincidentally timed as to "join in" the scintillation. Triggering a gate with a sample run-out a tad shorter than the decay time of the crystal, say 1uS, and then storing it in the buffer while the next pulse is collected may work. The false positives would get locked out if a pulse was unreasonably too large, and coincidental smaller ones also lose statistically if their average does not correspond to a valid element. A kind of "smart digital filter ", if you like.

If the pulse magnitude is within some limit tolerance, it's samples can be averaged with and stored as a sorted capture likely to belong to one element. A bunch of other sized shots, with such similarities, can be put in another bin as likely to belong to a different element. These "bins" are an array, with another dimension enumerating the number of pulses deemed to have contributed.

You get a set of averaged pulse samples, and a count likely to be proportional to the percentage % content of that element. The other array number represents the count (popularity) of that set. The plot could be Y-Axis scaled to suit whichever was the winner.

I expect that would most times be iron, but you never know. @RJSakowski might hold it up against one of his fish to check for mercury or something!

 

[RJSakowski]

If you look at the photon yield and decay time of the various scintillation crystals there seems to be a trade off between the two. A lower photon yield means more noise and a longer collection time but a longer decay time increases the probability of having two events occur in the same collection window.

We actually used the atomic absorption spectrometer for analyzing for ppb mercury in effluent. At the time, the company made mercury hearing aid batteries and mercury was added to zinc carbon cells so our Dept. of Natural Resources became very interested in our waste water.