Page 1 of 2

Scintillator counter questions

Posted: Sun Aug 12, 2001 11:02 am
by guest
I'm pondering the feasibility of hooking up a PMT to a computer via the serial ports and would appreciate any thoughts on the matter from the experienced neutron hands out there. The advantages I would expect would be the ability to log the data and measure the pulse height to screen out the junk signals.

I have a processor in mind, the Motorola 68HC11, some versions of which have a 10 bit A-D convertor with built in sample and hold that can do samples at a rate of 333 KHz flat out. By the time it runs any kind of code and pipes it out through the serial port it will be a good deal slower than that. With a 'HC11 it should be possible to build a 2-3 chip circuit for under about $50 US that would communicate with a PC, possibly running a visual basic program to give a neutron rate and pulse height and mathematically allow for distance from neutron source etc.

What I would like to know is what kind of pulse rates can I actually expect from a typical scintillator unit near a running fusor?. I understand efficiencies are a lot less than 100 % so I won't be getting high counts. I would guess if it was getting swamped I could bring it back from the neutron source till it counts ok again. Whats the currently accepted procedure? do you get a measured count and then plug it into a formula to get a real neutron rate inside the plasma?.

thanks in advance
Mark Harriss

Re: Scintillator counter questions

Posted: Mon Aug 13, 2001 8:47 pm
by Richard Hull
The answer to Mark's question hinges on his scintillator.

If a common piece of scintillation plasitc is used (As in Joe Zambelli's detector the counts will be in the thousands per minute BACKGROUND! The pulse widths are in the 100's of nanoseconds and a processor has to be listening during the active pulse interval to count it. The addition of fusor related counts which would be xrays + neutrons would really increase this figure. You would have your hands full catching all counts.

With a BC-720 Bicron Fast neutron scintillator, the count rate for background is about 1 or 2 counts/minute. The average fusor in operation at 20 cm would be only about 40 cpm. Overall, a much nicer scenario.

The effective neutron count from either a BF3 tube or a BC-720 is so slow and easy that a simple digital counter can follow the input and data reduction reduces to drudge work with a couple of simple equations for solid angle and counter efficiency correction.

The key to making the common scintillator plastic work with high counts is a known neutron calibration source. Joe Zambelli used the source at a local college. Without this calibration point, the Scintillation plastic is, effectively, un-usable.

A lot of info on the BC-720 can be found on the original Songs board (clickable on this sites login screen) in the early 1999 time frame.

Richard Hull

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 1:37 am
by DaveC
Are the high background counts for the scintillaor plastic due to UV, X Rays , or what?

To eliminate low energy background, one mfr of G-M tubes (LND) calls for using a 50 mm thick lead shield to establish background. He states the tube background is 10 cpm inside the shield.

Dave Cooper
Dave Cooper

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 3:20 am
by guest
The high rate low level background counts are probably due to the dark current of the PMT. Even a good bialkali cathode gives off a few thousand electrons/cm^2 at room temperature. S-1 cathodes are far worse. These counts are low level, so they can be rejected with a counter with a input threshold setting on a good or using a single channel analyzer in the integral setting. This would be a problem with any sensor using a PMT. A lead shield will also help get the background count down further. Cooling the PMT is one possible expedient for reducing the dark current, but may be too much trouble. Fortunately, with the fusor, there is a combination of factors that help the amateur experimenter to get around the relatively low neutron flux levels:
1) If you are using a properly shielded metal vacuum chamber, you will be getting neutrons and not much else. If you were trying to measure the same neutron flux level from an accelerator based system, the accompanying gamma, hard x-ray and other garbage would make it much harder to measure the neutron flux. This is the reason why the BC-720 was developed. It is rare to get a situation where you get neutrons without a lot of gamma rays as company.... We get relatively soft x-rays, which are mostly stopped by the vacuum chamber walls. A lead shield around the neutron detector will get just about everything else but the cosmic rays.
Richard Hull makes a good point about the need for calibration in a plastic-scintillator based neutron detector system. I am working on a plastic based detector, and plan to make it self-sufficient enough to be able to send it to an establishment like Ludlum to get it calibrated with a known neutron flux. The price is right (~$60, if I remember correctly), though it may take some cajoling on my part to get them to work on something non-standard. Another possibility is to make friends with someone at Stanfor d or SLAC who has acess to a neutron source. I'll cross that bridge when I come to it...

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 1:06 pm
by guest
Hi Richard, I was kind of banking on doing some pre digital processing like pulse stretching and discrimination, also lots of lead shielding with some 50 year old lead I have obtained (old phone cable sheath). I'd planned to try some cheap scintillator plastic first in the hopes it could be shielded enough to screen out the dross. The BC-720 sounds very usable with less than 100 counts/sec or so.

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 1:19 pm
by guest
Hi Richard Hester, Dave mentions 50 mm of lead for a shield around a GM tube but I thought they did not pick up anything much apart from beta, gamma and x-rays, would 50mm (2") of shielding be a good ballpark figure to work on?. How exactly does an analyzer work? it should be possible to program a PC to work as one with a suitable interface shouldn't it??.

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 2:22 pm
by guest
Both Mark and Richard Hester's comments are well thought out and I can re-iterate the need for old lead in shields. Old lead of the same thickness can reduce background by a factor of 2 or 3! Especially, if you are using a pancake counter or other alpha GM tube. I use tungsten plate to quiet my counters to check for "life left". (old counter tubes - surplus - may have most of the main gas or quench component just flat used up) This ups their count rate and shortens the gieger plateau or GM region to near zip.

The lead I have came for a 200 lb mass of buried lead I got from a friend who dug it up at an old confederate foundry site a few years back. I have sold or traded most of it now and have only a few lbs left. Regular lead is about 20 cents per pound. The market for old lead can see the stuff sell for $2.00/lb.

Korff has probably written more practical, sage advice on GM tubes than any living sole. His work is so good, that I am tempted to make a bunch of the tubes for myself!

I would probably go unquenched and quench electronically with some personally desired fast quench circuitry which would allow a huge increase in count rate over commercially quenched tubes which are often limited to 10k-30k cpm.

Geiger tubes can indeed detect alpha with proper mica windowing. A superb example is the classic 5679 mil spec GM tube. LND makes a superb 2.5" pancake detector that is the best I have ever used. It is almost too sensitive, being swamped by moderately strong sources. Therefore, it is best used to sniff out trace amounts of alpha or beta radiation.

GM tubes are terrible Gamma counters with Scintillation counters being the detector of choice here for survey work and limited spectrography. GM tubes, being gas amplification detectors, are worthless for spectroscopy or any form of energy determination, although I have seen some really slick, calibrated blocker targets with a critical GM assembly give amazing results on beta particles.

Plastic scintillators will flash with Gamma or X-radiation in the 15kev region! Cosmic radiation and normally un-counted x-ray backscatter is a real problem. Of course good lead shielding would allow for a more quiet neutron counting environment.

Remember the BC-720 is just an undoped piece of "buy it at any store" chunk of polystyrene, so the gamma scintillation is way down in the mud as opposed to the anthracene doped stuff used to photon scintillation. Bicron adds the special ZnS-(Ag) phosphor and this makes recoiling protons that hit it look like atomic air bursts compared to a gamma photon's firefly like whimper. This make discrimination abyssmally easy being over two orders of magnitude appart in some situations. You gotta' see it on a scope to believe it.

Richard Hull

Re: Scintillator counter questions

Posted: Tue Aug 14, 2001 10:57 pm
by DaveC
Mark - I believe LND are just giving the condtions used to establish a 10 count/min background. Their same tube, without the extra shielding, gives 16-17 avg counts/min as bkgnd. So the 6-7 cpm difference is what 2 inches of lead will do to reduce bkgnd. For an 80KeV photon, thats about 10E-57 attenuation!!

Data I have on an end view LND 712 G-M detector indicates it can be used for Alpha, Beta and Gamma detection. ( site is www. LNDinc.com). I have used this tube to detect 1.7 kEV Xrays from electron bombardment of Silicon. (One mil of Al foil stops these!) The LND tube has about 6 um thick mica window.

For an Xray detector, I like to have this level of sensitivity in a survey meter, since if it sees nothing, the energy level is extremely low.


On Analyzers.( presumably PHA's or MCA's. ) You can build a pretty simple device with either a TTL or ECL high speed comparator, and a DC reference for the discriminator. Assuming your PMT or rad detector output is above 10 mV , its output can go to one input of the comparator. The DC reference goes to the other, giving an adjustable threshold. Comparator output goes to a decade counter. (Build or buy this as you see fit.) Counter output and the discriminator voltage voltage ported to the computer via A/D, D/A interface gives you the guts of a useful MCA or PHA function. By sweeping the discriminator voltage, from background cuttoff to full cutoff , one gets a plot of "count rate for voltages (energies) greater than", from which with a little differentiation by the computer, will give the pulse height distribution function, as a function of the detector (PMT, GM..) output level. But only with the PMT will you get any significant distribution of pulse heights. The GM is all or nothing.

Years ago we built an analyzer like the one outlined above, and operated it successfully at 380 MHz rate, with an input signal in the 1 - 10 mV range. We did not sweep the threshold, though.

Today, you can get up to 2.5 GHz chips to do that for a couple bucks.

Dave Cooper

Dave Cooper

Re: Scintillator counter questions

Posted: Wed Aug 15, 2001 9:40 am
by guest
Hi Dave, Thanks for that information about the lead and the GM tubes, does the mica behave like a crystal and transmit radiation along certain axes?. I think that idea about sweeping the discrimination voltage is absolutely brillant!!. I'd have no need for a fast A/D converter, the 68HC11 has 4-5 channels of pulse width modulation I could generate the swept voltage from with a lot of filtering and I could use something like a 74HC4040 14 bit binary counter to count the pulses up to 40 MHz or so and pipe the count up to the PC every now and then. Can you recommend a few of those IC part numbers for suitable devices?.

Thankd Mark H

Re: Scintillator counter questions

Posted: Thu Aug 16, 2001 5:57 am
by DaveC
Mark - Regarding the mica window: It probably does have some diffracting capabilities, but they would not be apparent the way it is placed on the end of the tube. The radiation just goes through and ionizes the gas inside. The LNDinc.com page, as Richard Hull has noted, is well worth a visit. You will find some typical operating voltages, pulse rates and I think even the fall (quench) times for typical G-M tubes. These will help you with your design.

Regarding what chips to use.... I did not find much data yet for ECL comparators,although Motorola's list shows ECL categories, but two chips made by National Semiconductor are probably fast enough... LMV7219 which has a transition time of 7 nsec and only draws 1.1 mA. Runs on (+/-)5 volts and has rail to rail output.
The other is the LM360, which has a 1.3 nsec transition time. Check out the National Semiconductor site... their data sheets are available in PDF with more complete info. One of the two above is only available in SOT (surface mt) pkg, the other can be gotten also in DIP.

One thing comes to mind regarding the pulse height distribution scheme, and that is the discriminator sweep has to be slow relative to the count rate, so you get the average counts at or above the reference level. If the count rate is real slow... then you could do the A/D conversion with your uP 68hc111, directly. The method I suggested tends to work better with a fast pulse rate.

Dave Cooper