correct voltage for HARSHAW NAL(TI) 12S12/3 SCINTILLATION DETECTOR 3"

[ilinty]

Hi to All

I have bought in Ebay this:
HARSHAW NAL(TI) 12S12/3 SCINTILLATION DETECTOR 3"x3''

And I have build and resistor partitor found on the GeigerCounterEnthusiasts group on Yahoo...
I have used resistor of 4,7Mohm each...
I'm doing experiments with "gamma spectacular" for gamma spectrometry. I have had good results, but, I want a "fast" spectrometry, so what is an good voltage for this detector ?
Previously I used 900v but it takes 10 hours to get a good spectrum of the background.
With1200v I get good results in 1-2 hours. You think, I can go up to 1500/1800 volts? What is a good voltage for a fast spectrometry with my detector ? Can anyone help me ?

The PMT is a RCA8053...

[Richard Hull]

With increased voltage comes increased noise. I would rarely wish to push a gamma-spec tube much over 1000 volts unless I knew the tube had 12 or more dynodes. (most have 8 to 10 dynodes.) Apparently you are looking at some low level stuff if it takes that long to grab a spectrum. Needless to say, I would not even approach the design limit of the tube which is 1500 volts. I would not push beyond 1200 volts.

Richard Hull

[Carl Willis]

10 hours....?! An unshielded 3x3" NaI detector in good condition should produce thousands of counts per minute from background in the ~>30-keV range, so something's wrong. Here are some general suggestions for troubleshooting a scintillation detector:

-Check your supply voltage with the socket assembly connected (using an electrostatic voltmeter, an HV probe, or some other reliable method.) Make sure it is what you think it is.

-Examine the audio output on the PMT circuit with an oscilloscope for pulses. There should be many of them, even from background. Does the frequency increase when a test source of some kind is approached to the detector? Are there light leaks (turn on and off room lights, check seals with a flashlight, etc.)? Is the shaping acceptable? Is there evidence of periodic (repeating) noise?

-Pay attention to cable lengths between the PMT and the audio input device. Signal-carrying cables should be as short as possible to reduce capacitive loading.

-Since you indicated you are using the PRA / InTune software with a PC sound card, consider the software settings for mic volume, pulse shape analysis, and pulse-height threshold. In my experience, passable results with a scintillator are dependent on slavish attention to these features--and the larger the detector and the higher the count rate, the more challenging it is to tune and the lossier the system is due to pileup rejection. USE A TEST SOURCE to adjust these settings and verify the presence of clear spectral features. A 3" detector will be quite difficult to use.

-Be careful of raising voltages above about 1 kV. Standard BNC coaxial connectors are only rated to 1200V, and can only be used reliably to about 1500V before they break down.

[Jeroen Vriesman]

Hi,

found some very good documentation about PMT's, you can download it from my blog:
http://nuclearfusionreactor.blogspot.co ... ector.html
under "Everything you ever wanted to know about photomultipliers but were afraid to ask is ..."

[Steven Sesselmann]

Ivan,

Pushing up the voltage with GS-900 and PRA means that you are increasing the volume, not the count rate. Increasing the volume, will give you more resolution at the low end, and may distort the high energy pulses.

PRA is not the right tool if you want to measure very high count rates, because it works by discarding improperly formed pulses, which includes double pulses.

Over a certain count rate PRA will plateau.

If you want a spectrum in less time, the best way is to start with a high resolution detector, and carefully calibrate your pulse shape using Intune. If you can get your CS-137 peak to within 8 or 9% FWHM you will be able to get a spectrum quickly.

I can get a Cs-137 spectrum in 20 seconds with PRA from a 0.1 uC source.

See video here...Gamma Spectacular

Steven

[John Futter]

Ivan

I think that your probe has become contaminated with water in the NAI crystal

What you are reporting is consistant with that.
Most Harshaws are difficult to disassemble to check the condition of the NaI crystal.
This can only be done at the crystal /PMT interface and albeit quickly.
The optical grease will stop water in the air getting at the crystal interface at this point.
If you can do this the crystal should be clear as good glass, any discolouration means it is shagged, this discolouration can go from a milky appearance to a nasty green yellow colour.


slightly milky means no good for energy discrimination but maybe OK for counting, all other conditions it will not work in the nuclear field --ie salvage PMT for something else.

[ilinty]

Hi to All.

Thanks for all the answers.
I think I've got a good result with 1200V.
This is the spectrum of bg, obtained with 10 hours of sampling.
And as you can see, to meet Carl, I get more than 160 pulses per second (9600 pulse / minute). Unfortunately I have no sources at my disposal. Now I should try to sample for less time, for example, only half an hour (1800 sec). Having sources even ten times the level of the bg is pretty obvious that I should be getting a much shorter time. (But 20 sec of Steven seem to me very less!)
The nice thing is that the bg spectrum obtained in Aosta in Italy is almost equal to that achieved by this gentleman in the U.S.: http://www.randomuseless.info/spectra/r ... index.html
With an other equipment!
So also to answer John seems that the detector works quite well.

Someone can explain the peaks where I put a question mark?

[ilinty]

And this peak ???

The spectrum is the same as before, but the "channels" scale is much larger....

Bye

[Steven Sesselmann]

Ivan,

Possible that the peak right at the high end may be some noise from a cable or possibly double up pulses. If two pulses fall exactly on top of each other, you will get a double height pulse.

Your spectrum contains peaks from NORM or Normally Occurrring Radioactive Materials.

Similar to what I got when I took a spectrum of a ceramic toilet bowl.

viewtopic.php?f=13&t=6177#p40955

Your 3" detector is getting a better resolution at the high end, than my 1.5" ....your doing well!

Steven