Diagnosing Scintillators

[glhansen]

Hope you don't mind being bombarded with questions.

What should I think when a scintillator, like a 44-2, seems to have a lot of noise? Like, plateau it, the "as found" voltage was on the slope, I put the voltage where I think it should be, and it counts vigorously in a low-background area where another will barely count at all.

Can you look at the shape of a plateau and immediately tell if there's a problem? When I struggle with a scintillator, it seems like the plateau sort or resembles a nearly-square trapezoid with the left corner chopped off. I just don't like the shape of it. Do the locations of the knee and the high-voltage point where the count rate rises, or distance between them, give useful information? Does the count rate on the plateau give useful information? Or efficiency, I suppose.

Does the deadtime of a scintillator increase over time for some reason? For a 44-2 I want to see 9 microseconds. 10 or 11 makes me nervous.

When I tee off a meter to a scintillator and to the pulser, the voltage drops when the scintillator is added. That's expected, since the PMT draws some current, and the meter has a low-current power supply. But sometimes it drops a few volts, sometimes a few dozen. Does the voltage drop say anything about the PMT?

I'm looking for some guidelines to tell when the equipment is healthy, when it's on the way out, and when it's neither good or bad but just different from the last one. I don't know why it me took so long, but I'm thinking of which data to start collecting, or to look up in the records, so I can make comparisons.

[Sesselmann]

Greg,

The recommended voltage for the 44-2 seems to be extremely wide according to Ludlums.

https://ludlums.com/products/all-produc ... model-44-2

I think it would be reasonable to operate it at 650 to 700V but before you set the voltage you need to understand the voltage drop.

When you are operating a detector with one coaxial connection, the load resistor in your spectrometer forms part of the divider, therefore you will always see a voltage drop on the anode when you connect the detector.

Start by finding the impedance of your divider, you can either measure it via the BNC connection or open it and look at the value of the individual resistors.

Say for arguments sake the divider is 12M Ohm and and the load resistor in your spoectrometer is 1M ohm, then the voltage drop will be 1-(12Mohm/13Mohm) = 0.923

When I tee off a meter to a scintillator and to the pulser, the voltage drops when the scintillator is added. That's expected, since the PMT draws some current, and the meter has a low-current power supply. But sometimes it drops a few volts, sometimes a few dozen. Does the voltage drop say anything about the PMT?

Not quite sure what you mean by tee off? I trust you realise it is not possible to measure the voltage with a regular multimeter, this doesn't work for the same reason explained above. If your multimeter has an impedance of 2M then the voltage drop would be 1-(2/3) = 0.33

I'm looking for some guidelines to tell when the equipment is healthy, when it's on the way out, and when it's neither good or bad but just different from the last one. I don't know why it me took so long, but I'm thinking of which data to start collecting, or to look up in the records, so I can make comparisons.

Your best option is to post images of spectra and count rate histograms, then we would be able to judge the condition of your detector.

Steven

[glhansen]

What I mean by tee off. Per the procedure I learned from Ludlum, a tee connects the meter, for instance a 2241-2, both to the probe, usually a 44-2, and to the pulser, a 500-2. So the meter has a tee on the external connector, two cables attached to it. The pulser has a high-impedance input, displays the voltage, and can also make pulses, but pulses are not used for this. For this purpose I just use the voltage display. The process then is to repeatedly set the meter voltage and do a count. (I have a voltmeter, but the impedance is way too low for that purpose.)

When I unplug the probe, I see the voltage go up. I have always assumed that's because the PMT draws a lot of current compared with the pulser, so the voltage I find with the probe plugged in is the correct voltage because the meter will be used with the probe plugged in. I have always assumed that the pulser draws a negligible current. The manual says the voltmeter impedance is 1000 megaohms.

[Sesselmann]

Greg,

If the PMT draws enough current to affect the voltage display, then you should be able to measure the resistance with a regular multimeter.

PMT Voltage dividers range from 4 Mohm to 120 Mohm depending on their use.

Steven

[iRad]

Ludlum 44-2 probes are 111 Mohm resistance.

[glhansen]

I've figured it out. I had wondered why, when I got a scintillator that had been worked on by someone else, my plateau voltages tended to be higher than the "as found". Then, when I set it where I thought it should be, sometimes there was a problem with noise. I always wondered whether that was the equipment, or me.

Way back when, I had been taught to use a tee on the meter body to connect both the scintillator and the pulser at the same time, and the pulser was just used as a voltmeter. (We had a voltmeter, but it draws way too much current.) My assumption had always been that the pulser doesn't affect the count rate because of the high-impedence input.

Well, that was wrong. When I unplugged the pulser, the count rate went up. So I did a pair of plateaus, using the pulser to see the voltage, getting the counts, then unplugging it to get a second count. My plateau voltage was much lower the second way. In retrospect I think I see what was happening. The pulser may have a high-impedence input, but it must also have capacitance. It must have been cutting the amplitude of pulses from the PMT, so the voltage had to be turned up higher, the amplification increased more, before the pulses were big enough to get past the discriminator. The count rate on the plateau was the same in both cases; once the pulses were big enough to get past the discriminator they were all being counted.

I have a scalar now, and I repeated it on that and got a comparable result.

I don't think the amount by which the voltage falls when the PMT was plugged in has any meaning. A Ludlum tech said that the older meters have a stronger power supply than the newer surface-mount versions.

I feel like I'm reinventing the same wheels that everyone else has been using for years. But I think that's one more problem cleared up, which is nice.