Third order IMD in photomultipliers--weird behaviour

Hm, one other thought is, do you have enough capacitors to stabilize the voltage on the later-stage dynodes? If these dynode voltages are fluctuating, that could cause all sorts of unwanted modulations on the tube gain.

Most people don't run PMTs with periodic sources, so they don't see these effects in the frequency domain.

Jon

G. Sauerbrey wrote a fairly comprehensive paper on photomultiplier linearit y in Applied Optics volume 11 pages 2576-83 in 1972.

It was in German. I translated it at the time, and could scan my translatio n and e-mail it to you. You probably have it - it's the only paper I recall that mentioned the non-linearity arising due to space charge between photo

-cathode and the first dynode.

Stabilising the voltage drop across the last stage of the dynode chain with a zener diode is always a good idea. The last stage - final dynode to anod e - hasn't got any gain and the current not flowing through the last resist or raises the gain across the active part of the tube. Lush published this back in 1965, but in the UK Journal of Scientific Instruments, which no Ame rican ever seems to have read.

Someone else has mentioned the usual gottcha with two LED sources - every L ED is also a photodiode (though not a very good one) so what may come out o f the second LED may vary - a bit - with the light falling on it from the f irst LED.

It worried me when I thought about replicating up Sauerbrey's expensive ill umination scheme with a pair of LEDs - not that I ever did it.

Hi Phil

Have you checked for voltage sag at the last (highest current) dynodes? Tried to stiffen them up? Maybe even run them from separate stabilized power supplies connected in series instead of the usual resistor chain?

If the frequency is not too high, could you "compensate" by driving the last one or two dynodes actively, proportionally to the anode current (sort of like the "ultralinear" transformer tap in tube amplifiers is used to drive G2, but in this case controlled and driven by fast solid state amplifiers instead)? It's an old linearization trick, but if you can get fast enough drivers, maybe worth a try...

Regards Dimitrij

For the few led's I looked at it was only linear over a small range, like 1-10 mA. low currents had weird 2/3 dependence and what I assumed were heating effects rolling it off at high currents.

George H.

I think I read in a later post that you don't see it with a photodiode as monitor... if that's the case then it has to be in the PMT, at least mostly.

George H.

Am too ignorant in that to give direct help, but stupid enough to give possible "work-around". Since they have a "clever scheme to make each line come out at a different RF frequency", have them not harmonically related (like the DTMF frequencies as used by Ma Bell). Maybe there is really no such thing as "inherently evenly spaced"; that it just looks that way..

Use our Codatron(R) High Voltage shunt regulator (drop-in replacement to the Victoreen Corotron)?

On a sunny day (Wed, 18 Nov 2015 15:35:28 -0500) it happened Phil Hobbs wrote in :

Spectral components resulting from frequency modulation _can_ appear on places where cubic products would.

Thanks. I'd need 11 of them, though, and the bias is only 60-100V per stage, so a stack of FCX596TA PNP emitter followers is better than good enough. It'll also have a couple of LND150s for current limiting, so the tube doesn't turn to lava if somebody opens the box in room light.

Cheers

Phil Hobbs

About 25 mA p-p at 20 MHz.

Cheers

Phil Hobbs

Well, I'm hitting it pretty hard, and seeing about 50% modulation with a photodiode, so it's staying in high level injection all the time.

Cheers

Phil Hobbs

Yes, but quadratic behaviour is a bit better fit than cubic, though it looks pretty linear.

Cheers

Phil Hobbs

But they're still cubic at low modulation indices.

Thanks

Phil Hobbs

It's multiple passes through the same acousto-optic cell, iirc.

Cheers

Phil Hobbs

On a sunny day (Thu, 19 Nov 2015 10:24:14 -0500) it happened Phil Hobbs wrote in :

Yea, that points to something else..

I did feed the PMT dynodes directly from the voltage multiplier stages to reduce current and stabilize the voltages there.. like here:

I would offer you help, buy my customer is competing with your customer. Want a bunch of bad ideas?

I've got enough of those on my own. ;)

Thanks for thinking of me.

Cheers

Phil Hobbs

There is a surprisingly small fraternity of people who are honest, open, competent, and helpful by nature. We need to stick together, so that more good stuff gets done.

Unsurprising that John Larkin has exploited an opportunity for self-congratulation. He probably does think that he fits the description, which shows how much the "honest" is worth.