Magnetic field sensitivity of PMT

There will be thick aluminum but the socket area will naturally be very exposed because the circuit board has to plug in there. And then this optics stuff usually has the dreaded black anodization for a cool look where electrical contact will be anyone's guess.

Yup. The black anodizing on the threads helps a lot with light-tightness, though--there are only a couple of turns' worth on lots of gear.

The anodizing on the case and heatsink of Mini Circuits connectorized power amps led me on a merry EMI chase one time when I was a grad student. I replaced all the cables, re-routed everything, and there was still this huge feedthrough of the acousto-optic cell's drive getting into my photodiode.

I finally took apart the ZHL-1 amp and dremelled off the anodizing so I had a continuous shield. The feedthrough immediately vanished.

Cheers

Phil Hobbs

Joerg,

You should read PMT_handbook_v3aE.pdf chapter 5.4 Housing

Best regards, Habib.

That's because optics guys don't use thread dope which could stain their shiny lab coats :-)

My main tool during EMC hunts is a Dremel. Sometimes when things are really messed up an industrial strength angle grinder. I also often carry a really old pair of jeans and my 25 year old lumberjack shirt. "Can you provide me with a work space as far away as possible from everyone else, no carpet and with an exit door close by?" ... "Why? ... "You'll notice very soon".

Found it at Hamamatsu, can be ordered. But that takes too long, I have to start the actual design work today because of back-to-back assignments. Else I won't be able to wrap up in time. The circuitry is more or less done but, as John Larkin says, 50% of engineering is packaging. I'd say it's more like 70%. Finding parts, making CAD models, writing documentation, arranging stuff, cabling, connectors, all that grunt work.

If your SMPS switching frequency is a multiple of your A/D sampling frequency (I'm assuming everything turns to bits at some point), won't that at least turn the interaction into a constant error?

Not really because there is a fairly long travel time from cathode to anode, via lots of dynodes. Like a flight with ten stop-overs. The ADC has to be much faster than that and its sampling rate will generally be way higher than the converter frequency.

Hard to do when the signal is 20-50 MHz though, and it's a multiplicative interaction rather than straight pickup. (Pickup can just be filtered out, since Joerg's using a sine wave converter.)

Cheers

Phil Hobbs

And the classic RCA PMT manual from 1970 can be found here:

Not dogeared, held together with tape and coffee stained like my paper one.

Thanks! I like the "optional price" statement on the cover.

Some of mine have beer stains, too.

Dang. Well, I guess ignorance didn't work...

ome

or

that

Transit times depends on the tube and the interdynode voltages, but a ball- park figure for the electron velocity is around 10% of the speed of light, and it increases in proportion to the square root of the voltage.

My fast focussed tube has a roughly 20nsec transit time. In that project I was worrying about electron transit times in the electron microscope as wel l - and there was as much from the blanking plates in the column down to th e target integrated circuit, and rather more from the target back up to the scintillator.

I seriously under-estimated it at the early design stage, which was odd - m y rather earlier patent on a wide-voltage range electron beam blanking set- up had been all about electron transit times. My guess is that it was a sid e effect of an unrealistic schedule, which involved getting the specificati on written as fast as possible, which I'd complained about at the time, to no effect.

There was a UK slogan from the period - "we never get the time to do the jo b right, so we always have to find the time to do it again".

rs

at

be

I always thought "burning a little time with reading is far from a wasted t ime" -- Yoda

;-) Good Luck Joerg, Habib.

It's not as fast as that. Secondary electrons are emitted at low energy, so they have to accelerate nearly from rest at each stage. The mass of an electron (mc**2) is 511 keV, so even at 100V per stage (mv**2/2 = 100eV), the electron arrives at

V= c * sqrt(2 * 100eV/511 keV) ~= 0.02 c.

And since s = a t**2/2, the transit time per stage is sqrt(2) times longer than you'd calculate from the peak velocity.

So the average speed of electrons in a PMT is only 1% to 1.5% of c.

Cheers

Phil Hobbs

ll-park figure

so they have to accelerate nearly from rest at each stage.

Some secondaries are emitted at low energies. Electron microscopes sometime s differentiated between secondaries that were kicked out with an appreciab le fraction of the incident electron's energy, and thermalised secondaries which came out with only a couple of electron volts. The electron beam test ers that I worked on were interested only in the low energy seconary electr ons, because only they told you anything about the voltage at the emitting surface.

ger than you'd calculate from the peak velocity.

If the field is uniform. The point about fast-focussed tubes is that it isn 't (whence the focussing).

The smallish fast focussed tubes that I knew well had a ticket transit time of about 20nsec, but a fair bit of that was in the photocathode to first d ynode space, where we did keep the voltage drop to the first dynode as high as the spec sheet let us.

You might get close to 10% in an RCA 8850, with the GaP secondary emitter o n the first dynode, and 800V between photocathode and first dynode.

Thanks!

I've decided to split the board into a quiet one at the tube and a noisy one in the basement.

Absolutely true but sometimes there's schedule pressure.

Thanks. Will need it when it comes to magnetics such as the CCFL transformer because those things are skimpily documented.

Have you tried the Guinness Blond? Tasty beer, and the stains won't be as dark as from regular Guinness.