Re-imagined Baxandall photomultiplier power supply

That is a heaping pile of dog crap if I ever saw one..

I don't even know why I wasted my time copying and pasting that into Ltspice..

I can think of a much better approach to make a PS for a photomultiplier tube.

Jamie

On Sunday, 16 February 2014 02:14:32 UTC+11, Maynard A. Philbrook Jr. wrot e:

kcroft-Walton voltage doubler to reduce the number of turns on the secondar y. It turns out that Greinacher invented it a decade or so before Cockcroft and Walton independently re-invented it.

, and - with the doubler - this gives 1.63kV which is close enough to the m aximum allowed voltage - 1.65kV - for the XP2982 photomultiplier I had in m ind. In reality we never needed more than about 800V, but a really low gain tube might have needed more.

d an interlayer capacitance of about 200pF - the nominal length of a single layer would be 53mm, the width 19.4mm and I had room for 0.12mm of Mylar t ransfomer tape between each layer. Figuring on a dielectric constant for th e tape - Mylar and acrylic adhesive - of about 3 gives 200pF. Seven layers in series gives an end-to-end capacitance of about 29pF.

a lot - the inter-layer capacitance goes down by a factor of four - to 50pF - and there are four times as many layers leaving an end-to-end capacitanc e of 1.8pF.

though they are probably still around if you are prepared to buy enough of them.

by with four layers, with room for 0.5 mm of insulation between each layer , dropping the interlayer capacitance to about 50pF, and the end-to-end cap acitance to 10pF, equivalent to a resonant frequency of 65kHz. Loaded, the circuit below runs at 55kHz. A minimally banked winding on a two section f ormer - which one can still buy - could drop the end-to-end capacitance to

rlapping drives for M1 and M2. The arrangement shown works well enough for a simulation.

pply to current limit at something like 150mA. I ran a simulation with 150R in series with a 24V power supply to get a similar effect, and start-up wa s perfectly tidy.

LTSpice..

Presumably you hoped that you might learn something. Since you've still got quite a lot to learn, this would have been a reasonable idea, if you'd had any hope of making any sense of it.

Do show us.

Baxandall is reputed to have invented the Class-D oscillator as a means of taming the inconvenient inter-winding capacitance of the multi-layer second aries in the inverters used for stepping up 12V supplies to voltages high e nough to drive photomultipliers.

This doesn't show up in his 1959 paper

but a guy who did his apprenticeship under Baxandall's supervision at the R oyal Radar Establishment at Malvern (England) told me that that had been th e problem that Baxandall been trying to solve when he (Baxandall) did the w ork

It's not a particularly cheap solution, but it can be quite efficient.

Jim Williams popularised it in his - very popular - series of application n otes on driving cold-cathode back-lighting lamps, Linear Technology's AN45, AN49, AN51, AN55, AN61, and AN65.

We wait for your "much better" solution with bated breath.

Yeah, that's awful. 6 inductors, one with 5 windings! 19 zillion total parts!

These days, people usually use a C-W stack and hang dynodes on the multiplier stages. That makes more sense than using a resistive divider off a single HV supply.

Depends what you're trying to do. The C-W approach is good for photon counting but noisy in analogue applications. Also it's easier to do things like collapsing the supply to save the tube if somebody opens the box at the wrong time.

Cheers

Phil Hobbs

That would be noisy. Here's how one of my medical electronics clients does it...

...Jim Thompson

Something upside down?

My LtSpice did a readable schematic, but it did not get anything out.

Sure, that's the way it was done 70 years or so ago.

You don't care about the supply noise in photon counting. The low multiplication factor at the first dynode makes the pulse heights vary a lot--you might average 5 secondaries, but the 1-sigma limits are from 3 to 7, and that carries through all the way to the anode.

You can do a bit better by using a stack of NPNs with the resistor string setting the base bias. Costs more than resistors, but then the tube is probably $1k.

Cheers

Phil Hobbs

Paraphrasing Longshanks in the movie, Braveheart, who is this prick, Larkin, who dares to speak in my presence ?>:-}

And who can't read the date on the schematic, presented as the client's method... not mine. ...Jim Thompson

ckcroft-Walton voltage doubler to reduce the number of turns on the seconda ry.

ft and Walton independently re-invented it. The secondary has 540 turns whe re the original version had a 1000 turns, and - with the doubler - this giv es 1.63kV which is close enough to the maximum allowed voltage - 1.65kV - f or the XP2982 photomultiplier I had in mind. In reality we never needed mor e than about 800V, but a really low gain tube might have needed more.

ad an interlayer capacitance of about 200pF - the nominal length of a singl e layer would be 53mm, the width 19.4mm and I had room for 0.12mm of Mylar transfomer tape between each layer. Figuring on a dielectric constant for t he tape - Mylar and acrylic adhesive - of about 3 gives 200pF. Seven layers in series gives an end-to-end capacitance of about 29pF.

a lot - the inter-layer capacitance goes down by a factor of four - to 50p F - and there are four times as many layers leaving an end-to-end capacitan ce of 1.8pF.

, though they are probably still around if you are prepared to buy enough o f them.

t by with four layers, with room for 0.5 mm of insulation between each laye r, dropping the interlayer capacitance to about 50pF, and the end-to-end ca pacitance to 10pF, equivalent to a resonant frequency of 65kHz. Loaded, the circuit below runs at 55kHz. A minimally banked winding on a two section former - which one can still buy - could drop the end-to-end capacitance to 2.5pF.

erlapping drives for M1 and M2. The arrangement shown works well enough for a simulation.

upply to current limit at something like 150mA. I ran a simulation with 150 R in series with a 24V power supply to get a similar effect, and start-up w as perfectly tidy.

rts!

One transformer, two off-shelf inductors and three ferrite beads. Two of th e windings on the transformer are just there to provide a circuit that's ea sy to simulate in LTSpice - in real life I'd probably use a digital scheme to get non-overlapping operation of M1 and M2.

lier

HV

As was mentioned in the previous thread. People who are serious about photo

-multipliers don't have equal voltage divisions down the whole dynode chain . The voltage between the photocathode and the first dynode should be fixed a t the maximum allowable, and the voltage between the last dynode and ground should again be fixed at a value that's sufficient to avoid space charge b etween the last dynode and the anode. The voltage across this stage doesn't affect the photomultiplier gain - there's no useful electron multiplicatio n at the anode.

There's also the point that there's synchronous ripple on the output of a C

-W stack, which modulates the gain of the tube.

It's not a solution that makes sense in every situation, and it does requir e "zillions of parts" most of them running at higher voltages than you want to stick directly on a printed circuit board. Most of the components in my last divider chain were mounted on Teflon-insulated pins.

This seems to be the tube

They weren't around 70 years ago. the dynode chain does seem to incorporate some of the extra features that I talked about.

Oh, I never expected that it was original with you. You pretty much do as you're told.

On Sunday, 16 February 2014 06:06:33 UTC+11, Tauno Voipio wrote:

It worked for me, but start-up is horrible, and it took a while before anyt hing showed up on the output screen.

This starts up better - but in a rather odd mode of operation. After 2.5mse c it starts looking more like what was intended. Using a digital scheme to drive M1 and M2 looks better and better.

Version 4 SHEET 1 1992 864 WIRE -992 -288 -1072 -288 WIRE -832 -288 -912 -288 WIRE -720 -288 -832 -288 WIRE -656 -288 -720 -288 WIRE -320 -288 -656 -288 WIRE 304 -288 -320 -288 WIRE 304 -240 304 -288 WIRE 944 -224 896 -224 WIRE 1040 -224 1008 -224 WIRE 1120 -224 1040 -224 WIRE 1296 -224 1184 -224 WIRE -64 -128 -112 -128 WIRE 96 -128 16 -128 WIRE 608 -128 96 -128 WIRE 720 -128 688 -128 WIRE 1040 -112 1040 -224 WIRE 896 -80 896 -224 WIRE -720 -64 -720 -288 WIRE -672 -64 -720 -64 WIRE -560 -64 -608 -64 WIRE -416 -64 -480 -64 WIRE 64 -64 -416 -64 WIRE 96 -64 96 -128 WIRE 96 -64 64 -64 WIRE 896 32 896 -16 WIRE 1040 32 1040 -48 WIRE 1152 32 1040 32 WIRE 1296 32 1296 -224 WIRE 1296 32 1216 32 WIRE -416 48 -416 -64 WIRE -112 48 -112 -128 WIRE -32 48 -112 48 WIRE 48 48 -32 48 WIRE 608 48 112 48 WIRE 720 48 720 -128 WIRE 720 48 608 48 WIRE 1296 80 1296 32 WIRE 208 128 80 128 WIRE 304 128 304 -160 WIRE 304 128 272 128 WIRE 304 144 304 128 WIRE 368 144 304 144 WIRE 512 144 432 144 WIRE -416 160 -416 128 WIRE -112 176 -112 48 WIRE 720 176 720 48 WIRE -720 192 -720 -64 WIRE -832 208 -832 -288 WIRE 1296 208 1296 160 WIRE -32 256 -64 256 WIRE 80 256 80 128 WIRE 80 256 48 256 WIRE 128 256 80 256 WIRE 304 256 304 144 WIRE 304 256 208 256 WIRE 320 256 304 256 WIRE 384 256 320 256 WIRE 512 256 512 144 WIRE 512 256 464 256 WIRE 560 256 512 256 WIRE 672 256 640 256 WIRE -416 288 -416 224 WIRE 320 304 320 256 WIRE 464 304 320 304 WIRE -320 336 -320 -288 WIRE -320 336 -352 336 WIRE 320 336 320 304 WIRE -1072 368 -1072 -288 WIRE 1216 384 1120 384 WIRE 1296 384 1296 288 WIRE 1296 384 1216 384 WIRE -416 400 -416 384 WIRE -416 400 -576 400 WIRE -576 416 -576 400 WIRE -416 416 -416 400 WIRE 320 448 320 416 WIRE 320 448 144 448 WIRE 144 480 144 448 WIRE 464 496 464 304 WIRE 320 512 320 448 WIRE 1296 512 1296 384 WIRE -416 528 -416 496 WIRE 80 528 -416 528 WIRE 1120 528 1120 384 WIRE -416 560 -416 528 WIRE -1072 672 -1072 448 WIRE -832 672 -832 272 WIRE -832 672 -1072 672 WIRE -720 672 -720 256 WIRE -720 672 -832 672 WIRE -576 672 -576 480 WIRE -576 672 -720 672 WIRE -416 672 -416 640 WIRE -416 672 -576 672 WIRE -112 672 -112 272 WIRE -112 672 -416 672 WIRE 144 672 144 576 WIRE 144 672 -112 672 WIRE 320 672 320 592 WIRE 320 672 144 672 WIRE 464 672 464 560 WIRE 464 672 320 672 WIRE 720 672 720 272 WIRE 720 672 464 672 WIRE 896 672 896 112 WIRE 896 672 720 672 WIRE 1040 672 1040 32 WIRE 1040 672 896 672 WIRE 1120 672 1120 592 WIRE 1120 672 1040 672 WIRE 1296 672 1296 592 WIRE 1296 672 1120 672 WIRE 1424 672 1296 672 WIRE -1072 704 -1072 672 FLAG -1072 704 0 FLAG -32 48 tank- FLAG 608 48 tank+ FLAG 64 -64 ct FLAG -656 -288 +12V FLAG 1216 384 Out SYMBOL ind2 -80 -112 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 4 56 VBottom 2 SYMATTR InstName L1 SYMATTR Value 0.22m SYMATTR Type ind SYMATTR SpiceLine Rser=0.022 SYMBOL ind2 592 -112 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 4 56 VBottom 2 SYMATTR InstName L2 SYMATTR Value 0.22m SYMATTR Type ind SYMATTR SpiceLine Rser=0.022 SYMBOL cap 112 32 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 46 32 VTop 2 SYMATTR InstName C1 SYMATTR Value 19p SYMBOL ind2 -576 -48 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName L10 SYMATTR Value 1m SYMATTR SpiceLine Rser=1.8 Cpar=63p SYMATTR Type ind SYMBOL voltage -1072 352 R0 WINDOW 123 0 0 Left 2 WINDOW 39 24 132 Left 2 SYMATTR SpiceLine Rser=0.1 SYMATTR InstName V1 SYMATTR Value 13.5 SYMBOL ind2 912 128 R180 WINDOW 0 36 80 Left 2 WINDOW 3 36 40 Left 2 SYMATTR InstName L5 SYMATTR Value 595m SYMATTR Type ind SYMATTR SpiceLine Rser=64.5 Cpar=10pF SYMBOL res 1280 496 R0 SYMATTR InstName R3 SYMATTR Value 3300k SYMBOL FerriteBead -640 -64 R90 WINDOW 0 -16 0 VBottom 2 SYMATTR InstName L6

SYMATTR SpiceLine Ipk=0.2 Rser=1 Rpar=785 Cpar=535f mfg="Wurth El ektronik eiSos" pn="742 792 79" SYMBOL res 1280 64 R0 SYMATTR InstName R1 SYMATTR Value 2.2k SYMBOL cap 1104 528 R0 SYMATTR InstName C2 SYMATTR Value 10n SYMATTR SpiceLine V=3k SYMBOL ind2 480 272 M270 WINDOW 0 32 56 VTop 2 WINDOW 3 4 56 VBottom 2 SYMATTR InstName L4 SYMATTR Value 0.108m SYMATTR Type ind SYMATTR SpiceLine Rser=0.02 SYMBOL ind2 224 272 M270 WINDOW 0 32 56 VTop 2 WINDOW 3 4 56 VBottom 2 SYMATTR InstName L3 SYMATTR Value 0.108m SYMATTR Type ind SYMATTR SpiceLine Rser=0.020 SYMBOL res 288 -256 R0 SYMATTR InstName R2 SYMATTR Value 10k SYMBOL nmos 672 176 R0 SYMATTR InstName M1 SYMATTR Value AP9465GEM SYMBOL nmos -64 176 M0 SYMATTR InstName M2 SYMATTR Value AP9465GEM SYMBOL res 304 496 R0 SYMATTR InstName R4 SYMATTR Value 1.3k SYMBOL res -432 32 R0 SYMATTR InstName R5 SYMATTR Value 8.2k SYMBOL diode -432 160 R0 SYMATTR InstName D5 SYMATTR Value 1N4148 SYMBOL pnp -352 384 R180 SYMATTR InstName Q1 SYMATTR Value 2N3906 SYMBOL res -432 400 R0 SYMATTR InstName R6 SYMATTR Value 100k SYMBOL cap -592 416 R0 SYMATTR InstName C3 SYMATTR Value 10n SYMATTR SpiceLine V=3k SYMBOL res -432 544 R0 SYMATTR InstName R7 SYMATTR Value 22k SYMBOL npn 80 480 R0 SYMATTR InstName Q2 SYMATTR Value 2N3904 SYMBOL cap 448 496 R0 SYMATTR InstName C4 SYMATTR Value 100n SYMBOL res 304 320 R0 SYMATTR InstName R8 SYMATTR Value 680 SYMBOL FerriteBead 896 -48 R180 SYMATTR InstName L7

SYMATTR SpiceLine Ipk=0.2 Rser=1 Rpar=785 Cpar=535f mfg="Wurth El ektronik eiSos" pn="742 792 79" SYMBOL ind 1280 192 R0 SYMATTR InstName L8 SYMATTR Value 47m SYMATTR SpiceLine Rser=52 Cpar=37.5p SYMBOL cap 432 128 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName C5 SYMATTR Value 25p SYMBOL cap 1008 -240 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName C6 SYMATTR Value 10n SYMBOL diode 1184 -240 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName D1 SYMBOL cap 1152 48 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName C7 SYMATTR Value 10n SYMBOL diode 1024 -112 R0 SYMATTR InstName D2 SYMBOL res -896 -304 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R11 SYMATTR Value 15 SYMBOL zener -816 272 R180 WINDOW 0 24 64 Left 2 WINDOW 3 24 0 Left 2 SYMATTR InstName D4 SYMATTR Value BZX84C12L SYMATTR Description Diode SYMATTR Type diode SYMBOL cap -736 192 R0 SYMATTR InstName C9 SYMATTR Value 220n SYMBOL res 64 240 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R9 SYMATTR Value 68 SYMBOL res 656 240 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R10 SYMATTR Value 68 SYMBOL cap 272 112 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName C8 SYMATTR Value 25p TEXT -680 744 Left 2 !.tran 0 40m 0m 10n TEXT -680 776 Left 2 !.ic I(L4)=-0.00 I(L5)=-0.000 I(L3)=-0.00003 I(l

PALwHUB... It was the client's circuit... they asked me to behavioral model it, which I did.

You are one sick puppy.

Hobbs said your method was noisy... until he noticed I said likewise. It's a shame he's become your sycophant. ...Jim Thompson

Everything is noisy. If the ripple of a c-w chain is higher than you like, make the caps bigger, or add a second layer of RC filtering. Quad r-packs and quad c-packs can be had for a few cents each.

The resistive string wastes power and regulates the last few dynodes poorly. It's OK in some places.

Do you have direct experience working with PMTs? When I was a kid, a 931A cost $1.50 from Fair Radio Sales, so I did all sorts of stuff with them. Scintillation, lidar, interferance experiments, listening to light. I'm looking at a big Philips jug on my bookshelf here... beautiful pice of glass.

Jim, you may have run out of bullets, but that's your problem. Leave me out of your stupid flame wars.

Phil Hobbs

Don't stick your nose where it doesn't belong. ...Jim Thompson

Right. It's not as if this is a public forum or anything like that.

I didn't try to rent you for some stupid enthusiasm of mine, dude.

Cheers

Phil Hobbs