1. Home
  2. Electronics Design
  3. PMT with TIA pulse amp.

PMT with TIA pulse amp.

I had to redesign the pmt pulse amp. The polarity needed to be flipped to match with an existing instrument. The pmt is R212 from Hamamatsu and the socket/ power supply is C8991 also from Hamamatsu. The first circuit looked like this.

|\ AD8001 anode 0----+--C1--+---|+\ | | | >-+-50R-> BNC out 3k R2 +-|-/ | ohm | | |/ | | | +--R4--+ GND GND | R3 | GND C1 was 0.01 uF R2 was 300 ohms R3 was 100 ohms R4 was 1k ohm There are back to back diodes in parallel with the 3k ohm =91DC path=92 resistor on the anode.

This worked just fine, but there was some (5 to 10 mV) of 100 Meg Hz and other interference crud on the output. The circuit was operated from a split supply.

The new circuit looks like this,

+--R6---+ | | | |\ | anode 0--+--C1-R5--+--|-\ | | | >-+-50R=97-C2--+-->BNC out 1k +-|+/ | ohm | |/ 1k | | ohm GND +-2.5 VDC | | GND C3 | GND

I made it all capacitively coupled so that it could run from a single supply.

All the C=92s are 0.01 uF R5 is 100 ohms R6 is 2k ohms

R5 is used to reduce a 140kHz interfering signal that comes from the C8991 power supply. It=92s a Cockroft-Walton HV supply.

This configuration has almost no signs of the interfering crud that existed on the first incarnation.

Here are some beautiful =91scope shots. I had the persistence set to 1 second.

The first is with the trigger level set just above the noise. I then increased the PMT voltage till it started to trigger (~500V). At the

2mV scale the =91scope bandwidth is reduced to 20 MHz.

formatting link

Then the same with 5mV/div and 200MHz BW,

formatting link

700 Volts
formatting link

800

formatting link

1k
formatting link

and 1.2kV

formatting link

Will it ever look this nice in production?

George H.

Reply to
George Herold
Loading thread data ...

That's sloppy of Hamamatsu. But lower anode load resistance might help there. Or fast AC coupling, of the signal will be short pulses.

Signal level goes up as some absurd power of HV, but the 140K noise will increase roughly linearly. So crank it up!

Board layout really matters here. If the proto is a hack, a clean multilayer board will be much better.

John

Reply to
John Larkin

ohm

=A0GND

Yeah, We told Hamamatsu about it and they acknowledged the issue. Cranking up the HV works just fine. This is all pretty much over- kill, since we just do pulse counting anyway. Still, it's a pleasure to have a nice clean signal. What I found most interesting is that the TIA had less HF pickup 'crud' when compared to the previous resistor to ground and then voltage gain amplifier.

"Signal level goes up as some absurd power of HV" From my 'scope pics... and guesstimating some average signal level.

HV signal

500 5mV 600 15mV (not shown) 700 35mV 800 100mV 1000 450mV 1.2k 1.5 volts

That makes a nice log - log plot with a slope of ~6.5

George H.

Reply to
George Herold

Hamamatsu makes great detector products, but their instruments could use a bit of jazzing up. (I wouldn't mind running the Hamamatsu racing division.) ;) They've had this particular problem for at least 15 years, which is why their packaged analogue-mode PMT modules have 20 kHz lowpass filters on them, or did the last time I looked.

Those Cockroft-Walton things are really great in principle, because you can use one C-W tap per dynode, and the impedance of the taps goes up as you get towards cathode potential (-1 kV or so), so the higher-current dynodes get stiffer supplies.

The stiffness of the lower taps improves the PMT's linearity in analogue mode, which is a big help. (The other common way of doing it is with Zeners in the last couple of stages, but then you can't adjust the dynode voltages to change the gain. C-Ws adjust all the taps at once.)

My current favourite method is to use a $5 unregulated fluorescent backlight power supply plus a bridge rectifier, then use a resistor string with 10 emitter followers in series. Filtering the first couple at the cathode end gives you a capacitance multiplier, and then the rest just loaf along--the transistors' beta makes all the taps stiff. It does waste a bit of power, but not nearly as much as a low-resistance divider string. You close the loop at the cathode end via a high voltage resistor, or for lower cost, close it near the bottom of the string.

You can even use the backlight supply as a half-wave doubler circuit, because it comes with a couple of high voltage capacitors in series with its output. That gets you twice the voltage with half the diodes. If you use the bridge rectifier approach, both ends of the transformer secondary have to float. so you have to disconnect the ground. The voltage doubler approach allows the secondary to stay grounded, but may overstress the capacitors if you're not careful.

(You can run small He-Ne lasers off backlight supplies too. Good medicine.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net
Reply to
Phil Hobbs

=A0 ohm

=A0GND

In fact the big linearity problem with resistive dividers is caused by the last resistor. The stage from the last dynode to the anode doesn't generate any gain, so as this voltage drops as the anode current rises, the voltage across the gain-generating stages rises. Replacing just the last resistor with a zener kills any super-linearity and makes sure that the gain decreases with increasing anode current - usually by quite a bit less than it used to rise, but this depends on the dynode design.

I tended to put a zener across the other end of the string as well - it pays to keep the photocathode to first dynode voltage as high and stable as possible. Maximumn gain at the first stage minimises the multiplication noise, and maximum voltage across the - rather large - photocathode-to-first dynode gap minimise the electron transit time through the tube. It also doesn't mess up the electron-optics through the subsequent dynodes, which are much closer together and see much higher field gradients.

Those two zeners don't inhibit gain-adjustment by changing the total voltage across the PMT.

-- Bill Sloman,Nijmegen

Reply to
Bill Sloman

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.