Advice Needed...PMT

[jack]

I'm working on a project now where I want to observe dark current
pulses on a PMT. The output from a PMT can be considered a almost pure
current source ( I've read) and I'm trying to figure out the best
electronics to put between the anode and my scope. I'm configuring the
tube as pulse mode to be used with a scintillator,so I have it as
cathode grounded, and for starters I have a coupling capacitor just
after the anode output. OK, several questions...
Do I need a preamp or should I go straight into an op amp and if so
what op amp is best? From my meager knowledge i have read a
transimpedence op amp may be what I'm looking for as it takes a
current and converts to voltage, and has the advantage of high slew
rates which I like as these current pulses are in the 5 to 20 ns
range. Before I go further with a scintillator,I would like to
understand the nature of noise in these tubes and would like to build
the electronics to do this. I have read the Hamamatsu handbook on this
but in my opinion they did not give as much attention to pulse mode
electronics as they did analogue operation. Does anyone have any
experience with this or any suggestions? Any help would be
appreciated. jack

 

[jack]

What is the gain of your tube? Which photmultiplier are you using? If
you can't identify the tube, at least tell us the number and type of
dynodes involved.

Your answer is rather different for a 14-stage fast focussed tube and
a ten-stage box and grid or venetian blind tube.The former produce
single photon pulses of several mA with a full width at half maximum
of around 2nsec, while the ten-stage tubes with the simpler dynode
structures have much less gain and spread the pulse out over more than
ten times the time.

Yes,sorry. I have a 10 stage linear focus tube with a rated gain of
..6x10^6 . It is a electron tube inc tube model9250B ,a 52 mm front
end plate. So my initial question is the amplitude and width of the
dark curent pulses. The rated dark current is 1.5 nA but I assume this
is not the same as the dark current pulses. The pulse rise time is 4
ns and the pulse fwhm is 6.5 ns. While not saying so directly I assume
these are SER times. Just to observe the dark pulses,could I just run
the anode line (after decoupling ) into my scope? Thanks for any help.
jack

 

[jack]

Thanks, now I'm going to show you how new I am too all this. What is
AoE and where can I get this article? Thaks for the reply. jack

 

[Michael A. Terrell]

Thanks, now I'm going to show you how new I am too all this. What is
AoE and where can I get this article? Thaks for the reply. jack

AoE is "The Art of Electronics" by Horowitz and (Win)Hill. Check
Amazon.com, used book dealers, or a good library.

 

[Eric Inazaki]

Perhaps related to this discussion (or not).

Some people I work with are setting up an ion
counter using an electron multiplier. Typically,
I've seen the EM connected to the preamp through a
high voltage capacitor. Is there any reason not
to directly couple the EM to the preamp, provided
that end of the EM is at ground potential? Also
what are the pros and cons of directly coupling vs.
capacitively coupling the EM? This is a topic of
some debate at work right now with each scheme
having an advocate. The current plan is to have a
shoot-off between the two designs but I'd like to
see if there are any informed opinions out there.

On an unrelated note, what ever happened to that rumor
about an AoE 4th ed.?

Thanks,
eric

 

[Watson A.Name - Watt Sun]

AoE is "The Art of Electronics" by Horowitz and (Win)Hill. Check
Amazon.com, used book dealers, or a good library.

You can probably beat spamazon's prices with the URL in my .sig,
below, or with www.abebooks.com.

--
@@F@r@o@m@@O@r@a@n@g@e@@C@o@u@n@t@y@,@@C@a@l@,@@w@h@e@r@e@@
###Got a Question about ELECTRONICS? Check HERE First:###
http://users.pandora.be/educypedia/electronics/databank.htm
My email address is whitelisted. *All* email sent to it
goes directly to the trash unless you add NOSPAM in the
Subject: line with other stuff. alondra101 <at> hotmail.com
Don't be ripped off by the big book dealers. Go to the URL
that will give you a choice and save you money(up to half).
http://www.everybookstore.com You'll be glad you did!
Just when you thought you had all this figured out, the gov't
changed it: http://physics.nist.gov/cuu/Units/binary.html
@@t@h@e@@a@f@f@l@u@e@n@t@@m@e@e@t@@t@h@e@@E@f@f@l@u@e@n@t@@

 

[Sir Charles W. Shults III]

Personally, I would prefer the isolation of the high voltage capacitor. For
one thing, it is very easy, even with a ground reference potential in the
system, to create extremely high voltages due to static charges.
This will manifest itself as discharge events that spike the counter at
random, indicating the presence of ions that are not there. Presently, as each
ion passes your detector head, it triggers a cascade of electrons in a manner
almost identical to that used in photomultipliers. The resulting signal is a
varying voltage with spikes representing individual events, and if a large
number of ions passes simultaneously, a barrage of events will produce a nearly
continuous DC signal when the system saturates.
The capacitor's size should be determined by the need to respond quickly to
any single event, but also to allow enough "head room" so that it will not
saturate. A DC coupled system would eliminate the saturation problem, but the
real solution is to make a set of three or four "filters"- all AC coupled, all
with varying sizes of capacitors, each with a Schottky rectifier directly after
each capacitor. All the filters would be wired to the same input- that voltage
created by the electron multiplier. Each of their outputs would be buffered.
All the buffered outputs would represent some part of the even spectrum, and
many would overlap- but that will be okay.
Now, sum the events with a simple op amp circuit and you get fast response
and head room at the same time. And, simultaneous signals step right on top of
each other, "erasing" themselves. A simple second stage AC amplifier will clean
the signal up and give you an accurate representation of the events, even if
they are coming very rapidly.
Just a thought.

Cheers!

Chip Shults
My robotics, space and CGI web page - http://home.cfl.rr.com/aichip

 

[Bill Sloman]

Yes,sorry. I have a 10 stage linear focus tube with a rated gain of
.6x10^6 . It is a electron tube inc tube model9250B ,a 52 mm front
end plate. So my initial question is the amplitude and width of the
dark current pulses. The rated dark current is 1.5 nA but I assume this
is not the same as the dark current pulses. The pulse rise time is 4
ns and the pulse fwhm is 6.5 ns. While not saying so directly I assume
these are SER times. Just to observe the dark pulses,could I just run
the anode line (after decoupling ) into my scope? Thanks for any help.
jack

That would be the Thorn-EMI Electron Tubes 9250B tube - unless there
has been a management buyout in the last few years. I worked at EMI
from 1976 to 1979 - not long before they got taken over by Thorn - on
medical ultrasound. I did have some contact with Electron Tubes on
photomultiplier non-linearity, but I don't know if the German paper on
the subject I sent over to them (with my translation) had any
immediate effect. The section on photomultiplier non-linearity in the
Thorn-EMI data book I've got is pretty sensible, but who knows where
it came from.

The simple-minded way of looking at dark current is to assume that it
all comes from the photocathode as thermionic emission - with the red
insensitive bialkali photocathode in the 9250B, cosmic rays and
potassium-40 decay in the glass of the photocathode are probably
nearly as important.

The typical dark current of 0.1nA (your 1.5nA is worst case for four
of the the other tubes of that family, but the 9250 is listed as 1.0nA
worst case in my catalogue)at the anode would be 1.7x10^-16 amps at
the photocathode, or about 1,000 photoelectrons per second, while the
catalogue gives 300 counts per second as typical for all three
bialkali photocathodes.

IIRR some of the dark current comes from delayed secondary electron
emission from the dynode surfaces, most of it from the later dynodes,
whose surfaces see many more electrons than the early dynodes. If you
do a pulse height analysis on the dark current, this proportion of the
dark current shows up as long tail of low amplitude events. Because
these events are small and numerous, the shot noise they contribute is
negligible.

If you want to look at the important - single photo-electron - dark
current events, you are looking at 1.6x10^-19 coulombs of charge,
multiplied by a gain of 6x10^5, 9.6x10^-14 coulombs, spread over
6.5nsec, a current of about 15uA.

You will only see that 6.5nsec pulse width if the tube is driving inot
a 50R load, so that current becomes about 0.8mV, probably too small to
see clearly with a fast scope (you'd want about 100MHz of bandwidth).

There are integrated circuits around that can offer a gain of ten or
twenty at around that sort of bandwidth. When I was working with these
sorts of pulses we used Comlinear and Analog Devices current feedback
op amps, which did the job nicely, but the parts I used then (late
1980's) are now obsolete.

The National Semiconductor LMH6624 voltage feedback amplifier is only
stable when used as a gain-of-ten (or more) amplifier, but it looks as
if it would do the job nicely, and offers rather fewer opportunities
for disaster than Win's circuit in "The Art of Electronics" ISBN
0-521-37095-7, which is in any event getting a bit dated these days -
it would be fun to re-do it with surface mount wideband transistors
(Farnell stocks PNP parts with bandwidths up to 5GHz and NPN parts up
to 10GHz, and Siemens has datasheets for 50GHz parts).

John Larkin will probably recommend a cheaper and faster amplifier -
he seems to be designing for these sorts of bandwidths at the moment,
and doing nicely out of it.

 

[jack]

This has been extremely helpful, thank you.

The simple-minded way of looking at dark current is to assume that it
all comes from the photocathode as thermionic emission - with the red
insensitive bialkali photocathode in the 9250B, cosmic rays and
potassium-40 decay in the glass of the photocathode are probably
nearly as important.

By assuming its mostly thermionic emission off the cathode,I guess
that sets this contribution as the upper limit in terms of pulse
amplitude .

The typical dark current of 0.1nA (your 1.5nA is worst case for four
of the the other tubes of that family, but the 9250 is listed as 1.0nA
worst case in my catalogue)at the anode would be 1.7x10^-16 amps at
the photocathode, or about 1,000 photoelectrons per second, while the
catalogue gives 300 counts per second as typical for all three
bialkali photocathodes.

This is the part I was fuzzy on,i.e. how these numbers are
calculated,.and it seems it was easier than I thought. It makes sense
though. I'm not sure why the dark counts disagree so much ,however.

IIRR some of the dark current comes from delayed secondary electron
emission from the dynode surfaces, most of it from the later dynodes,
whose surfaces see many more electrons than the early dynodes. If you
do a pulse height analysis on the dark current, this proportion of the
dark current shows up as long tail of low amplitude events. Because
these events are small and numerous, the shot noise they contribute is
negligible.

If you want to look at the important - single photo-electron - dark
current events, you are looking at 1.6x10^-19 coulombs of charge,
multiplied by a gain of 6x10^5, 9.6x10^-14 coulombs, spread over
6.5nsec, a current of about 15uA.

Makes sense..

You will only see that 6.5nsec pulse width if the tube is driving inot
a 50R load, so that current becomes about 0.8mV, probably too small to
see clearly with a fast scope (you'd want about 100MHz of bandwidth).

So for a cathode grounded scheme, and a decoupling capacitor,a 100 Mhz
scope should do the job.

There are integrated circuits around that can offer a gain of ten or
twenty at around that sort of bandwidth. When I was working with these
sorts of pulses we used Comlinear and Analog Devices current feedback
op amps, which did the job nicely, but the parts I used then (late
1980's) are now obsolete.

Analog make a current feedback op amp the 8011 with a very high slew
rate. If I can figure out the proper way to wire it as a
transimpedence amplifier....

The National Semiconductor LMH6624 voltage feedback amplifier is only
stable when used as a gain-of-ten (or more) amplifier, but it looks as
if it would do the job nicely, and offers rather fewer opportunities
for disaster than Win's circuit in "The Art of Electronics" ISBN
0-521-37095-7, which is in any event getting a bit dated these days -
it would be fun to re-do it with surface mount wideband transistors
(Farnell stocks PNP parts with bandwidths up to 5GHz and NPN parts up
to 10GHz, and Siemens has datasheets for 50GHz parts).

John Larkin will probably recommend a cheaper and faster amplifier -
he seems to be designing for these sorts of bandwidths at the moment,
and doing nicely out of it.

I'm still hunting down this book,should be interesting. Again thanks
loads jack

 

[Eric Inazaki]

Eric Inazaki wrote...

4th ed, we're still working on the 3rd!

Thanks,
- Win

LIke that's any excuse?

Seriously, I should have said 3d ed.

 

[Bill Sloman]

Analog make a current feedback op amp the 8011 with a very high slew
rate. If I can figure out the proper way to wire it as a
transimpedance amplifier...

You don't want to wire it as a transimpedance amplifier. The anode
structure in fast linear-focussed photomultipliers seems to be
designed to be terminaed with a 50R resistor, and you seem to get
better results by amplifying the voltage developed across a 50R anode
load than by making the anode a virtual ground.

This was not what I expected when I started fooling around with fast
tubes, and might not be true for someone with a really good feel for
complex impedances, but it was certainly true for a couple of
reasonably competent electronic engineers at Cambridge Instruments
some 15 years ago.

My rationalisation of this result was that the anode structure doesn't
just have 8-10pF of capacitance to ground (some of it via capacitative
coupling to the last dynodes) but also a significant inductance, and
you need the 50R to get a well-damped RLC circuit, but we never played
around enough to get a really good feel for what was going on.

Bill Sloman, Nijmegen