Page 2 of 3

Re: A semi-DIY preamplifier for radiation detectors

Posted: Mon Feb 25, 2019 5:54 am
by Richard Hull
Years and years ago. Carl Willis warned all of us about neutron preamp bias and I have taken his advice ever since. I always ramp up the bias slowly from zero to about 1600 volts and likewise I ramp it down to zero before I hit the bias supply on-off switch. This saves the input FET gate from possible destruction via capacitive pulsing when snapping on full bias or snapping it off. My advice is just do it!

As an old electronics engineer I knew how sensitive the ultra high impedance electrometer grade FET gates were, but in all my experience over the years, I never had 1600 volts DC in a gate circuit between the gate separated by only a capacitor. I only encountered this in my neutron detection scheme. and tend to obey the rules.

All the electrometers I ever used were paired 5886 electrometer vacuum tube front-ended. They would take a lightning strike! I would often, when measuring large static fields, actually have a spark leap to the input.....No problem. Do that with a modern FET front end electrometer, then reach for your wallet and a few weeks at kiethley.

Richard Hull

Re: A semi-DIY preamplifier for radiation detectors

Posted: Tue Feb 26, 2019 7:34 pm
by Rich Feldman
We agree that ramping HV bias up or down is good practice. Richard and John, how slow is your slow?

Designer at Cremat just agreed with me that ramping slow enough for preamp to stay linear is unnecessarily conservative.
I tried it while watching scope, not quite pegging CR-110 output.
Took about a minute to get from 0 to 1000 volts.

If my HV blocking capacitor were 10 nF, like most, it would take 10 minutes.

Thought of a feature that would be easy to add. Sort of like a transit lock, in the suspension of record turntables and other delicate instruments.
Put a switchable resistor or external clamp diodes between preamp input and ground. With that engaged, HV bias can change at a convenient rate w/o stress on preamp. Release the charging-current bypass switch before attempting to detect things.

Back to the primitive bias power supply:
Changing rectifer to bridge reduced voltage ripple to 780 mV (p-p, at amp output, with 1000 volt bias).
Augmenting the 0.01 uF output cap with a 0.95 uF microwave oven cap (w/ integral 10M bleeder) reduced ripple to 28 mV.
Reducing the HV blocking cap, and AC-coupling the preamp output, will help more.

Maybe not enough to make up for lack of regulation. When refrigerator starts, line voltage dips, so HV bias dips, and preamp output level moves substantially. Hmm -- just discovered that you can get 4500 volt transistors at Mouser.

I think detecting neutrons will be more rewarding than scratch-building a bias supply that's up to the task.

Re: A semi-DIY preamplifier for radiation detectors

Posted: Tue Feb 26, 2019 9:45 pm
by Richard Hull
I guess It takes me just under 45 seconds to manually ramp up and down my bias supply voltage. I often feel it is just the spike that kills. Ramping up and down are just a bit of warranty that feels comfortable. Certainly, I have not killed my first preamp in use since 1999 by just ramping up and down.

Richard Hull

Re: A semi-DIY preamplifier for radiation detectors

Posted: Wed Mar 06, 2019 7:05 am
by Rich Feldman
Broke down & got me an Ortec 459 Bias Supply ($75) and a powered NIM bin to put it in ($33).
Picked 'em up this morning, and in a quick test this evening both appear to work. Except +/- 6V power at the bin slot I checked, but the 459 doesn't use that.

The turns-counting dial on voltage setting pot has more friction than it should. I will need to open the box and switch HV polarity to positive. And buy or contrive a SHV connector, and clear plenty of bench space.

Re: A semi-DIY preamplifier for radiation detectors

Posted: Wed Mar 06, 2019 9:08 am
by Richard Hull
Good deal on the NIM bin, super cheap. The bias supply is about right on the price.

By the way, virtually no NIM bin has a +/- 6 volt supply in them. All NIM bins contain a +/-6v buss in them! A decent fraction of all bins have test probe, tip jack points on the power panel on the right side of the bin that include +/-24 v, +/-12v and +/-6v outputs so that the buss voltages can be easily read, verified and adjusted with ease.

Virtually 100% of all bins demand a +/- 6 volt plug in NIM module supply. This module then supplies the already extant buss to all plugs the +/- 6 volts.
In many cases, such modules only supply one 6 volts and might have a switch to change its polarity. A second 6 volt module will be required and its polarity selected accordingly to have both + and - busses powered. All such 6 volt plug-ins are heavy as the 6 volt requirement can be a good number of amps in other NIM modules demanding one or more of the 6 volt polarities. Such supplies often take up 2 or 3 front panel slots.

6 volts in NIM modules are effectively ancient history of long ago. 6 volt power supply plug-ins are very rare and are not common on the NIM surplus market. I lucked onto a total of two 6 volt supplies now. I purchased them 6 years apart and the last one I paid $135 for.

I only need one +6 volt module in my fusor NIM setup. It is one of the demanded voltages for an LRS custom pulse matching and shaping module.

Never expect to find or obtain an empty NIM bin that has a built-in +/- 6 volt capability.

Richard Hull

Re: A semi-DIY preamplifier for radiation detectors

Posted: Wed Mar 06, 2019 12:58 pm
by Rex Allers
Reminded me of an ebay purchase about a year ago.

Was looking for an adjustable supply to use for PMTs or neutron tubes. Happened on a cheap one listed.

A Canberra 3002, 500-3000V. Listed as for parts only. Seller said they measured no output when set in the 500V range.

It is a NIM module but from pics I could see it only used 115V input. Good 'cause I don't have a bin to plug it into.

I'm usually pretty good at fixing things and it was cheap so I bid. Got it for less than $30 with shipping.

When it got here I plugged it in and measured at 500 setting -- no problem. With a higher voltage probe I measured every step from 500 to 3kV and it worked fine. I switched the polarity and that worked fine too. So fully functional. From specs on a later version manual I found, I think it is good for up to 10 mA.

Looks like this.
Would have preferred one with a meter and one multi-turn setting knob, but this will be fine, especially at the price.

A bit back onto the topic of this thread -- I noticed that when I change the setting it takes a few seconds to get to the new output value. I guess that's part of the design to help avoid blowing what might be attached with fast voltage transients.

Good thing it worked because I looked inside and everything is in a bunch of completely sealed plastic modules. I doubt if I could have fixed anything but a loose wire.

Sometimes you get lucky.

Re: A semi-DIY preamplifier for radiation detectors

Posted: Wed Mar 06, 2019 6:14 pm
by Dennis P Brown
This is an excellent example of the "Art of Bidding" that matters for getting cheap but good equipment.

Newbie's should take note!

Having the ability to do limited repairs allows one to 'gamble' on equipment and get extremely good prices; it is no accident that he 'lucked out' and it didn't need any repairs. Luck is often made just like this; if you have some knowledge, good things will often occur because you can afford to take limited risks and sometimes, they pay off completely like this. That does also require getting non-working stuff in bids, often not too difficult to fix if one gets (develops) the skills required.

Re: A semi-DIY preamplifier for radiation detectors

Posted: Sat Apr 13, 2019 9:06 pm
by Rich Feldman
OK, I think I've seen a few neutron detections at last.

With bias from the Ortec, and preamp output connected to oscilloscope, there was a discouraging amount of low-frequency wandering. Much reduced with AC-coupling mode on the scope. It seemed to get better after I left the old bias power supply (by itself) on for a few days at 1900 volts, and worked the voltage-setting dial up and down a few times. AC coupling by user, with a much shorter time constant, will probably make it a non-issue. It was handy to use the scope's continuously-crawling display mode, with horizontal at 200 to 1000 milliseconds per division.

The BF3 tube is in a pile of 1/4-inch thick HDPE scraps from Tap Plastics, along with a sheet of beryllium and part of a small household safety appliance. Digital storage scope triggers about once every minute or two, reminiscent of Carl Willis's "counts per day" reports like this one from 2005: viewtopic.php?f=13&t=5566

Occasional trigger events resulted in weird waveform shapes, but most showed the expected 140-us-tailed positive pulses. Scope automatic measurement feature indicated peak millivolt values of 32.8, 34.4, 33.6, 36.0, 33.6, 38.4, 23.2, 91.2, and 47.2 in about 10 minutes.
Nothing smaller, even with trigger level at 6.4 mV (barely above the noise). To be re-visited after noise reduction steps. Most of the gamma flux is expected to be at 60 keV, which I suspect is strongly attenuated by the metal wall of BF3 tube. Got to find something radioactive that makes higher energy gammas. Maybe the test source on yellow CD geiger counter.

Am considering a couple ideas for gathering pulse-height statistics from an unattended setup, without buying a MCA or starting to deal with PC sound cards. Could get out a self-contained, battery-powered event logger.
elusb5.jpg (2.96 KiB) Viewed 4229 times
It can record the time and sign of state changes with 0.1 second resolution. Preamp would be followed by a breadboarded pulse-stretcher circuit that might turn the 91 mV pulse into an start and stop event a few seconds apart. It might be easier to point a digital camera at oscilloscope screen, taking pictures at regular intervals or soon after each scope trigger event. Or a USB camera -- now how do we electrically activate the "capture image" key in camera monitor program? "Motion" sensitive capture mode? Other ideas are welcome!

Re: A semi-DIY preamplifier for radiation detectors

Posted: Thu Apr 18, 2019 9:18 pm
by Rich Feldman
Here's a representative pulse from preamplifier output, into 1 MΩ scope channel with 20 MHz bandwidth setting, and bias at 1500 volts.
After cutting up eight 7-inch-wide sheets of HDPE, and one smaller sheet,
the moderator pile was reassembled more scientifically. Still far short of Fermi's original Chicago piles.
There's room in the middle for beryllium and alpha sources. Some might protest that there's no moderator between the predicted neutron source and the detector tube. And that the predicted neutron source is on the wrong side of a known "neutron reflector". Let the configuration experiments be done! Any requests? First need to automate the counting of pulses per hour or per day.

A forum FAQ thread about CAD has bubbled up lately. I found a simple 2-D drawing program, MS Visio, very helpful when figuring out how to cut up the plastic sheets. There's plenty of guesswork about which factors matter. Original lot of material is sufficient to fill a 4" diameter cylinder around the detector tube.

Re: A semi-DIY preamplifier for radiation detectors

Posted: Fri Apr 19, 2019 1:15 am
by Rich Feldman
Stacking the top half of pile differently resulted in, subjectively, a dramatic increase in count rate.
I guess by a factor of at least 3. Measured 9 pulses in about 5 minutes:
28.0, 2.4, 28.0, 23.2, 27.2, 20.8, 26.4, 28.0, 3.2 millivolts.

Now there's 1/2 inch of moderator between predicted neutron source and detector.
And the detector and alpha source are on the same side of 1/16" thick sheet of beryllium.
Is any reader in a position, with the inclination, to talk about neutron transport simulation?
For example, with the MCNP mentioned from time to time?
Or to present rules of thumb, seat-of-the-pants awareness, from practical experience?