Page 2 of 3

Re: Not Quite As Simple CSA

Posted: Sun Mar 28, 2010 3:08 pm
by Tyler Christensen
I'm getting 50mV pulses out of the 3He tube and several volt pulses out of the 10B tube, although that one is a direct function of the bias voltage so it has little meaning.

This is with the 2pF charge storage capacitor.

This is well above the 2-5mV noise floor

Re: Not Quite As Simple CSA

Posted: Sun Mar 28, 2010 7:27 pm
by Richard Hester
This suggests that you're are getting about 100 nC/event from your He3 tube at your chosen bias voltage. This is a plausible value. It also suggests that a charge storage capacitor of 1-2 pF will be a good starting value for other experimenters. It also sounds like a post amp and shaper located inside the same housing as the preamp would be useful. 50 mV is a usable signal - 0.5 to1V would be better.

Re: Not Quite As Simple CSA

Posted: Sun Mar 28, 2010 11:09 pm
by Richard Hester
Here is my first PSpice tinkering run for a post-amp and shaping circuit to go along with the preamp. Some of the components are in place merely as probe points, and would not be included in an actual design. The first 2N5457 and the 2N3906 constitute the post amp, and the other two 2N5457s are the shaper. Since the shaper only has two poles, its output doesn't come near an ideal Gaussian pulse, but the peak is broadened sufficiently to make discrimination easier, at the cost of a good deal of amplitude, which will be seen in the next post. The blue probe monitors the input pulse, the red is the output of the postamp, and the green probe monitors the shaper output.

Re: Not Quite As Simple CSA

Posted: Sun Mar 28, 2010 11:17 pm
by Richard Hester
The picture here shows the input pulse (blue), the postamp output (red), and the shaper output (green) from the PSpice simulation. As can be seen, the shaper smooths and stretches the pulse, at the cost of a good deal of amplitude.This is very similar behavior.to the kind of circuits I've been using for my scintillator-based All-In-One detector setups.

Re: Not Quite As Simple CSA

Posted: Fri Apr 02, 2010 8:50 pm
by Tyler Christensen
My build of this circuit just stopped working, I'm now getting 0.6V P-P noise straight out of the pre-amp. It just suddenly started doing this the other day. At first it was just a few loud peaks then it went into chaos. I've replaced all three active transistors and it behaves exactly the same. Checked all the resistors, none are burned up. Any guess on what's causing this? This is with the HV off.

UPDATE: Solved it, it was the input diodes (just leaving the message incase anyone has the same problem after driving too much voltage into the input). I didn't initial replace them because they test fine on my diode test meter, interesting.

Re: Not Quite As Simple CSA

Posted: Fri Apr 02, 2010 11:37 pm
by Richard Hester
The diodes are there to catch a bullet (like switching on the HV to the detector with the preamp connected) that would normally take out the input FET first thing. Bad diodes also happen, and they will inject noise right into the input. The diodes might have been degraded enough to get noisy in the reverse direction. This may not necessarily show up on a simple diode tester. It would be interesting to put the original FET back into the circuit to see if it survived whatever took out the diodes.

Re: Not Quite As Simple CSA

Posted: Thu Sep 23, 2010 3:15 am
by Richard Hester
I was looking at the postamp/shaper circuit I posted earlier on in this thread, and may have a solution for getting the signal amplitude higher while still keeping in the range needed to operate with a 9V battery. This is an offshoot of work I'm doing with jfet-based phono amps I'm documenting on a DIY audio site. More on that tomorrow when I can get to my simulator results. If I could get 0.5-1V output from a simple 2-pole pulse shaper things would be very happy-making.
A circuit for optically coupling the preamp output to a remote SCA would also be neat, but it may be more expedient to keep the SCA local to the detector housing and optically couple the resulting go/no-go pulse to a remote counter. This might require a pair of 9V batteries rather than a single, but that's still not a biggie.

Re: Not Quite As Simple CSA

Posted: Thu Sep 23, 2010 6:49 am
by Starfire
Richard - I'm getting good results with a 555 using the trigger as input and set up for pulse o/p.

Re: Not Quite As Simple CSA

Posted: Mon Apr 25, 2011 12:10 am
by Richard Hester
Attached is my second stab at a discrete postamp and shaper circuit to tie in with the charge sensitive amp. I relaxed my design constraints a bit and used a different fet for the input, which got me the gain I needed. As before, it runs off of a 9V battery.

Re: Not Quite As Simple CSA

Posted: Mon Apr 25, 2011 12:13 am
by Richard Hester
Attached are the input and output waveforms. From a teensy ~50mV input, you get a solid 500mV pulse with a rounded top that will make it easier for a discriminator to work without needing a blindingly fast comparator.

Re: Not Quite As Simple CSA

Posted: Mon Apr 25, 2011 12:49 am
by Richard Hester
Here's a simpler version of the postamp/shaper that works the same as the previous (same response).

Re: Not Quite As Simple CSA

Posted: Thu Apr 28, 2011 4:08 pm
by Richard Hull
Once again, nice amps and thanks for sharing. These with the usual input and bias circuitry added ought to do OK for a 3He counter tube.

Richard Hull

Re: Not Quite As Simple CSA

Posted: Thu Apr 28, 2011 11:42 pm
by Richard Hester
Keep in mind that the circuits I just posted are a postamp and shaper to take the output of a charge-sensitive preamp such as the one cited at the beginning of this tread, amplify the output, and shape the response to ease amplitude discrimination.They're not too useful without the preamp ahead of them. They are particularly germane to this thread, as the preamp that started the thread was aimed at He3 proportional tubes where the output pulse is fairly tiny, even with a small value for charge storage cap (~1pF) for the charge sensitive amplifier. The output of the charge sensitive amp is about ~50mV, so the postamp and shaper can make life much easier for a discriminator or SCA.

Re: Not Quite As Simple CSA

Posted: Fri Apr 29, 2011 9:54 am
by Richard Hull
I am sorry, I was commenting on the entire thread. Sometimes I assume too much of the reader, being an electronics guy, myself. The original "CSA" linked to your last or next to the last "postamp" circuit is what I was referring to. This combo should replace most any older NIM based preamp setup and can be attached directly to the tube head, itself in a small box, utilizing batteries or the common 9 pin nim plug using the 12 volt line with a regulator.

I like attaching the preamp/postamp right to the head of the tube. It is a real noise reduction system if well executed.

I didn't look at the FET and transistor limits but I'll bet NIM 12volts could be used with only minor alterations.

Richard Hull

Re: Not Quite As Simple CSA

Posted: Fri Apr 29, 2011 1:32 pm
by Richard Hester
If you really wanted to get nifty, the next step would be to gin up a low power discriminator circuit that you would adjust once to chop away all the gamma grass (assuming your HV to the tube is stable and repeatable). This would drive an LED coupled to an optical fiber, going to a remote counter. Having all the analog stuff in one box at the detector would go a long way toward making the detection system more noise immune.The optical fiber is perhaps gilding the lily, but it wouldn't be all that hard to do. Someone here (John Hendron, I think) mentioned using a 555 timer in the context of a discriminator. I'd vote for clever use of a CMOS 555 - I'll do some thinking.

I'll look at my simulations to see what 12V will do to the biasing in both circuits. It may not be much of an issue.

Re: Not Quite As Simple CSA

Posted: Fri Apr 29, 2011 4:33 pm
by Richard Hull
I figured once you had a 2 volt signal in hand, standard coax and good termination would obviate most of the noise in all but the harshest environments. However, it would, indeed be cool to run the signal thru fiber, especially if you are working a high pulsed energy fusion system.

I figured that 12 volts, NIM, might need some tweeking on the biasing networks. I imagine a linear regulator (LM317T) set for 9 volts would be a good, no bother, solution. Only a super "greenie-weenie" would worry about the 3V disappation losses with the few milliamp demand of the circuit.

Good work as always Richard....and thanks.

Richard Hull

Re: Not Quite As Simple CSA

Posted: Fri Jun 10, 2011 4:58 pm
by Richard Hester
Attached is a picture of a board incorporating variants of the charge sensitive amplifier and postamp/shaper circuits presented here, plus a simple discriminator/cable driver circuit constructed using an LM393 dual comparator. I've done the initial power-up and tweaking for DC operating conditions, and will try to mate this up with a detector system. Since I don't have any neutron tubes and sources configured at this time, my first step will be to feed this bad boy with the output of an NaI(Tl) scintillator with the PMT run at reduced voltage to bring the gain and pulse amplitude down.

My goal is to include the board inside of a detector assembly using one of the Russian corona tubes. I haven't figured out whether I want to use the B-10 or He3 corona tube to start with. If I interperet DC's results from his testing of a B10 tube correctly, I won't need to use a postamp with the B10 tube, so I would feed the output of the charge sensitive amp right into the shaper circuit. This may be eventually what happens with the NaI scintillator system two, as I will likely get a good, strong ~1V output from the charge sensitive preamp even with the PMT crippled by a low supply voltage. The post-amp will come in useful for people using He3 and B10 proportional tubes with their lower amplitude output.

Anyway, this one is still in the development phase, and I'll post schematics when I have a better idea of what's happening. The board shown is about 4" by 1 3/4", so it can very handily be mounted on the underside of a gallon paint can lid.

Re: Not Quite As Simple CSA

Posted: Fri Jun 10, 2011 6:13 pm
by Tyler Christensen
Regarding not using a post-amp, worked fine for me... my amp/shaper broke and I just fed right into my SCA directly from the CSA pre-amp box and it worked great for BF3, B10, He3, and Geiger... probably not "optimal" but I got the same count rates I'd expect on all detectors, so I didn't see any issues with it.

Re: Not Quite As Simple CSA

Posted: Fri Jun 10, 2011 6:31 pm
by Richard Hester
Maybe so in your case, but on this board, the discriminator is included. Without the post-amp, the threshold pot for the discriminator would have to be cranked all the way down to the low end, with out any real latitude for adjustment.

When all the homework is done, this should be a setup that woulld allow me to feed a detector head directly into a counter.

Re: Not Quite As Simple CSA

Posted: Mon Jun 13, 2011 2:48 pm
by Richard Hull
We have grown used to Richard's all-in-one solutions and the wait is worth while.

Richard Hull

Re: Not Quite As Simple CSA

Posted: Wed Jul 13, 2011 6:42 pm
by Richard Hester
I found my LED-Photodiode pulser in a box last night (I'd forgotten that I'd put in a nice case with a switch and output connector and all). At any rate, I hooked it up to the preamp/shaper/discriminator board today, and things worked first time around. I'll post some waveforms when I get a chance to take some pictures. When I find my little Power Designs HV supply, the next thing will be to hook it up to an NaI(Tl) scintillator with the PMT operating at reduced voltage.

Re: Not Quite As Simple CSA

Posted: Wed Jul 13, 2011 8:33 pm
by Richard Hester
For something to chew on, the next two posts feature the schematic for the preamp/shaper/discriminator board. The schematic here shows the charge sensitive amplifier, followed by a post amplifier (useful with He3 and B10-lined proportional tubes), and a shaper that takes the sharp-tipped exponential waveform from the preamp/postamp and rounds it off into a waveform more easily handled by a discriminator.

Re: Not Quite As Simple CSA

Posted: Wed Jul 13, 2011 8:40 pm
by Richard Hester
The next schematic shows the discriminator, constructed using an LM393 dual comparitor. One half of the comparator is used to compare the input waveform with a voltage reference generated using a TL431 shunt regulator. The pot is used to adjust the trip threshold of the input comparator. The second comparator of the 393 is configured as a monostable multivibrator, which puts out a fixed ~5usec pulse when the input comparator fires. A couple of transistors are used to buffer the output of the monostagle, so that one can drive a long cable if necessary. I plan to mount one of these boards right inside a neutron detector head. The output would directly drive a counter.

Re: Not Quite As Simple CSA

Posted: Thu Jul 14, 2011 4:43 pm
by Richard Hester
Attached is a digital oscilloscope picture of the waveforms from the charge senstive amp, postamp, and shaper portions of the circuit, driven by a led/photodiode pulser circuit. The top waveform is the output of the charge sensitve amplifier (probed at the junction of C8 and R17) at 50mV/div. The peak value of the signal is about 35mV. The center trace shows the signal at the output of the post-amplifier (probed at the junction of R10 and R14) at 500mV/div. It has an amplitude of ~1.7V. Finally, the shaper output (probed at C7) is shown on the bottom trace, at 500mV/div. The shaping process has reduced the signal amplitude to ~0.7V. This is the signal that gets fed to the discriminator circuit.

Re: Not Quite As Simple CSA

Posted: Thu Jul 14, 2011 4:57 pm
by Richard Hester
This next oscilloscope picture shows the input signal to the discriminator comparator (middle trace, pin 4 of U2A), the DC level that his circuit is compared to (top trace, probed at pin 3 of U2a), and the resulting output pulse (bottom trace, probed at OUT), The "OUT" signal is used to drive a counter.The perturbations on the pin 3 trace should not be there. These will get fixed, and a new schematic and waveform set shown. Having said that, the circuit is still functioning as intended.