Pulse generator Questions

[Tom Dressel]

I was thinking about making a simple pulse generator to debug the NIM daisy chain and to get the feel for how the NIM components work togather.

Could one use say a 555 timer chip to generate a chain of square wave pulses, then use a 741 op amp run as a differentiator to get a series of voltage spikes and a diode to filter out the negative pilses? Are the typical Radio Shack op amps fast enough to simulate the pulses from a PMT?
Looking at the spec sheet for a 353 dual op amp, it has a slew rate of 13 V/micro second.

Tom Dressel

[guest]

A 741 op amp is nowhere near fast enough to make a decent pulse in even in tne microsecond range. You would be better off using a CMOS 555 such as the TLC555 or LMC555 as your astable oscillator (faster edges than the original bipolar 555, and less noise on the power supply due to cross-conduction of the output stage). These CMOS 555s also make dandy square wave astables with a resistor from the output to trigger and threshold, with a capacitor from there to ground. The differentiated square wave could then be sent to an emitter follower buffer to beef up the drive level. What sort of pulse polarities and amplitudes wetre you looking for?

[Tom Dressel]

I am not interested in doing anyting fancy, I just want to check out the NIM components in the daisy chain. I thought that if I could get the system to work with some sloppy pulses that it wouldn't be to hard to get it to work with real pulsed from a PMT. I realize that the pulses may not be as sharp as the actual PMT pulses, but I just want to get a feel for how the NIM components work togather.

Tom Dressel

[DaveC]

Tom -
If you just want to exercise the pulse counters and pulse height bins, etc... you might think about using the calibrator output from your oscilloscope... (rashly assuming you have one)...They are usually a TTL class square wave... (50 nsec transition time...typical)

Just use the calibrator as input to your NIM hdw direct nothing to build.

Dave Cooper

[Tom Dressel]

Yes, I do have an old analoge scope. That sounds like the best way to go for generating a string of pulses.

Tom Dressel

[Richard Hull]

Ideally you should be looking for a NIM pulser as it will serve well for a myriad of purposes associated with the effort.

The advantage of the nim pulse is the magical reduction of a known and calibrated signal without the horrible noise associated with big resistor chain dividers. It is magic.

The nice features of the pulse are.

A real sharp mechanical switch closure at usually a 50-100 hz rate (unadjustable) from a vibrating mercury wetted relay. (clean and noiseless)

There is a prcision DC supply and calibrated pot dial to allow you to vary your output signal from 0-5 or 0-10v, normally.

There is a stepped ladder of precise toggle switches that will let you divide or reduce the signal by 4 decades plus by 5 and by 2 and finally by 1.414 (Square root of 2)or any combination thereof. All the time, and regardless of the switch settings, the output impedance is maintained at 93 ohms. I have checked and with a 1 volt signal dialed in and reduced by 2000 a clean, noiseless 500 uv signal spits out the BNC!! Stunningly amazing.

I bought my nim pulser (two nim width unit) off e-bay for $55.00 (worth hundreds more). At a hamfest in Northern VA, I picked up that killer box seen in image du jour. It too is a nuclear pulser by Tennelec but with many more features than can be crammed into a single or double wide nim module. I paid $5.00 for it!! Darned if four weeks later I picked up an Ortec nuclear pulser box at another hamfest for only $3.00!! I traded the Ortec box to a friend in the nuclear biz for three Ortec Nim modules, all spectroscopy amps!! That three bucks really went somewhere!!

The nuclear pulser is a very specialized item and is the rosetta stone to nuclear instrument testing and calibration at the professional techie level.

Yes, an absolute equivalent could be made with modern stuff, but I would hate to have to assemble one from scratch and retain the noise figure into the microvolt level.

Interestly, I have never seen a nuclear pulser made with all electronic components, which stuns me as a practicing electronics engineer. All still use the vibrating reed mercury switch, precision DC supply, and special toggle switch ladder divider/impedance matcher. The only reason I haven't tried to build one is I don't need one.

It turns out that frequency response is rarely an issue in the nuclear instrumentation biz. More important is the ability to supply regular pulses, noiselessly, with fast rise and fall times coupled with ultra low level signal levels. Therefore, tester/pulsers are designed to mime low input, low noise sources to test amps and delay lines etc. If you can handle the rise and fall times, pusle rep rates are a non-issue outside of coincidence problems.

I think I might throw up a couple of images of my pulsers this evening to give all a chance to few them in the flesh.

Richard Hull

Richard Hull

[DaveC]

One additional thought on using the scope calibrator...

To get a good pulse, you should use a simple differentiator network on the calibrator output - a capacitor in series with a resistor to ground. The pulse output to the NIM cct, is taken across the resistor, which should be either 50 or 75 ohm. (The resistor should match the characteristic impedance of the coax for the cleanest signal.)

If the signals have a small amount of noise, then a pulse height analysis will show a small distribution around the real signal level.

The differentiator circuit will transmit a charge pulse whose magnitude will be deltaQ = (1/C )x(delta V),
where C is capacitance in Farads, Q is the charge transferred, and delta V is the pk-pk voltage of the square wave.

For a 1 volt pk-pk calibrator signal, with C of 100 pF, the charge transferred per cycle (pulse) is
1 x 100 = 100 pc
Incidentally, this is a standard method of generating a claibrating pulse for use with partial discharge measuring circuits.

Hope this helps a bit.

Dave Cooper

[DaveC]

Whoops!!! --- NOT Delta Q = 1/C delta V !!! :

Delta Q = C delta V

The numbers are right in the example. The formula was scrambled. Apologies for the typo.

Dave Cooper