Needing more than a spark test?

[homebrewed]

This is a point on which I looked hard for what is known. I know when the photon arrives, it might miss. Most of what is inside of what we think is solid hard stuff is mostly empty space. There is a probability of collision involved. If it misses, there is more to hit beyond. If the stuff is thin enough, it can go right through. It's higher energy X-rays stuff (Gamma), after all!

If it hits, it excites the element electrons into a higher energy state. They don't stay that way. They drop back into their "normal" state. The amount they took in to get to the excited state is released as a fluorescence photon, having a new wavelength determined by the energy change involved, and Planck's constant. It comes out as X-Rays.

So - what happens to the excess?
A 60KeV photon hitting iron (Fe) uses up only 6.4KeV and 7.6KeV to excite the K-shell, and presumably, at the same time, uses up another 705eV and 718eV getting the L-shell electrons into a higher state. That total is 15.4KeV, leaving another 44.57KeV yet to do anything.

Does the leftover energy simply excite the same atom electrons another couple of times until it can't quite manage a last K-shell event?
Does the remainder go on to keep working the same atom L-shell up and down again and again until even that runs out?
The wavelengths are getting pretty long for the small energies. Would that be into infra-red?
Does it end up shaking the atom about somewhat, as in it "warms the stuff up" a bit?

Does it not happen that way at all? Does the remainder 44.57KeV keep going to strike some other atom instead?
------------------------
All of the above is about the fluorescence scintillation. The next bit is about what happens in the photodiode detector.

Although I have trawled many videos about what happens in photodiodes, particularly from the advanced set of lectures from a Indian university, I have not seen a clear explanation of exactly how a arriving photon turns into a current amid the conduction band carriers in a material. All the equations are about a energy flux of lots of photons. There is an efficiency involved. We do not get a current of energy equal to 100% of the incoming. Some of it, I think, ends up as heat.

In our design, we have a transimpedance amplifier capable of seeing a current started by only one photon, which is then amplified.
A whole bunch of other noise currents will be amplified along with it. Some of it is thermal generated, but I don't propose cryogenic amplifier design. Some is induced from outside fields, but we can shield it from such interference by design. We even can deny magnetic interference, and we can get up to circuit design tricks that can cancel differential noise, and avoid common mode noise. In the end, this will be about signal to noise ratio, and the only way to preserve that, locked in, is to start with extreme low noise amplification with very high gain, so that any noise from later stages is dwarfed by comparison to our (amplified) original noise.

I am pretty sure that we will run into pulse and real signal situations we did not anticipate, meaning the difference between theory and practice. We just have to give it our best shot. You are doing all the right things to anticipate most everything. The aluminium plates I had first thought was overkill, but I revise my opinion.

You have some good comments here. I'm thinking that the incident photons do experience a sequential drop in energy as they excite fluorescence in atoms they hit and also pump carriers into the conduction band that eventually contribute to the current pulse height. The fluorescence x-rays also do their bit to add to the current pulse, but since they can't excite any more fluorescence x-rays all they can do is move carriers up into the conduction band. So in a fast cascade the incident photon is converted into back-radiated fluorescence photons, hole-electron pairs and heat (phonons). For detection efficiency we mostly want hole-electron pairs.

The energy needed to get a carrier into the conduction band of doped silicon is pretty low, down in the single electron-volt range so it should be possible to suck most of the energy out of any spare x-ray photons flying around the crystal lattice.

 

[graham-xrf]

@homebrewed - Thanks much for some more physics insight. I tried to take in as much as I could from my "YouTube education" binge, but there came a point where I overloaded.

Pulse pattern recognition
Then I started to get practical. A simple discrimination strategy is to note the pulse duration characteristic from the PIN diode in recovering and recombination. For the X100-7 salvage diode, the largest energy pulses decay back in about 10-mS to 13mS. A new (good) pulse from the same element, arriving as late as 9mS will make an "extended" duration total pulse, but will have a near-repeated amplitude. The whole smeared pulse would be about 20mS. Maybe several pulses would make a very long smear, but we count from a peak.

The next case is where we have a second pulse arriving within the 13mS window. It might be a smaller one. It smears the pulse to (say) 20mS or more. We still see one good measurable peak.

So what if the next one within the window happens to be a bigger pulse? This is actually OK. It ruins the measure of the first pulse that started it all. It also has it's own measure ruined, because we get a summed new peak looking like something impossible, or maybe unfortunately spoofing a false positive on a third element. If it happens to land so exactly timed coincidental to make a false pulse, then it will be rejected for being stupidly too high, or at worst, will falsely increment the bucket for a third element, but so rarely it will not make a significant showing.

Therefore, I think we might get a quite good result using a ridiculously crude and simple discrimination filter. A pulse qualifies if it happens within a 13mS window, or perhaps more subtly, within a 10mS window after the peak.

Add to that, if a pulse amplitude is beyond a maximum for the highest energy pulse from Tungsten (W -> 59.3KeV), plus a little bit, then it is a combination coincidence, so gets rejected. There are, of course, some combinations of lower energy pulses that can spoof a higher energy pulse, but they would also have to happen within 10mS of a single peak. Most would get rejected, and the coincidence ones would happen very rarely, so have low bucket counts.

The filter also rejects various kinds of interference noise. The low side threshold would be set to exclude noise, and the s/n noise ratio we hope to get such that a 220eV pulse is still seen. Such a discriminator is a brutal simple thing, and it might extend the "gather samples" period, but it might also perform as well as, or better than a ton of software smarts in pulse pattern recognition, software compensations, and other good stuff.

It may also just be wishful thinking on my part.
For this reason, I urge you to continue your smart software approach.

 

[rwm]

I like it Graham.

 

[homebrewed]

I get it: "it should be as simple as possible, but no simpler" . It makes for an elegant design and I'm all in favor of that!

 

[rwm]

Can you do a primitive mock up and see what the pulses look like in terms of pulse rate, decay rate, number of random summed events etc? Could you look at this with an oscilloscope and get any idea?

 

[homebrewed]

Can you do a primitive mock up and see what the pulses look like in terms of pulse rate, decay rate, number of random summed events etc? Could you look at this with an oscilloscope and get any idea?

Acquiring some raw data will be the first step for me. I will sample the output with nothing in the sample chamber (other than the unavoidable aluminum enclosure) and get some basic statistics -- exactly what you're talking about. Aluminum's XRF signal is at about 1.5Kev, 'way down in the detector's sensitivity range, so I expect mostly background stuff.

I also have a number of pure elemental samples I can use for calibration and reality-check purposes. That's where I expect to see pulse overlap issues arise. Thanksgiving activities have delayed things a bit, but I _think_ I'm pretty close to firing the whole thing up in an EMI-free environment. Basically, wiring up a connector and double-checking it so I don't let any of the magic smoke out.

I'm hoping the 12-bit A/D on my Teensy4.0 will be adequate for the job. It maxes out at about 2MSPS, although at that sample rate its ENOB is closer to 10 bits. The pulses coming out of my signal conditioning circuit are pretty slow so I may be able to do some averaging to increase the effective number of bits (that's what ENOB stands for).

 

[homebrewed]

For anyone who is thinking about going down the Teensy route for an XRF system, my additional investigations into ADC performance vs. Teensy versions indicate that Teensy 3.6 may be a better choice. Even though its system clock is slower, its ADC resolution and sample rate are better than 4.0 or 4.1. 3.6 has a built-in FPU just like 4.x so doing stuff like polynomial fits will still be pretty speedy. For some reason the 3.6 is a bit more expensive than 4.0 (but still less than $30USD).

Since I have a Teensy 4.0 I will go ahead and see how well it works. Since I'm using the Arduino IDE it would be easy enough to recompile my code for a 3.6 if I find it necessary to use one.

To keep from overdriving the Teensy's ADC I've decided I need to add a diode clamp to the output of my signal conditioning board. My board can output up to +10V, well above the Teensy's 3.3V max. I'll tap into the Teensy's 3.3V supply for the clamp -- it's brought out to a pin so no need for yet another power supply or voltage regulator.

 

[homebrewed]

I finally had some time to fire up my XRF setup to get a preliminary idea of how it's working. Nothing special, just the 'scope looking at pulses. It appears that my enclosure is keeping 60Hz noise out just fine, but I discovered the switching regulator on the PocketGeiger is injecting a fair amount of noise into the transconductance amplifier (or perhaps into my signal conditioning board, which is nearby). This may have been another reason for having that copper shield on the board -- the 10x10mm detector appears to be a great antenna, in addition to being an x-ray detector.

The noise is not due to an oscillation problem in my signal conditioning board, because it goes away when I turn the pocket geiger off.

Some experimentation is needed to see if I can put some shielding around the switcher/inductor, or if I just need to bite the bullet and use another external supply to run the detector board. I threw the copper shielding away but I have some copper sheet I can cut and form. The main thing is to avoid shorting anything out, while not covering the detector.

One potentially big advantage of using an external supply is the option of easily varying the bias voltage on the detector. The detector diode's dark current is a substantial noise contribution, and it strongly depends on the bias voltage. This alone is enough to make me lean toward an external power supply.

This is one of the reasons I've been calling my setup a "test bench" -- it's to identify and wring out all the issues that invariably come up in a project like this.

 

[graham-xrf]

Hi Mark - and good on ya! Great that you are trying out the diode. I have been reviving ambitions about it, and two nights ago I was sorting out what the good lady calls "all that radioactive crap" For a little longer, XRF is still has to be the project behind a few other things, but even now, every bit of unknown metal I see is crying out for it!

Re: the switcher inductor noise
You can Faraday shield it without creating a shorted turn, by using some Kapton tape in a foil capacitive overlap. Pot-core type inductors keep all the local magnetic fields trapped in the high permeability ferrite that surrounds everything, but even a small one of those is probably too big as a single component for that size board.

Even if the switcher inductor were perfectly shielded by a deity, that it exists on the same circuit board as the transconductance amplifier means that all the ways it can pollute the signal, beyond by radiative (near fields) across the airspace between it and conductors affected, are all still present. It injects it's waveform into every PCB trace, especially the common-mode return.

When trying for low noise amplification at the kind of gain we are using, layout more like the style used in electrometers or pH meters is probably appropriate, except this time, we throw in the need for reasonable bandwidth as well. I think that if the voltage must be created with any kind of switcher, then first, it gets banished from the board. It can be software-settable, but however it is made it should be filtered, and post-switcher final regulated, where the regulator has sufficient bandwidth to follow the pulsations, and regulate them away. Fortunately, this can be done with one low noise op-amp. The trip to the amplifier can be by screened twisted pair, with the screen grounded at only one end.

Then, before it gives any life to the transconductance circuit, add a common-mode filter (tiny toroid), and a balanced filter (two more tiny toroids + capacitor), and a final storage capacitor at the amplifier board. If you get the amplifier signal clean enough, just measuring it may need some care. I would use two probes to a scope, in differential setup, with channels summed, one inverted, and the gains balanced to null the signal of the scope test square wave.

 

[homebrewed]

Hi Mark - and good on ya! Great that you are trying out the diode. I have been reviving ambitions about it, and two nights ago I was sorting out what the good lady calls "all that radioactive crap" For a little longer, XRF is still has to be the project behind a few other things, but even now, every bit of unknown metal I see is crying out for it!

Re: the switcher inductor noise
You can Faraday shield it without creating a shorted turn, by using some Kapton tape in a foil capacitive overlap. Pot-core type inductors keep all the local magnetic fields trapped in the high permeability ferrite that surrounds everything, but even a small one of those is probably too big as a single component for that size board.

Even if the switcher inductor were perfectly shielded by a deity, that it exists on the same circuit board as the transconductance amplifier means that all the ways it can pollute the signal, beyond by radiative (near fields) across the airspace between it and conductors affected, are all still present. It injects it's waveform into every PCB trace, especially the common-mode return.

When trying for low noise amplification at the kind of gain we are using, layout more like the style used in electrometers or pH meters is probably appropriate, except this time, we throw in the need for reasonable bandwidth as well. I think that if the voltage must be created with any kind of switcher, then first, it gets banished from the board. It can be software-settable, but however it is made it should be filtered, and post-switcher final regulated, where the regulator has sufficient bandwidth to follow the pulsations, and regulate them away. Fortunately, this can be done with one low noise op-amp. The trip to the amplifier can be by screened twisted pair, with the screen grounded at only one end.

Then, before it gives any life to the transconductance circuit, add a common-mode filter (tiny toroid), and a balanced filter (two more tiny toroids + capacitor), and a final storage capacitor at the amplifier board. If you get the amplifier signal clean enough, just measuring it may need some care. I would use two probes to a scope, in differential setup, with channels summed, one inverted, and the gains balanced to null the signal of the scope test square wave.

I don't think the switcher's inductor is a toroid or pot core. It looks like a plain old solenoid style, which should be great at broadcasting its mag field. A thick-enough layer of copper can attenuate that (but enough???). I also note that the designers of the PocketGeiger put a lot of bypass caps near the chip, prompting me to suspect that they ran into noise problems due to switcher transient currents on the Vcc trace.

It's likely that the PocketGeiger will get a pair of external power supplies but I will at least explore the possibility of adding shielding to get rid of the switching noise.

Right now it looks to me like most of the circuit noise is coming from the switcher rather than the diode dark current. So it's well worth the effort to address this.

Oh, and BTW, Happy Holidays!

Mark