Needing more than a spark test?

Well, noise could have a big impact on the energy resolution, since each pulse must be processed as a separate event -- there is no way to average pulses to improve the SNR, because you don't know the photon energy beforehand. That's why I'm hoping that a least-squares fitting scheme will improve things. However, starting out with the best SNR possible can't hurt.
About Noise
It is a serious limitation. For going after these (low) energy X-Rays directly from silicon, without a scintillator, the whole thing can fall flat on it's face without we pay attention to noise.

Incoming noise
Without full shielding, can't work in the face of the common electromagnetic racket in the average room, even without someone putting switching currents into the wall AC mains by using a welder. So OK, we know it needs screening, and by design, made to not respond to all the EMC threats, but we mean down to a new level way below what might be OK for average electronic kit. So good EMC immunity.

We are using the Am241 gamma to provoke a little racket of our own that we are interested in, and it must not be contaminated by "other stuff". I think it can only work if the gadget uses really good shielding, such that all of the path from the Am241 to the test material, and into the sensor sees only lead, and the test material. One mode uses lead sheet with the sample on it, and XRF à la HM gets plonked down over it. The other mode uses a smaller shield pushed onto the end, and it is pressed up against the test material. With a little care, it can still work even if the lead shield is not complete, say for example putting it up to a rod too long to be in a kind of "lead lined test box".

Amplifier device noise
With just my first forays into trying various op-amps in the circuit, and also exploring the circuits with discrete FET front-ends, I see that for "large" signals, given enough gain to get near 4V for the ADC, in the bad scenario noisy cases, we get enough "noise" of various kinds to have about 80dB SNR.

The lower noise circuits can get the racket down to below 160uVp-p at output, and I even have one at less that 90uV. That best one would have SNR at 93dB. I just threw in an example 80nA as the "biggest" signal, and I use 500pA to 5nA as the smallest signals. I do suspect the real signals might be as little as a third of that.

If we consider a "smallest" signal might from a 500pA signal, which pushes 440pA into the amp, and delivers 26mV at the output. Will that be seen above the noise racket? [The spice simulation does not show a transient analysis simulation of white noise added]!

500pA sig.png

The answer is "it might be", but perhaps at the cost of limiting the bandwidth, slowing it down, having a big phase delay. Also doing something about the shot noise, adding in noise filter stages. Ultimately, living with the "pulse" ending up as a much reduced, stretched thing with a loose approximation to the amplitude of the original, and an area under that has a very smudged relationship to the energy. This is where we get a threshold region below which we have the bumps from pulse that get going before the previous has finally died.

Here is where we come up against the noise figure of the amplifier. We have to deal with both voltage and current noise specs for the device (it seems there are two things accountable).
For most selection tables, "low noise" is anything <10nV/√Hz. Those "best" TIAs mostly have about 6nV/√Hz.
The (expensive) LTC6269 has 4.3nV/√Hz and 5.5fA/√Hz.
The "ultra low noise" FETs have specs a NF noise figure 1dB, which I would know as 75K
Low noise RF pHEMTs I used to work with have noise figures 0.4dB (28K). I would like that without 14dB of gain at 2GHz!

TIA3-noise.png

When we take a circuit that uses one of those low cost FETs, like about £0.54 each, or £4.30 for ten of them, we can have 0.9fV/√Hz, and lost a whole lot of shot noise.

Shot noise and the like.
I have not (yet) done much with the elaborate possible noise sources I posted earlier. The op-amps internal .MODEL have a noise contributor, as does all the resistors. Simply exploiting that we have no interest in slow signals down to DC, and AC coupling, loses noise. Then, if after the main sgain stage, we use a band-pass filter to start at (say) 100Hz, and drop off at (say) 200kHz, a huge amount of white noise is lost, as well as 50/60Hz remnants.

Real hardware
Everything about this so far is sloppy approximate - just about good enough to head for the main aim of deciding the basic amp with some gain, and hooking up a real circuit, and measuring some noise. I am not even using the right currents yet.
The integration of that 500pA current pulse over 10uS is 109.19pA x 10uS = 1.0919E15 Coulombs.
Dividing by the electron charge e = 1.602176634E-19 says this involved 6815 electrons.
That many, distributed over the 230pF would try to lift the voltage to 4.74uV at the input.
It would of course fail. Some would be lost in the 40Meg leak.
Pretty much all the rest would try to alter the state of the coupling capacitor, and fail again as the TIA zeros it out with the feedback, generating the voltage to make this happen at the amp stage output.

Where I get unglued is that the event that started this was a photon that could only have released one electron-hole pair when it smacked the sensor. I don't understand the physics process by which the thousands of electrons to make this current happen, came about. The pulse we get comes from a single photon arrival, not thousands in a bunch.

Maybe some experts here can set out what happens, but it may involve some extended lectures from Prof. M. R. Shenoy, Dept of Physics, Indian Institute of Technology, Delhi. About 46 lectures each 42min to 58 min long. The be-all and end all of photodiode semiconductor tech.
--> Photodiode Physics Course

Do I know what I am doing?
I think what actually happens is the incoming photon hits, and expends 4.15eV in liberating the electron, and the remaining energy is put into giving it KE kinetic energy. That speedy electron will ping around and (maybe) slam into atoms, trying to go right through, but in some way, liberating another electron, which ends up that bit slower. I don't know if the approach is correct, but I thought to find out how many work functions gets expended in this. Starting with (say) a 9.6keV photon from zinc, divided by 4.15eV silicon work functions, that might generate 2313 electrons.

2313 electrons is 3.7E-16 Coulombs.
If it delivered over 8uS, the average current would be 46pA.
Given the pulse shape we had, it's peak might be at 187pA.

Compared to my 500pA guess, the 26mV pulse might be nearer 11mV. It could be contaminated with 120uV of noise, making the numbers sampled jiggle about some unless we filter noise, or let computer averages do it for us. The very lowest energies might have us trying to sample 3mV pulses.

We are in the right ball-park!
Meantime, I do not have to understand the physics down to the last electron to make a better amp. So just working back from what the LSB from 16 bits can give, starting with a reasonable Vref, I get to the values I simulate with. Basically trying for minimum signals that sit up at least 10dB to 15dB above the noise at the input. We can do no better, and we have enough gain to see some volts at the output we can count, with still a dynamic range 80dB to 93dB. I expect it will work at least as well, and likely a good deal better than all those you and I have seen so far.
 
Last edited:
@homebrewed:
Hi Mark. I did not really want to invest in a T962 IR reflow oven, partly because before even using it, you need to take it apart, replace paper-based tape with Kapton, fix ground connection, add open source cold junction compensation improvement, and re-flash the software for something open source that actually works, including key debounce. UK eBay £126.00 from Czech Republic, though it started in China.
That said, with free postage - it might be worth improving.

T962 Reflow Oven1.png

YT is full of videos of various projects to "improve" it.
The eBay.com link with prices in dollars seems to range £299.00 + $34 shipping, to $662.00 + $100 shipping :eek 2:.. What! ??
--> eBay T962A Reflow Oven

I did not miss that the "cheap" one is called T962, and the $300 one is called T962A
That's like = 1 lathe + DP + extra 4-jaw !

Electric Iron?
I do have some rescued PID controllers, thermocouples etc, but I thought it may be enough to get an upturned electric iron set to "low", and pre-heat the sensor board a bit, to say about 80C, then crank it up to something high, and just keep some slight tension on the diode using tweezers or something. Then, as the solder melts, just lift it off, and let it cool in air. It would amount to a "custom" temperature profile, but I thought it might be all it needs.

Maybe a temperature settable hot air gun just applied to the underside until the sensor lets go?
A lighter-fluid micro-torch might also work?
 
I was thinking more along the lines of modifying something like this but it sounds like it may not be just a matter of replacing the temperature controller. A "real" reflow oven would be nice but it would take a consortium of hobbyists to justify getting one.

Our lab had a home-made hot plate that used two soldering iron heater elements installed in an aluminum block. The temperature controller was thermocouple-based. We didn't use it for soldering circuit boards, we used it to heat epoxy-molded IC's up to where the plastic turned soft so we could extract the die from the package without damaging it (much). It easily got up to solder-melting temperature. It was very reliable -- we used it for over 30 years with the same heating elements and controller.

You can get temperature-controlled heat guns for not much money these days. I got one from HF that seems to work OK (although I haven't tried using it to assemble/dissasemble PCBs). Its controller puts out all kinds of nasty EMI though -- when I turn it on all the LED-based lights on that circuit start flickering like crazy!!

We eventually bought a small PCB rework station that also used hot air; and for some difficult jobs that needed a blast of heat from both sides of the board, we used a small butane soldering gun to heat the top side. It had a catalyst grid in it so no flame came out the end, just very hot air.
 
OK - given that only the sensor is of any interest, I can use my Aldi "workzone" digital settable hot air gun, set to (say) 300C+, and just cook it from the underside. As the solder melts, off it comes, and cools in air.

I do like your toaster idea, because I have the suitable PID controller and some N-Type thermocouples. Infra-red bulbs are not a lot different to toaster spring-spiral wires hung in air, so capable of quick response to a temperature profile recipe. Not the same as the thermal delay from heater elements wound on porcelain.

One needs a fan to come on to accelerate cooling, or even just to be wasting some heat all the time, so the controller can keep the temperature on the profile. Last time I meddled with this stuff, though on a bigger scale, I needed a thermocouple in the middle of the heater elements, and another coupled to the workpiece. Operated in cascade, with one control loop feeding the other, to allow the workpiece to "catch up" without putting a in huge pulse of stored energy, and avoid the elements basically at maximum during the heat-up. Cascade control is a bit OTT for this stuff.

Hot air gun then - but then we come to the new situation of having to get the new electronics built. I am thinking that gets done with soldering iron, and maybe no components smaller than 0805. The mind has just imagined this with 1206 :)
I can manage 0603 with a bit of a struggle. Anything 0402 requires I buy extras to cover the "oops-ping - poof - and it's never seen again" situation!
 
I use 0603's but that's under a stereomicroscope and using SMT flat-tipped tweezers to handle them. Yesterday I finally got around to stuffing my signal-conditioning board and managed to do it with losing just one of the 0603-format resistors. I made the mistake of grabbing the resistor first and then fumbling around for the soldering iron, and squeezed the tweezers too hard. Under the force the tips became non-parallel and the resistor vanished with a barely-audible "snap". No Harry Potter-style apparating required, just ham-fisted me :rolleyes: .

When I ordered the SMT resistors for the board I stupidly ordered some 0201's for some of the resistors. THOSE are about impossible to handle, not to mention they don't even come close to matching a 0603 footprint. Fortunately I remembered I had an assortment book of 0603's I'd gotten for some pre-retirement work projects so I could go ahead with the assembly process.

One of my work-based designs required 0201's in order to minimize their parasitic capacitance. It was for an active probe to measure RF signals on the internal nodes of IC's running upwards of 2.4GHz and I needed as low a parallel capacitance as possible. I used a Mini-Circuits HEMT based amplifier for the gain. Our technician who assembled that board really complained about the small size of those resistors! I had initially thought that 1206's would have lower parallel capacitance due to the wider spacing between their terminals, but our circuit-modelling group did a FE analysis for me and showed that the the capacitance was lowest for 0201's. Hence her pain.

Back to the signal conditioning board -- I haven't fired it up yet. Hopefully this weekend.
 
@homebrewed Hi Mark
Re: ADCs data transfer - awkward choices
I have been juggling ADCs for a whole now and one of the considerations is the data interface.
The full parallel interface would be a big bunch of wires , but no speed issues. The lower cost ADC only uses SPI serial.

There are, of course, other things that affect choice. Trying for only 20 samples out of a 10uS time for a XRF pulse to rise and decay requires 2Msps.
Then for 16-bits in a surface mount package from Analog/Linear, yields 7 devices with prices ranging £18.65 up to £41.49 (all Mouser).

So - if one chooses (say) a AD4001 (lowest cost), which will only use SPI.
2Msps requires a 70MHz SCK clock.
A Raspberry Pi 4 can, in theory, manage 125MHz

I don't much fancy the chances with 4 loose wires in a bundle, and I don't know much about high speed SPI over a range more than the size of a PCB. I have done RS485, LVDS, CANbus, etc. This may be OK on a transmission line trace across a PCB. Somewhat more awkward to get it between the measure head and the display, unless the Pi or whatever lives in the same box as the sensor.
 
Yup - you have the same space-time anomaly like at my place. The one that moves about somewhere between near left foot, to out front past right shoulder. Anything tweezer-spring powered passing that zone is gone forever, never to be seen again!
Do tell about your signal conditioning board.

I have started ordering proto 2 ADC's, op-amps, and various bits. The lowest cost option will require an external reference, but has the differential drivers built in, saving more op-amps. Actually building stuff right now is awkward, and a bit chaotic, but that's mostly my fault. With tree stumps gone, I have been spending time outside with my new eBay Huepar laser level kit, getting the positions figured out, and some profiles in place. It will need a tracked vehicle digger, hard core, steel, etc.. I think the USA expression is "pulled the trigger on it". Essentially, I called in the same construction firm that I have used before, and their head guy will be visiting this week. Until that outhouse is built, things are only going to get worse in my garage, before it gets better.

Related to your mention of Mini-Circuits HEMT, some mini-circuits product includes things like DC-2GHz amplifiers. It had me thinking that a RF pHEMT with 0.4dB noise figuremight serve for the first device along with that exceptional op-amp circuit. I perished that thought straight away! When a transistor has a whole lot of gain at GHz, it's got even more at audio, and it's quite hard to shut it down with feedback without it turning into a burping RF oscillator. They used to call it "squegging" or "motorboating", but mostly, all you know is the poor transistor is expiring in heat because it's doing stuff at a frequency the scope can't see. Low frequency stuff using high frequency bits still needs really careful attention to layout, and stability.

I have simulated that low noise FET circuit, and it's great at 0.95nV/√Hz, compared to 4.3nV/√Hz for the straightforward TIA op-amps. The higher level is likely "good enough", without needing a op-amp loop to control FET bias. I still have ordered some of those 1990's style low noise audio JFETs, just because less than $5 bucks gets you ten, and I might still want to play if there is a spare op-amp left in a quad-pack.

Shot noise is "low frequency" only in the sense that it happens at a low rate. When it does happen, the waveforms are fast enough to walk right in to circuits limited at low frequency. Regardless, I am tempted to have the gain stages also be filters. A band-pass starting at about 115Hz would inherently limit 50/60Hz pickup. At the other end, rolling off at 200kHz or so would lose a lot of the white noise that jangles the sample numbers. This would be at the price of phase delay. Not taken to extremes of "pulse stretching", but maybe OK to lose a bit of fuzz off the trace.

While I am trying to be minimalist with this, a few extra passives seems not to excessive.
 
@homebrewed Hi Mark
Re: ADCs data transfer - awkward choices
I have been juggling ADCs for a whole now and one of the considerations is the data interface.
The full parallel interface would be a big bunch of wires , but no speed issues. The lower cost ADC only uses SPI serial.

There are, of course, other things that affect choice. Trying for only 20 samples out of a 10uS time for a XRF pulse to rise and decay requires 2Msps.
Then for 16-bits in a surface mount package from Analog/Linear, yields 7 devices with prices ranging £18.65 up to £41.49 (all Mouser).

So - if one chooses (say) a AD4001 (lowest cost), which will only use SPI.
2Msps requires a 70MHz SCK clock.
A Raspberry Pi 4 can, in theory, manage 125MHz

I don't much fancy the chances with 4 loose wires in a bundle, and I don't know much about high speed SPI over a range more than the size of a PCB. I have done RS485, LVDS, CANbus, etc. This may be OK on a transmission line trace across a PCB. Somewhat more awkward to get it between the measure head and the display, unless the Pi or whatever lives in the same box as the sensor.

Yes, at those clock rates you're looking at a controlled impedance situation. Even the Teensy audio codec board has to employ a series damping resistor on its clock line, and it's running a lot slower than your ADC will. Even with controlled-impedance lines, at some point the prop delay will cause problems since the data is aligned to the (delayed) clock edges the ADC sees.

Regarding things like phase delay, I'm thinking it doesn't matter too much, else the Theremino MCA wouldn't work. It uses a highly modified pulse and still outputs reasonable spectra.
 
With tree stumps gone, I have been spending time outside with my new eBay Huepar laser level kit, getting the positions figured out, and some profiles in place. It will need a tracked vehicle digger, hard core, steel, etc.. I think the USA expression is "pulled the trigger on it". Essentially, I called in the same construction firm that I have used before, and their head guy will be visiting this week. Until that outhouse is built, things are only going to get worse in my garage, before it gets better.
Good luck with your project! It sounds like it's going to be a really NICE outhouse :D . Our barn-rehab project is to the point that we're talking about what color of wood stain to use on it (although I still have a couple of studs that need to be rebuilt). The rainy season has arrived so we have plenty of time to make that decision.
 
@graham-xrf I've been searching around for low-noise JFETs that are actually available (for a different project) and found some you might be interested in. The ones I found on Mouser have just so-so input capacitance for what I want, but might be OK for a TIA:

IF1330, 1nV/sqrt(Hz), Cin ~ 10pF, $3.90 in singles
2N5397, 3nV/sqrt(Hz), Cin ~5pF, $11.73
2N6550: .9nV/sqrt(Hz), Cin ~30pF, $11.94.

For comparison, the LTC6268-10, which has a 4GHz GBW, is specified to have En = 4nV/sqrt(Hz) and .45pF input C. Its price is comparable to the more expensive JFETs.
 
Back
Top