Needing more than a spark test?

[WobblyHand]

Something else to consider on the first stage is the cascade of 22Meg resistors. They are in series, which reduces their capacitance. There are parasitic capacitances in that circuit which probably change the behavior. Real resistors have both capacitance and inductance which changes how the circuit works.

 

[WobblyHand]

Designed and printed a punch for the flashing, that would form the lead to the right shape. I will use it to further explore this idea. The punch can be made on a lathe. Just finished printing it. I made it in mind that there would be 1.6mm lead between the pieces. The punch has a 6.4mm hole through it that I am hoping I can insert a metal punch and form the pockets. This idea might flop, but it has been a fun exercise.

The previous piece I printed could be used to do lead lost wax casting, I suppose. Would be a more direct way of forming a solid lead structure, although I think it would be overkill.

I priced out some lead the other day, it's fairly expensive unless you buy about 50lbs at a time. I wanted to buy some lead sheet at 1/16" thick. McMaster has 1' x 3' x 1/16" at $112 + shipping. The home repair places seem to only sell the 1mm stuff these days and it comes in giant rolls for over $150.

Rotometals has 1/16" lead sheet at an ok price. Unfortunately, their shipping rates were terrible. They have a minimum buy of $149 to get free shipping. If I bought 1' x 3' of sheet lead and did not have $149 of product in the cart, they wanted to charge me $120 shipping for the $36 worth of lead. However, if I had $149 worth in the cart AND added the sheet lead, then the sheet lead would have shipped free. So to buy about 3 sq ft of lead, I need to spend $200! On the plus side, not only do I get the sheet lead, but I get an extra chunk (62 lbs) of 99.9% pure lead. I suppose I could start casting again, but $200 to buy a little bit of lead sheet is a bit much. Got to see if I can find an alternate source.

So on a whim I checked eBay. Have no idea if it is pure lead, or if the dimensions are true, but 1'x2'x1/16" is $44 with shipping. That's worth a try. 2' is enough to try and have enough to mess around with.

 

[homebrewed]

Designed and printed a punch for the flashing, that would form the lead to the right shape. I will use it to further explore this idea. The punch can be made on a lathe. Just finished printing it. I made it in mind that there would be 1.6mm lead between the pieces. The punch has a 6.4mm hole through it that I am hoping I can insert a metal punch and form the pockets. This idea might flop, but it has been a fun exercise.
View attachment 434586View attachment 434587
The previous piece I printed could be used to do lead lost wax casting, I suppose. Would be a more direct way of forming a solid lead structure, although I think it would be overkill.

I priced out some lead the other day, it's fairly expensive unless you buy about 50lbs at a time. I wanted to buy some lead sheet at 1/16" thick. McMaster has 1' x 3' x 1/16" at $112 + shipping. The home repair places seem to only sell the 1mm stuff these days and it comes in giant rolls for over $150.

Rotometals has 1/16" lead sheet at an ok price. Unfortunately, their shipping rates were terrible. They have a minimum buy of $149 to get free shipping. If I bought 1' x 3' of sheet lead and did not have $149 of product in the cart, they wanted to charge me $120 shipping for the $36 worth of lead. However, if I had $149 worth in the cart AND added the sheet lead, then the sheet lead would have shipped free. So to buy about 3 sq ft of lead, I need to spend $200! On the plus side, not only do I get the sheet lead, but I get an extra chunk (62 lbs) of 99.9% pure lead. I suppose I could start casting again, but $200 to buy a little bit of lead sheet is a bit much. Got to see if I can find an alternate source.

So on a whim I checked eBay. Have no idea if it is pure lead, or if the dimensions are true, but 1'x2'x1/16" is $44 with shipping. That's worth a try. 2' is enough to try and have enough to mess around with.

The lead sheet I have was purchased from ebay. It sure looks & behaves like pure lead.

 

[WobblyHand]

Here's the die set with a steel pocket forming punch. Not sure if this will work. I'll find out once I get that lead sheet.

And the underside where the punch comes through.

6.35mm steel rod for the pocket forming. Somehow I have to clamp things together and form the pockets (8 times!).

 

[graham-xrf]

I know this is an old post, but, I think my answer may help others going forward. Graham, your answer was in your question. You said, “only cost-effective if they are identifying exotic stuff all all day long.” so, bring it to a place that checks “all day long”. We always relished any opportunity to help the “general public” who were brave enough to come in and ask for help. It was just good public relations. Any decent aerospace defense company will have one to verify every stick of material they receive. I’d imagine material handling companies and mills also have one.
even if they did choose to charge you for it, it wouldn’t cost much as it takes literally about 5 seconds of their time.

I did consider that, but the intermittent nature of the need, and what it takes to arrange even one of such trip in UK is a pain. Simply identifying a metal merchant within travel range (they all have these detectors), in a service economy like ours, is quite a task. Having a low cost gadget will get me there, made of re-purposed smoke detector elements, and some electronics that one can access from a smartphone browser, is very cool.

I do other stuff also that may not be considered "cost effective", like I own a small mill, and lathe(s), and a metal cutting bandsaw, various electronic instruments, and chemistry items, and I go flying sailplanes. While I am still around, I mean to act like I am still alive, even if I will be giving it all away someday. Folk here want to play with this stuff, and I am OK with trying to help when I can.

 

[WobblyHand]

The lead sheet I have was purchased from ebay. It sure looks & behaves like pure lead.

Found a source on eBay that claims 99.9% lead, so I bought it.

 

[homebrewed]

Something else to consider on the first stage is the cascade of 22Meg resistors. They are in series, which reduces their capacitance. There are parasitic capacitances in that circuit which probably change the behavior. Real resistors have both capacitance and inductance which changes how the circuit works.

Yes, it's quite possible that they are put in a string to reduce the parasitic C. 0201 SMTs would have the lowest possible capacitance. I wouldn't want to try to hand-solder those -- they are REALLY tiny. I used a string of 0201 resistors for a passive high frequency probe. Hand made, so I know exactly how tiny those things are.

 

[graham-xrf]

Something else to consider on the first stage is the cascade of 22Meg resistors. They are in series, which reduces their capacitance. There are parasitic capacitances in that circuit which probably change the behavior. Real resistors have both capacitance and inductance which changes how the circuit works.

The 3 x 22M resistors are just not viable in this TIA circuit, in my opinion. The capacitor shown is intentional, and not parasitic. In all these opamps connected as a current input mode device (TIA), the capacitor is necessary for stability.

We have the detailed simulations for these circuits. We know that the input current of 2 to 4 femto-amps is the input offset current, and the FET inputs have resistances of tera-ohms. Even electrons across the surface of the PCB would be a major source of error without electrometer-style guard rings. We know that an input current has no place to go up one of those inputs, other than to go up the 66MΩ.

For me, it strains credulity. The simulation shows it doing what one would expect, slowly working a CR low pass filter time constant which has a roll-off frequency below 1Hz. If that gain is 66 million for real, followed by a stage gain of x100, the gain is already 6.6 Trillion! If we imagine we have a "biggest" pulse that delivers 2V, it started with a current of about 300pA. More gain downstream on Mark's conditioner board reduces that to about 3pA.

If these are truly what happens with a photon measuring TIA, then I am wrong, and I have to say sorry, but right now, I think this scenario unlikely. Every TIA amplifier circuit I have seen has never had a Rf greater than 1MΩ . Many deliberately start with 20K, or 100K, laid down as two in series, and the compensation capacitor is similarly made up of two in series, trying for values between 0.5pF and 2pF. The gain, as needed, is added in later stages. The first stage tries to have sufficient gain to lock in the signal-to-noise ratio, decided only by the input noise performance, and not by a madly high value feedback resistor total. By all means have Rf made up of a series cascade, but I have trouble believing it should be 66MΩ.

 

[homebrewed]

Oh dear!
If the little chip was taken to it's max and beyond, would you maybe expect it took some damage and should be replaced.

I am confused about how much of the noise you see is quantization noise, and why you have only 500mV signal. That said, as much as 500mV would be a fully expected good return that you would want to measure amplitude to high resolution. That would be about 16,000 in a 16-bit counter, but I don't know how many effective bits you have right now. I think the 16-bit ADC board would be much more than "interesting". I think perhaps you under-state things. Probably, you can hardly wait for the package to arrive!

The signal conditioner provides a x1.001 gain buffer, and then a (1+9) = 10x in U1. Then another (1+9) = 10x in U2. The other U2 output is not used. Thus, the board has gain about x100 already. Going back into the Pocket Geiger, before the TIA, there is another x100.
So far, it's 80dBV, before we take in the gain of the TIA stage.
Taken at face value, as published, the 66MΩ, for a current input amplifier has gain 66 million. Total gain now 6.6e+11
A full sized maxed out 2V at the ADC would have been made from a current of 3pA !

At this point, I run out of ways to think about it, because that little current is not what I would equate to a "maximum".

The offset and noise floor of the signal conditioning circuit look OK so I may have lucked out. But I'll check to see if I have any more LMC662's I can swap in. I also may have a lower-noise alternative, although I don't think that would make much of an improvement -- the largest gain is right up front in the amplifier chain.

Regarding the relatively low pulse amplitude, given the large overall gain, has me wondering if perhaps there's a problem with my pocket geiger -- the detector or maybe even an incorrect component value. It will be good to see what results other folks get with a similar setup. FWIW, the pocket geiger's voltage divider string establishes a +200mV detection threshold on comparator U3B, relative to U1's common signal "ground".

However, I do occasionally see pulses that are _much_ larger in amplitude. They show up regardless if I have the Am241 sources present or not so they have to be from the environment or are cosmic rays.

 

[WobblyHand]

The 3 x 22M resistors are just not viable in this TIA circuit, in my opinion. The capacitor shown is intentional, and not parasitic. In all these opamps connected as a current input mode device (TIA), the capacitor is necessary for stability.

We have the detailed simulations for these circuits. We know that the input current of 2 to 4 femto-amps is the input offset current, and the FET inputs have resistances of tera-ohms. Even electrons across the surface of the PCB would be a major source of error without electrometer-style guard rings. We know that an input current has no place to go up one of those inputs, other than to go up the 66MΩ.

For me, it strains credulity. The simulation shows it doing what one would expect, slowly working a CR low pass filter time constant which has a roll-off frequency below 1Hz. If that gain is 66 million for real, followed by a stage gain of x100, the gain is already 6.6 Trillion! If we imagine we have a "biggest" pulse that delivers 2V, it started with a current of about 300pA. More gain downstream on Mark's conditioner board reduces that to about 3pA.

If these are truly what happens with a photon measuring TIA, then I am wrong, and I have to say sorry, but right now, I think this scenario unlikely. Every TIA amplifier circuit I have seen has never had a Rf greater than 1MΩ . Many deliberately start with 20K, or 100K, laid down as two in series, and the compensation capacitor is similarly made up of two in series, trying for values between 0.5pF and 2pF. The gain, as needed, is added in later stages. The first stage tries to have sufficient gain to lock in the signal-to-noise ratio, decided only by the input noise performance, and not by a madly high value feedback resistor total. By all means have Rf made up of a series cascade, but I have trouble believing it should be 66MΩ.

Not disagreeing with you. The 1 pF cap is intentional. Maybe the extra 0.3pF due to the resistors capacitance are doing something - and as far as I can tell, that capacitance is not in the simulated circuit. Did you select actual component models, like actual part numbers for the resistors and capacitors? Many times they include parasitics - and that can change the behavior.

There's a ridiculous amount of paper gain in this circuit, but I don't think we are getting that gain. I'd wager there's something else that is limiting the performance, like maybe slew rate limiting or some other effect. It could be the sim models are not accurate, or don't model something important, it wouldn't be the first time. It appears the physical circuit is not behaving according to the sim. The circuit sort of works, or at least after Mark fixed it.

When simulation and reality don't line up, you have to go with reality. Because that is measurable and real. Not saying we can't improve things, but sometimes sims just can't quite get you all the way there, and you have to resort to working in the lab. I have many spice sims that have gotten me close - but to get to the finish line it took playing at the bench.