Needing more than a spark test?

[graham-xrf]

@WobblyHand
Hi Bruce
Again I say - tenacious!
I did not look deeply into stability of L662, because I was never going to use them.
I do see the capacitance of the reverse biased D1. It gets to about 50pF at 33V
It gets to about 85pF at 10V.
I calculated that the charge electrons spreading into it, would have made a 85/50= x1.7 as much "voltage", if it were ever allowed to charge the diode capacitance, which it never does, because it goes up into the TIA.
For me, I use 10V because that is where I want the dark current noise to be at. My amplifier gain can make up the signal to match without hurting the SNR ratio.

On the question of the appropriate value across the Rf for stability. I guess I will have to check out the app notes.
On my TIA, the GBW is 4GHz on the LT6268-10, and 500MHz on the plain ole' LT6268
The GBW of LM662 is 1.4MHz !

OK - I have found it!

Cs is predominantly the 50pF. Cf is acting as a (strange) integrator. Indeed, there is an integrator within the opamp.
Rin would be 40MΩif it means the bulk resistance of the photodiode.
Going with Rf = 66MΩ. (initially) CF should be .. ( 50pF x 40MΩ )/60MΩ = 3.333e-11, or 33pF
Wow. I guess that is what happens if you have a pole that happens early!

I know I had misgivings, perhaps even some justified prejudice when I first started looking at Pocket Geiger circuits, but given I did not know much about transimpedance operation, I tried to keep an open mind. While I am sure the LM662 might be useful in some medical applications, the deal breaker for me is this ..



It starts rolling off at somewhere less than 10Hz !

At one stage I considered whether the comparator LM393 pinout tracking matched any useful opamps, and there are some that do, like LTC6244, but sadly, no place to put feedback resistors and suchlike.
------------------------------
Re: The LTSpice LM662 simulation.
I do hope that model file I sent you is not the culprit! The original files from TI (.MOD) are in the zip. You need the second one that specifies the 6mV imput offset, and change the subcircuit name to match whatever you use on the symbol. I made a symbol too.

Dinner is called, I gotta go,
I think I should get up some more LTSpice TIA amplifiers, posted here, that simply use the available models that come in the installation update

 

[WobblyHand]

Everything is pointing to the LMC662 being not a good choice if one wants a good representation of the pulse. Perhaps it could be used for some slow speed stuff. The low GBW and slew rate are really limiting. It's one advantage is it has CMOS inputs and very low Ib and Ios.

For what it is worth, TI is calculating (in one of their app notes) the required loop capacitance in a different way. Roughly
Cf = sqrt(Cd / (GBW * Rf), where Cd is the biased diode capacitance. I think they are making approximations for certain conditions, but I'm not sure if those conditions are applicable. They do note that the required capacitances are less than determined the traditional way. One more reason to see if these circuits can be simulated.

 

[homebrewed]

The pseudo-nulling scheme on scopes works OK up to the clipping range of the amplifier chains but due to the inevitable slight mismatches in phase between the two channels you aren't ever going to get a really high rejection ratio.

Once you enclose your PG in a box a lot of your noise problems will go away. It was a pain to make my enclosure but it's been pretty useful for that kind of thing, and robust enough to handle frequent dissasembly w/o mishap.

I used round DIN connectors for power and signal. I hate filing out square-ish holes for DB type connectors....

 

[homebrewed]

Perhaps bullheaded. Or stubborn.

Anyways, I honestly see no effect of the source. I do see some effect of local capacitance, even getting my hand 3-4 cm away. At first I thought that was due to the source, but I was mistaken, it's just my hand capacitance.

I will fiddle with this some more tomorrow. Will drag out the microscope, to make it easier on me to solder stuff.

The Am241 is predominantly putting out two kinds of radiation. The first is alpha particles, which won't make it through the epoxy covering the PIN diode. The second is 60Kev gamma rays, and the PIN diode has a very low probability of detecting them. Try covering the smoke detector source with some copper foil and putting that near the PIN diode & then check the output of the amplifier chain. The foil should convert some of those higher-energy photons to 8.8Kev photons which are close to the maximum-probability peak of the detector.

the Am241 source is on a little steel carrier that's emitting _some_ ~6Kev photons but from what I'm seeing it isn't a lot.

 

[WobblyHand]

The pseudo-nulling scheme on scopes works OK up to the clipping range of the amplifier chains but due to the inevitable slight mismatches in phase between the two channels you aren't ever going to get a really high rejection ratio.

Once you enclose your PG in a box a lot of your noise problems will go away. It was a pain to make my enclosure but it's been pretty useful for that kind of thing, and robust enough to handle frequent dissasembly w/o mishap.

I used round DIN connectors for power and signal. I hate filing out square-ish holes for DB type connectors....

Would a plain metal bud box be sufficient? How good does the grounding have to be all the way around? I have some circular connectors I could use. DB connectors are no fun to file out, and the punches are ridiculous.

 

[homebrewed]

On a slightly different subject, I think I have discovered why my aluminum focus ring is dumping so many aluminum counts into my MCA, and there isn't any good way to fix that -- other than getting rid of it!

The basic problem is that the little peened-in thingie still is larger than the Am241 source dot. In my OpenSCAD model I was using the actual size of the dot. Once I used the unavoidable carrier itself, the size of the focus ring needed to accommodate the carrier expanded, to the point that a substantial part of the ring AND the carrier can be "seen" by the detector. I'm NOT going to try machining the carrier, I don't care to think about the possibility of releasing Am241 dust in my machine shop. So the only option for this kind of structure is to go wtih a plastic source holder. Period.

Here's my proof.

Photo 1 showing the focus ring and full--sized sources from the top:


And now a view from the detector side. Remember that the gray color represents the aluminum focus ring:

So, QED. No aluminum allowed.

 

[graham-xrf]

Hi Mark
Yes - the X-rays in the aluminium come in all sorts of directions, and may well be headed towards the diode. They keep on going through the relatively transparent and the arrive at the diode.

At least, that is what I think you might mean.

Sure, if you go with plastic, I guess you still need a lead ring fitted into it, surrounding the diode, to stop the initial Am241 radiation from simply heading sideways into the diode, releasing whatever they find in the ceramic carrier, the connection metals, etc.

Maybe I should be looking at 3D-printers?

Or is it that we just get some plastic rod, and turn that?

 

[WobblyHand]

On a slightly different subject, I think I have discovered why my aluminum focus ring is dumping so many aluminum counts into my MCA, and there isn't any good way to fix that -- other than getting rid of it!

The basic problem is that the little peened-in thingie still is larger than the Am241 source dot. In my OpenSCAD model I was using the actual size of the dot. Once I used the unavoidable carrier itself, the size of the focus ring needed to accommodate the carrier expanded, to the point that a substantial part of the ring AND the carrier can be "seen" by the detector. I'm NOT going to try machining the carrier, I don't care to think about the possibility of releasing Am241 dust in my machine shop. So the only option for this kind of structure is to go wtih a plastic source holder. Period.

Here's my proof.

Photo 1 showing the focus ring and full--sized sources from the top:
View attachment 435951

And now a view from the detector side. Remember that the gray color represents the aluminum focus ring:
View attachment 435952
So, QED. No aluminum allowed.

Not sure what I am seeing. Are you saying that your sources as mounted are shining on the inner ring? So they are releasing aluminum xrays? Can you recess the sources into a lead dish? Or are you saying the sources no matter what you do are washing the aluminum ring structure with 59KeV xrays? If the latter, yeah, that's no good. Then one needs to get rid of the aluminum, or recess the sources in lead so they only radiate in a limited cone, basically limited by the depth and diameter of the lead pocket ID.

If you can design up something in plastic - convert the source holder model to an stl - I can print it for you. Take me a couple of hours to print and I can send it to you. If you have mounting holes or other features, let me know (via the model) where they are. It will get you pretty close. Although it's not structural like aluminum, you can thread into it and install it to things.

 

[WobblyHand]

Hi Mark
Yes - the X-rays in the aluminium come in all sorts of directions, and may well be headed towards the diode. They keep on going through the relatively transparent and the arrive at the diode.

At least, that is what I think you might mean.

Sure, if you go with plastic, I guess you still need a lead ring fitted into it, surrounding the diode, to stop the initial Am241 radiation from simply heading sideways into the diode, releasing whatever they find in the ceramic carrier, the connection metals, etc.

Maybe I should be looking at 3D-printers?

Or is it that we just get some plastic rod, and turn that?

I would think lead pockets would reduce the xrays to the detector, and have the unobstructed part shine through the pockets aperture. Poor man's xray flashlight.

 

[graham-xrf]

I would think lead pockets would reduce the xrays to the detector, and have the unobstructed part shine through the pockets aperture. Poor man's xray flashlight.

You are right about the pockets, so long as they are not too deep. Going for lead the right shape seemed easier than multi-material fabrications.

It may be possible to just set down a circular array of sources on a flat disc of lead sheet, exploiting the fact the rays go in all directions, and rely on the ring around diode to stop all except those useful ones coming back from a test material from getting in. This is a much simpler, easier to make, arrangement. I am tempted. I thought it was the scheme Mark used at the beginning.

Re: the FreeCAD design. I got carried away. I thought my logic was hardball !

Any Am241 photons not destined to reach the material under test are not only useless, and unproductive, they can only cause problems if allowed to make it onto any other electronics or material anywhere. They are utterly banned. Unfortunately, we cannot "reflect" X-rays, like a flashlight.
I know Mark @homebrewed uses the term "focusing ring". I know what he means, but I can't really call it that. The most I call it is "collimating", and even that term now yields to "shielding". Yeah, I know, too picky by more than half!

I looked at the geometry. This with a view to getting the available pile of useful direction photons onto the test material in absolutely the closest distance it can. We allow all these Am241 photons that would be useful the widest possible solid angle onto the test material, and stop all others. Of course, this motivated the angled mounting, and the circular array. I have thought to double the number of sources, but not yet.

Regarding the the shadowed route to up to the diode. It is designed to be as unobstructed as possible. The only role of the shield is to stop initial Am241 photons that would not have yielded any useful results from the material under test anyway, from coming at the diode sideways. This includes any secondaries the unproductive photons it might have spawned. These considerations set the height of the shield, to do this, and no more.

The whole shape is designed to maximize the returns from the material under test, intent on allowing every possible useful Am241 ray to do its stuff, and to allow in the maximum of it's test material results to make it to the detector. I thought the arrangement, in lead, was uncompromising. Hopefully, I have not fumbled.

I found it heartening that when I displayed Mark's SCAD model in FreeCAD, we seemed to have ended up with pretty much the same concepts, and the look of the thing ends up much the same.

I must admit I am thinking to go with a scheme that does not involve turning any lead, even though the end result would look very much better if machined. If all one needs is to cut some sheet, and possibly roll a bit around a Sharpie, the notion has its attractions!