A silly (?) question to you chemists out there

All Recipes and Chemistry!

Moderators: richardh08, Boophoenix, Lloyd

sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Microcontrollers are renowned for their noise, they throw out and suck up all sorts of current. You just need to decouple it.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

In answer to Lloyd's comments, I'm using Arduino's own development system. The language is based on C/C++ and is compiled to the chip's native code. I have no idea how efficient the compiler is.

Obsessed with this topic as I am, I don't have any intention of learning the chip's instruction set and/or attempting to write assembly language code.

I don't see the noise unless my circuitry is both attached to the Arduino's ports and powered up, so I think that Sam is correct in his interpretation of its source. It seems clear to me that, when designing and building the Mk2 circuit, I will need to pay much more attention to a) the op-amp's offset voltage, b) the layout and c) effective decoupling.

On the positive side, I tried hooking one of my grains with an embedded thermocouple to the Mk1 board. It gave me a signal - and it popped, without destroying the junction! So all the individual parts of the system work, at least in principle. The downside is that I can't make sense of the recorded data:
RealGrain.png
Those two large dips - especially the second, which drops to precisely zero for 0.3 sec - are worrying. They imply either a drop in temperature to way below ambient or a malfunction somewhere in the system.

I also don't know how the measurements relate to the behaviour I've seen in the earlier video records, particularly where in the recording the actual 'pop' occurred. Further tests will have to wait until I've got a more reliable circuit. It's also clear that I will need to run a video recording at the same time as measuring the thermocouple output, at least until a correspondence is established.
You do not have the required permissions to view the files attached to this post.
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Richard I think you're making great progress. There is no garbage collection on the arduino, not like with the Raspberry Pi which is a much more evolved system but the AtMega is built for its simplicity and to be used effectively out of the box. Could those dips be a transient short. Will the contacts effectively short if the grain becomes molten ?

I've got to say it, I'm really loving this - for me this is the way forwards. Shimizu did some similar work with his glitter tests in Pyrotechnica XIV.

Sam.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

The only way I can think of to remove the possibility of a 'short' would be to lightly coat the junction with a refractory compound that is not conductive at high temperatures. I might have mentioned that in a previous post.

But... when I think of thermocouple junctions... the whole idea is to have them "shorted-out"; one metal to the other. I'm not sure what's in that composition that could actually cause a direct short, nor how it could also be removed (just by flame action) in such a short time.

I, too, am loving this. I WISH I had the time from client projects to pursue such "pure research"!

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

At first I thought of a 'short', but that can't be the cause. Because of the uncorrected offsets, the Mk1 circuit gives an output of about +1.3V in that circumstance. I also wondered if I were overloading the ADC, but that is unlikely as the output saturates at about +4.9V and, in any case, the signal in the neigbourhood of the 'anomalies' isn't that close to saturation.

About the only things I can think of that could give a reading of zero are an interruption of the circuit's power supply or disabling the ADC. But both of those seem unlikely to me.

To be on the safe side, I don't intend to run any more tests until I've built and tested the Mk2 circuit.
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

I'm sure it will be on the analogue side Richard. A loss of current through your op-amp from whatever cause. It's a little difficult to be any more helpful I'm afraid as I'm not completely sure how you have this wired up. Good luck with MK2 though !

Sam.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I suspect there was an intermittent fault in the Mk1 circuit.

Mk2 is now up and running. Here's the result of a preliminary test with a bare junction more or less randomly heated with a lighter flame:
Mk2Test.png
It's a marked improvement in that the offset is reduced to zero and the decoupling seems to have eliminated just about all of the noise I was seeing in the previous circuit.

The software will need a few tweaks to improve the synchronisation between the run length (which is limited by the amount of memory on the board) and the grain's burn time. I'll also probably need to adjust the amplifier's gain a little, once I see the level of output I get from a real grain.
You do not have the required permissions to view the files attached to this post.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Richard,
Considering the low 'heat mass' of that junction, that's an amazingly stable output! I think you have it 'licked'.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I think I still have a couple of problems to solve.

I think I'll need more frequent sampling to catch what happens near the point of explosion, which means using more memory than is available on the board. I'm thinking of investigating data-logging to an SD card.

Also, it's beginning to look like my hope that the data would be auto-calibrating isn't going to work. I'll therefore have to come up with some means of calibrating the output to get actual temperature readings.

Either way, it looks like I'm going to have to make a LOT more thermocouples :(
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Hi Richard,

it's looking really good. How much memory do you need? I have some ram chips for the atmega somewhere or another if they are of any use to you ?

At that kind of voltage range can you not store your output as a byte, or do you need greater resolution than 255 intervals ?

Sam.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Hi Sam,

That's kind of you, but I've decided to go the SD card route. I thought about dropping down to 8 bit resolution, but it could still mean having to juggle the data if I find I need to up the sample rate or extend the recording time. Writing to external data storage will allow me to forget about such problems.
Even when I'm wrong, I'm convincing.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

The circuitry and software is now up and running, recording the data to an attached microSD card. To allow for the fact that the data to be measured is still a little uncertain, I currently record outputs from each of the three final stages of a 4-stage amplifier. The actual gains in each stage are x79 (fixed), x16 (variable), x2 (fixed) and x2 (fixed).

Timing tests show that writing to the SD card is the limiting factor. A typical cycle takes around 700µs, but every few cycles that increases to about 5400µs (5.4ms), presumably as buffered data is actually flushed to the card. Explicitly flushing the data in every cycle makes matters much worse, as each cycle then takes about 13.5ms. With the current software it seems that the highest sensible rate at which I can take measurements is once every 10ms, i.e. 100 times per second. I can see that, if necessary, I might be able to improve on that by a factor of 10 or so.

I've also run some rough calibration tests, which proved to be a little trickier than I expected, but they seem to give reasonably consistent results. They tell me that the sensitivity of the W/Mo junction is about 3.9µV/K, which is quite low compared with the output from more conventional thermocouples, but doesn't seem unreasonable to me. With my present setup I ought to be able to measure temperatures up to about 1700K with moderate accuracy. If necessary, the maximum can be shifted up or down by tweaking the amplifier's gain.

Further progress must wait upon the making of another set of real samples, to see if their output is within the current range of measurement.
Even when I'm wrong, I'm convincing.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Here's the result of the calibration of a single W/Mo thermocouple junction.
Calibration.png
The three points are: room temperature (293K), condensing steam (373K) and solidifying KNO3 (607K). It's based on the assumption that the response is more or less linear with temperature, which the results seem to confirm.

Trials with a couple more examples suggest that they give reasonably consistent outputs and that I can trust the resulting temperatures to within ±15K or better.

A timing test shows that I can, indeed, take a reading every 10ms, which is four times faster than I can get from my video recordings (25fps).

All I have to do now is get some real recordings (and understand what they tell me)!
You do not have the required permissions to view the files attached to this post.
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Very impressed, it's looking really great.
You might be able to increase your sample rate further by "simultaneously" writing to RAM and the SD card. Store alternate readings in each then write your RAM buffer to the SD card at the end - depending on of course how many values you have.

looking forward to following this some more.

Sam.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

There are undoubtedly ways to increase the sampling rate, but I probably won't put much effort into that aspect until the current software is shown to be inadequate. I'll wait until I get some reliable data from real grains before deciding how to proceed.
Even when I'm wrong, I'm convincing.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I’ve got some initial results from my thermocouples which are, to say the least, somewhat confusing. They don’t seem to fit with anything that I’ve previously seen. They clearly show features, but I can’t relate any of them to the more obvious physical or chemical properties of the mix. One feature is truly weird and doesn’t appear to make sense.

At the moment, I can’t even relate the results to what I see in a video recording of the same event. Either there’s something wrong with my measurements or the mechanism is even stranger than I thought. I even got to the point where I began to doubt whether the timescales were compatible, but I’ve checked both the thermocouple software and my camera (!) and the assumed data rates appear to be correct, and also agree with each other to a very high degree of accuracy.

I need to make more measurements, but I first need to devise a means of synchronising the thermocouple and video data, and to tweak my circuitry to make it sensitive to a wider range of temperature. On and off, it took me most of yesterday to make four (yes, only four!) thermocouple assemblies and I have yet to add the crackle composition, so it will be a while before I’m ready to get more data.

I’ve just finished re-reading “Fermat’s Last Theorem” by Simon Singh. The proof of the theorem eluded mathematicians for over 300 years and was only resolved after many failures and trips down blind alleys. Although my search for the crackle mechanism is nothing like as erudite or theoretical, I do see parallels. One quotation caught my eye as being highly relevant:

An expert problem solver must be endowed with two incompatible qualities – a restless imagination and a patient pertinacity. Howard W. Eves.

The fact that I’ve been working at the topic for the best part of three years suggests that I have the second of these qualities in spades, but I’m not so sure about the first. I clearly don’t have the imagination to deduce what the initial thermocouple data is trying to tell me!
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Easy there fella - stop being so harsh on yourself. You clearly have both of those qualities. Can you flash an LED sequence via the IO pins and pick those up with your camera. Say red for every 5 miliseconds and green for every 10 milliseconds ?
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

GOOD idea, Sambo! It would take a TINY bit of code to add that, and would guarantee synchrony between camera and event.

It would probably require that the I/Os be programmed for "one-shot" service, in order to lengthen the pulse enough to be seen without 'eating' too much loop time, but most micro-controllers permit that.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I didn't mean to go quiet for so long, but just after my last post we had a power cut. It's the first we've had in years but power wasn't restored until about 5am this morning! It makes me realise just how much everything we do these days is totally dependent on a mains power supply. I haven't looked yet, but I imagine the contents of the freezer are 'toast'.

I like the LED idea and was already thinking along those lines. Lots of free pins available and still plenty of program space. Since I've convinced myself that the two are measuring on the same timescale, I wouldn't need to keep flashing at regular intervals (although it wouldn't take up much time or programming space to do so). Just marking the start (and perhaps the end) of the recording thermocouple session would be enough to bring them into synchronisation and wouldn't mean further burdening the measurement loop. K.I.S.S.
Even when I'm wrong, I'm convincing.
sambo
Posts: 957
Joined: Sat May 16, 2015 6:04 pm

Re: A silly (?) question to you chemists out there

Post by sambo »

Brad really kindly sent me The Best of AFN VII, amongst other things. In fact, Brad if you are reading this, it has arrived and I am eternally grateful to you. I owe ya big time!. Anyhow, the point is, Lloyd Scott Oglesby wrote an article regarding the conductivity of crackle grains which is featured in that issue. In fact it features a couople of pages ( P55 ) after your articles. He describes it as an electrochemical reaction that goes into short circuit and talks about demonstrating this by connecting two wires to a grain of the crackle mix and measuring the resistance changes at the smolder stage, the heat up from that which produces a sudden lowering to nearly no resistance at all which results in the explosion. Are these effects what you are seeing with your data Richard ?

Sam.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I haven't seen that article, Sam, so I can't really comment. But given that the crackle composition obviously gets quite hot, it wouldn't surprise me if the resistance dropped dramatically - for one of any number of reasons. I don't want to sound disrespectful, but I haven't always been totally convinced by some of LSO's explanations.

I just don't yet have the data to make any conclusions about what might be happening. I'm still trying to get to grips with the properties of the thermocouples themselves, as well as the characteristics of my new circuit. I'm also only at the very earliest stage of how to embed them in a grain of the right size. So far I've only got two results, with great uncertainty about which features are real, and how they relate to the different stages of the reaction.

Even if there is a drop in resistance, I can't see how a thermocouple could pick that up.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

"I don't want to sound disrespectful, but I haven't always been totally convinced by some of LSO's explanations."
-=======-
Thank you, Richard! Me, too! And there's that unfortunate similarity of first names! <grin>

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

PS --- "But given that the crackle composition obviously gets quite hot, it wouldn't surprise me if the resistance dropped dramatically - for one of any number of reasons..."

Ayup! Molten substances with metals in them display that characteristic nearly across-the-board. As for the 'number of reasons'... many salts demonstrate the same property.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

The really weird feature I'm seeing is a sudden dip to zero output, as shown in my earlier post of a recording with my Mk1 amplifier. At first I wondered if it were either a circuit fault or some bizarre endothermic event but, now that Ive seen it with both the Mk1 and Mk2 amplifiers, I believe it to be some kind of effect that is generated by the twisted (W/Mo) pair I'm using. Given the characteristics of my amplifier, it could be explained if the thermocouple momentarily went open circuit (i.e. effectively infinite resistance across the input of my amplifier, which is exactly the opposite of LSO's fall in resistance!).

I now have a single set of correlated data that ties this behaviour very precisely to the moment of the explosion, and to a duration of 0.27 seconds (which agrees with the about 0.3s duration of the previous recording). I'm struggling to understand what is causing it, let alone how to fix it.
Even when I'm wrong, I'm convincing.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I suppose I’m making progress, but it’s very slow. Over the last week or so I’ve learned more about thermocouples themselves, rather than what I’m trying to measure with them.

The main problem I have is the apparent break in contact at the moment of explosion (I’ve convinced myself that’s what is happening). I’ve been looking at alternatives, including other materials that are easier to solder. Unfortunately, with lower melting points and less resistance to oxidation, the junctions turned out to be just too fragile. Also, soldering turned out not to be a good idea. However, just before the thermocouple died, I did manage to get a reasonably good estimate of the melting point of my solder!

I think the best solution will be to look into the possibility of impact welding the W/Mo junctions. Getting a reliably consistent technique will be the aim.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Richard,
It's a delicate operation on wires that small, but they typically 'braze' those thermocouples together. Eh... it's still 'soldering', but high-temp.

I don't know how they do it, but imagine they're using a dry-in-place water-solvated paste flux loaded with the 'soldering metal' in powder form, then doing the joining in an oven. I can't see how you'd do it with a torch.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Lloyd wrote:...but imagine they're using a dry-in-place water-solvated paste flux loaded with the 'soldering metal' in powder form, then doing the joining in an oven. I can't see how you'd do it with a torch.
I guess I won't be doing that, then! My only realistic options seem to be impact welding and/or a thin, inert coating.

Anyway, I have another batch in preparation so I might have more results in a week or so.

The fragmentary data I have up to now seems to be indicating some problems with my current theory. However, it also suggests that there are just as many problems with any other theory about how crackle works.
Even when I'm wrong, I'm convincing.
User avatar
biffo
Posts: 278
Joined: Mon Oct 13, 2014 9:25 pm

Re: A silly (?) question to you chemists out there

Post by biffo »

Hi Richard is there any way you might be able to melt some brazing rod in a small stainless pot in a makeshift hearth and dip the ends in flux and then into the molten brass aloy. If you are able to post one or two assemblies to me I could spot weld or at least try to if these metals can be spot welded ,with a borrowed spot welder.if I can be of assistance please let me know .Regards B
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

I tend to believe that thermocouples won't perform properly if the wire is welded in only one spot. I believe (can't justify it, though), that they need 'contact area' larger than a single spot.

Otherwise, why would they 'traditionally' be made by twisting several turns, instead of just touching the two wires together at the ends, and spot-welding?

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Thanks Biffo. If my next round of trials fail, I might take you up on your kind offer.

Lloyd, the following link suggests that simply welding the tips does work:
http://www.thermosensedirect.com/acatal ... uples.html.

It's a shame that they are just too massive for my needs.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Richard,
Thanks. I'll have to educate myself a bit better on how they work. I thought the multiple-twist contact area was important. Perhaps it's only important for the current the thermocouple can deliver, but not the voltage. With modern, high input-impedance amplifiers, the current-delivery wouldn't be so important.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
dave321
Posts: 388
Joined: Sat Apr 25, 2015 11:53 pm

Re: A silly (?) question to you chemists out there

Post by dave321 »

Richard,

maybe
http://www.macgregorsystems.com/

would be willing to micro spot weld you a few "test" thermocuples for you to try out ?
always worth asking

dave
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Thanks Dave, I'll bear them in mind. But I'm not sure either of these last two suggestions would be viable long-term solutions.

Ken and I have been discussing whether it would be possible to use a pressure welding technique and have come up with an idea. Time will tell whether it works or not.

In the meantime Ive been trying to learn a bit more about the thermocouples themselves. I've made up a batch of 10 twisted pair assemblies. It takes a long time because they are very small and delicate. I can't keep up the necessary concentration for long, so the work is done in short bursts, with long relaxation periods.

Today I ran a quick test on each of them, recording the output when they are held in the steam emerging from a boiling kettle. It's not hugely scientific, but the consistency of the results is a lot better than I expected. On an average ADC reading of 108, the standard deviation across all 10 samples is 3.5 - which isn't much more than the spread of values (about 2.8) that I see in the output from each individual device.
Even when I'm wrong, I'm convincing.
User avatar
biffo
Posts: 278
Joined: Mon Oct 13, 2014 9:25 pm

Re: A silly (?) question to you chemists out there

Post by biffo »

Hi Richard just had an idea which you might have already tried but what about twisting the wires in the middle of its length then bringing both pairs back to there respective polaratives,so you have no breaks in the cables to blow apart.would it be viable? Regards B
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I'm not convinced that the cut ends of the twisted pair have anything to do with this problem. The vast majority of the samples I've tried so far survive intact, with the wires apparently twisted as tightly as they were at the start - and still functioning with about the same sensitivity.

Since the crackle grains produce a very sharp explosive sound, it seems clear to me that they must generate a brief, but very intense, pressure wave. All I can think of is that this pressure forces reaction products into the (previously non-existent) space between the wires, and that afterwards the wires' natural elasticity restores them to their original shape and electrical contact.

However, it all seems a bit unlikely and I admit that I'm clutching at straws.

At the moment I'm just hoping that our impact welding idea will work. Otherwise we'll have to think of something else, or put up with losing about a quarter of a second of data around the critical instant.
Even when I'm wrong, I'm convincing.
User avatar
Boophoenix
Posts: 968
Joined: Wed Sep 09, 2015 11:49 pm

Re: A silly (?) question to you chemists out there

Post by Boophoenix »

Richard, why your grasping keep in mind most metals expand with heat I do believe. This could intensify the effect from the wave and / or make diagnosis difficult.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

"Since the crackle grains produce a very sharp explosive sound, it seems clear to me that they must generate a brief, but very intense, pressure wave. All I can think of is that this pressure forces reaction products into the (previously non-existent) space between the wires, and that afterwards the wires' natural elasticity restores them to their original shape and electrical contact."
----------------
Richard, IF that is happening (I'm not convinced), one possible way to solve that would be to 'seal' the union with some sort of gas-impervious material. Since brazing things of so little mass is out of your ken, two ways come to mind. One is 'ordinary' soldering, with a good active flux. The other might be to simply apply a very thin coating material (maybe even just a thinned lacquer), so that the union area is gas-impervious.

I'm leaning toward thinking that the superheated metal-bearing gas (eh... probably not 'plasma') forming around the junction is conductive during its brief hot spell, and disturbs the 'bi-metallic' nature of the union. I'm not clear on how one would prevent that from happening.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I'm not sure I believe it either, Lloyd. Your explanation is at least as convincing as mine. Perhaps the hot conducting gas (I agree it is highly unlikely to be a true plasma) provides a short circuit on the (cooler) lead wires rather than at the twisted junction itself, but the conductive path would have to be pretty good to reduce the output to precisely zero, which is what I currently see.

I might be able to modify my amplifier in such a way that it could distinguish between a short and an open circuit, which would push the answer towards one or other of these two possibilities. Knowing which way it goes might help to work out if a fix is possible.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

"I might be able to modify my amplifier in such a way that it could distinguish between a short and an open circuit..."
00000
That would be pretty easy to do! Add a series resistor to the 'couple, and provide a 'bias current' through the junction. Then just adding another op-amp to read the voltage across the resistor would tell you if the current ceased or went 'ballistic' during a short.

Adding a small bias current wouldn't disrupt the voltage readings. It would only 'offset' them some. Were it me, I think I would (first) choose the resistor and bias current to have a voltage across the resistor approximately-equal to the maximum voltage the junction produces during the heating. I would make the voltage-reading op amp's inputs very high-resistance compared to the value of the 'dropping' resistor, in order to avoid additional offsets caused by junctions in the amp.

Lloy
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

It's not so much how to modify the circuit, but whether I have enough room for the additional components. Fortunately I have some spare op amps already on board, so I only need to find space for two or three resistors. But it's built on a piece of stripboard and already packed quite tightly.

Also, I now have the little impact welding tool that Ken and I dreamed up - thanks very much for making it Ken.

So it looks like I have a fair bit of testing and readjustment to do before I can get back to making measurements.
Even when I'm wrong, I'm convincing.
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

The circuit is now modified, with an additional output that will distinguish an open circuit condition. While working on the board, I took the opportunity to make the thermocouple's connection to the amplifier more reliable - the original connection was somewhat fragile so I've beefed it up.

I've also started playing with the impact 'welding' tool. It is quite fiddly to get the junction in exactly the right position (I've already broken one junction assembly by hitting it in the wrong place) but, with care, it seems to work as expected. Here's a picture of a raw junction and two that have had the treatment.
Squashed.jpg
There appears to be little or no change in the junction's resistance, so only time will tell if the wires are actually welded and/or if it is a fix for the problem. It will take several days, possibly a week, before I might be able to get some answers.
You do not have the required permissions to view the files attached to this post.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Hey! You're fast! Does the addition offset the voltage, or did you do what I failed to mention last time, and put the thermocouple on the 'ground side' of the network, and read the series resistor's voltage on the 'high side'?

LLoyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Lloyd wrote:Hey! You're fast!
Well, I've nothing better to do!
Lloyd wrote:Does the addition offset the voltage, or did you do what I failed to mention last time, and put the thermocouple on the 'ground side' of the network, and read the series resistor's voltage on the 'high side'?
In the end, I did neither.

I took advantage of the fact that the output of the first stage of my amp is driven close to Vcc- when the input is open circuit. It more or less had to go one way or the other and it's about time I had a bit of luck. So, all I've done is add a simple inverter to the o/p of stage 1. I had to reverse a couple of breaks in the copper strips, and it was a tight fit to add the extra resistors and connecting wires, but it seems to work OK. The new output is close to Vcc+ in o/c, zero in short circuit, and small and negative (effectively zero to the ADC) if all is well. A true short circuit is identified by zero volts on all outputs.

The bonus in doing it this way is that there is zero impact on the rest of the circuitry.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Bingo, Richard! Good thoughts! As Ned says, there are lots of ways to skin cats, and some of them work pretty darned well!

<smile>
Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

I ran a few trials over the weekend with interesting but slightly puzzling results.

When I originally found a sudden fall in output at around the time of a grain’s explosion, I could only think of three possible (but fairly unlikely) causes. The first was that it indicated a real drop in temperature, suggesting some kind of violently endothermic event. I eliminated that possibility by adjusting the amplifier’s offset so that the quiescent output was at 2.0V. I still saw the output drop to zero, which would correspond to a physically impossible negative absolute temperature.

I was left with two remaining options – the thermocouple was either going open circuit or was somehow being short-circuited. The above result, showing an output lower than the quiescent (short-circuited, 0.0V input) one, suggested that the open circuit option was the more likely of the two. However, my newly modified hardware – specifically aimed at picking up the condition – fails to show any indication that the input is going open circuit. There also appears to be no difference in performance with either a simple twisted pair or one that has been impact welded. The balance of probability seems to have swung towards the short-circuit option, but there is enough conflicting evidence to leave some remaining doubt about the true cause.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Richard,
I almost posted this yesterday, then thought better of it, because 1) I don't have a solution, and 2) my explanation sounded 'silly' to me.

But... it's a theory.

First, I couldn't figure out how a thermocouple might 'short'. Some are actually brazed or welded together along their entire twist-length. But I realized that those that are, are soldered only at the junction contact areas; they're not 'fully surrounded' by brazing alloy. That might be the secret to what's happening here.

We know that superheated metal vapors (not even in a plasma state) can be highly-conductive. It's just possible that the cloud of vapors around the elements are providing a temporary short-circuit of the bi-metallic nature of the thermocouple. It wouldn't take much.

But... like I said; it was such a weird idea, I hesitated to post it.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

Lloyd, I'm just as confused.

Part of the calibration process involves dipping a junction into molten KNO3, which has to be a good conductor of electricity, and it shows no sign of shorting out in that case.

On the other hand, I've tried two different crackle compositions, one that produces a very sharp report and one where the explosion is gentler and softer. Both showed the dip in output, but it was wider and more pronounced with the more slowly reacting sample. That seems to agree with your suggestion that the effect is connected with the hot gases surrounding the thermocouple.

As I briefly pondered a little earlier, I'm wondering if the gases might be providing a short between cooler sections of the wires. Both metals are not especially good conductors of heat, they are very thin wires and the reaction only lasts for a few seconds. As shown in the picture earlier in this thread, a typical junction has around 5mm of exposed wires, and there is a further 5mm or so that is partially exposed. There is no sign of charring near these sections, so they clearly don't get that hot.

That wouldn't happen with molten KNO3 because everything in contact with the liquid would also be heated by it.

If I'm right, insulating the exposed wires might improve matters. To test this out, I'll just have to find a suitable means of applying an insulation that can survive the reaction.
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

"Both showed the dip in output, but it was wider and more pronounced with the more slowly reacting sample."
-----------
AH! That's a HUGE clue, Richard! I don't have much constructive help to offer, but your idea of insulating them might be spot-on.

If I might suggest an insulator -- a thin slurry of a ceramic dust with a 'minor' binder might do the job. Perhaps dip in the slurry, then gently heat it to dry/solidify/carbonize the binder, as required. (*eh... 'carbon' might not be the ideal thing! <grin>)

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
User avatar
richardh08
Site Admin
Posts: 2226
Joined: Wed Aug 07, 2013 1:54 pm
Location: Bedfordshire

Re: A silly (?) question to you chemists out there

Post by richardh08 »

What to use and how to apply it are indeed the issues. Dipping would be ideal, except I don't think I can do it that way. The thermocouples are too unmanageable and fragile before they are mounted, and afterwards I can't see how I could dip them without also coating the twisted section, which is something I want to avoid.

I have an idea, but it might turn out to be either stupid or impractical, so I'll try it before talking about it!
Even when I'm wrong, I'm convincing.
User avatar
Lloyd
Posts: 1902
Joined: Sun Aug 09, 2015 1:43 am

Re: A silly (?) question to you chemists out there

Post by Lloyd »

Richard,
I'm of the opinion that it is the twisted section you must protect from shorting... which would mean you must coat it.

I recommended a 'thin slurry' as a means of minimizing the mass of the insulation.

Lloyd
"Pyro for Fun and Profit for More Than Sixty Years"
Post Reply