Independence Model version...

Well, I braved the mosquitoes and measured the temperature of the air exiting the chamber where the Octenol is placed and it's just about 38-40 C, which I believe is within normal range. I also measured the air exiting the sides of the Magnet, after passing the heatsink, and it's just about 36 C, so nothing out of the ordinary. And this is with the thermistor indicating 200 C on the cat chamber. This IS probably on the high side as I probably didn't do as good a job on the thermal grease on the TEGs since this was a temporary setup, but in looking at my history, my normal cat chamber temperature for this ambient temperature is probably more like 180-185 C.

Digikey DOES have several 40x40mm TEGs that will operate at 200C, but they are at least $30 each. Not worth the expense unless you REALLY want the portability.

Here's some data I have on the Independence OEM thermistor. It can help you if you have to replace the thermistor and cannot find the exact replacement or if you need to translate the ADC voltage to temperature:

FWIW, the datasheet for the supposed OEM module on the MM 1000 says 271C solder, so 200C operating point seems reasonable. I expect the Independence/Liberty would need high temp like this too.

I have just yesterday ordered three of these 40mm 5W/5V TEG for experimentation. They are good to 330C continuous. I don't think they will have enough 'oomph' for the job (for the MM 1000), but we'll see what they do (I don't think we'll get to the optimal operating point). The problem is TEGmart don't have alot in stock right now.

Just for giggles, I also ordered some 66W TEC (Peltier) modules which are not really advertised as high temperature but boast a 'Inter melting point' of 200C - I assume that means the internal solder. I'm not expecting these to actually be suitable, but I wanted to do some testing with them.

> I have just yesterday ordered three of these 40mm 5W/5V TEG for experimentation. They are good to 330C continuous.

TEGmart does have some interesting modules. I think I'm going to do some actual measurements on my working unit, just to get a better idea of what they are actually producing. That way, I'll have a better idea of what I'm looking for.

> Just for giggles, I also ordered TEC1-12710 some 66W TEC (Peltier) modules

Ha ha, I did the same thing... we'll see how they turn out...

Sorry for my late answer:
> -If you connect the heatsink to combustion chamber the system not able to reach the necessery heat.
> No solid green -If you just let it loose (no real connection between the two) than in a couple of minutes you get
> an overheat (green/red)

I'm not sure I understand. You say that you have it running for 2 days now, but based on your statements above, it either doesn't go green, or it overheats.

I put back the faulty TEGs (just for the heat compensation) and tethered the system. In this way it runs smootly. I am really interested in the final TEG type solution...

I did some measurements today on my Independence with the working TEGs. Basically measured the current out of the battery:

- Upon startup, the battery (4 x 1.2V Nimh cells) measures out at 5.1V

- After ignition is detected, i.e. thermistor voltage dropping 0.2v, the fan kicks into full speed and I am measuring about 890ma of current consumption from the battery. This is about the maximum steady state of consumption of the unit. My controller draws about 100ma MORE than the PIC due to the WiFi module and faster Atmega chip, so the standard Independence probably draws about 790ma.

- As the unit heats up, and the TEGs start to come up to voltage, that current drain from the battery steadily reduces, but even after reaching steady state operating temperature (185C on this unit), it never drops below 14ma or goes negative, indicating some charging. At this point the TEG voltage is about 5.85V and it is producing (890-14 = 876) ma, so it's producing slightly over 5.1W, less than I estimated earlier.

So, with my controller+Wifi, the TEGs never produce enough power to fully power the unit. It's just a little bit shy but will probably still run my controller when the battery runs low, but at a slightly lower voltage. And in fact, it does continue to run when I remove the battery, but with a slightly lower Vcc driving the fan FET, you can detect a slight drop in fan speed. It probably handles the original PIC controller fine, with a couple of ma spare to top up the battery. This just points to how finely tuned the Independence was in terms of operating temperature, power consumption, TEG capacity, etc.

Does the Defender only measure air temperature?

The Defender measures exhaust temperature before cooling. The flow is: 11 inWC regulator, solenoid valve, nozzle and forced air intake, ~1" flame area, ~4" cylinder with catalyst mesh screens with the igniter inserted 3/4 way, thermistor, ~3" tapering exhaust chamber with fins cooled by the suction fan, exhaust tube with attractant, outlet.

The catalyst supposedly converts unburned propane and CO to CO2, and is thermally sunk to the chamber case, so there might be some cooling or heating from any catalytic reaction.

This is such a simple system. The purpose of the thermistor is to detect combustion failure and, rarely, overheating. The problem is that combustion failure has so many causes. Fortunately, none of them are that hard to troubleshoot.

My NodeMCU Defender also measures incoming air temperature and humidity. The combustion should cause a certain temperature rise, depending on humidity. This is a somewhat useful tool to measure the burn. At some point I will try to measure the combustion case temperatures, but, right now, it is catching mosquitos.

Temp=121.2C (0) at 5:171108 F=1 I=0 G=1 S=1 E=0 T=40.6C H=27% M=8896 (238/1011:5796) R=70~140 W=-73 B=1.0 V=3.4

Ok, thanks for the description. I finally found a picture of the Defender innards, so have a better idea. The thermistor seems to be placed AFTER the catalytic converter, whereas on the Independence, the thermistor is placed right at the middle of the converter. It is also pretty much one piece, whereas the Independence is comprised of several pieces screwed together. Since picture of the innards seem so rare, I'm posting what I have of the Independence here, starting with the overview:



Starting from the left, it's the solenoid valve and nozzle, feeding into the combustion chamber carburetor. The black plastic shroud around the burn mixing chamber gathers some air from the suction fan blowing against the finned heat sink to feed air into the burn mixing chamber. The burn chamber leads into the catalytic converter, which is where the igniter is located. It's a spark igniter that is inserted right into the converter. The thermistor is also fastened to the side of the converter. To the right of the converter, there is a cooling trumpet. It actually has two flows, one from the catalytic converter, and one from the finned heatsink. The flows converge on the downspout of the trumpet.

Here is a more detailed view of the burn chamber carburetor:



The left picture shows where the nozzle is inserted and screwed down. The plastic shroud is removed at this point. There are two air holes on the side of this chamber for air to feed in via the shroud. The right picture shows the other side of the chamber that feeds into the catalytic converter, through this mesh. In thinking about it, I now don't think any combustion actually takes place on the nozzle side. The igniter is after the mesh, and you wonder if the flame will pass through the mesh to the nozzle side.

Here is a view of the catalytic converter:



The left picture is the side where the gas feeds into converter. You can see the igniter screw on the upper left, and the thermistor screw on the upper right. They are pretty close to the 'burn chamber' carburetor side. However, you can see that there are no signs of combustion on the output side of the burn chamber or the input side of the catalytic converter, so I think all the combustion takes place within the catalytic converter. And if fact, if you look at the right picture, which is the output of the catalytic converter, you can clearly see signs of combustion. Hence you can expect the converter to be the hottest part of the unit.

(continued on next post...)



Edited 2 time(s). Last edit at 08/06/2020 05:50PM by Independence.

Continuing after the catalytic converter, it feeds into the cool-down trumpet:



The left picture shows the part of the trumpet that screws into the output of the catalytic converter. You'll notice that there is another pathway on the trumpet, that collects suction air from the finned heatsink and pipes it into the hot air path. The right picture shows where they converge on the downspout of the trumpet. Must be some finely tuned way to get the right temperature of the exhaust CO2.

That's all the pictures I currently have. What I don't have is pictures of is how the finned heatsink sits in the case, just above (or underneath, in normal orientation) the suction fan. But basically, the suction fan sucks in air from the top chamber, which is connected to the outside of the trumpet/funnel. This suction air is blown downwards on the finned heatsink, and then exits the side of the unit.

Edit: While I had the unit apart to install the new TEGs, here is a picture of the heatsink.



That should give you a complete view of the unit. The suction fan is underneath the heatsink (or on top, in normal orientation).



Edited 1 time(s). Last edit at 08/07/2020 04:13PM by Independence.

...you wonder if the flame will pass through the mesh to the nozzle side.

Good point, likely not. What would be the purpose of the mesh except to prevent the flame going back into what is more looking like a carburetor and not a combustion chamber? This sort of mesh is called a flame arrestor, used to prevent back propagation of a flame. I would guess that essentially all combustion takes place on the other side of that mesh. The input of the catalytic converter is amazingly clean looking, especially compared to the flame arrestor mesh in the first chamber. The catalytic convert exit perforated metal looks to be another flame arrestor, presumably to prevent flames from exiting the unit (and scaring the mosquitos).

Thanks for the photos. This topic deserves a new wiki page.

> .. prevent the flame going back into what is more looking like a carburetor and not a combustion chamber?

I think you're right. It IS where air and gas mix prior to combustion, so a carburetor would be the right term.

> especially compared to the flame arrestor mesh in the first chamber.

Does the Defender not have any flame arrestors in it's design?

> > Just for giggles, I also ordered TEC1-12710 some
> 66W TEC (Peltier) modules
>
> Ha ha, I did the same thing... we'll see how they
> turn out...

Well, I received my units yesterday, and they seem to be constructed about the same way as the other TEGs (i.e. cheaply). I bench tested them both in cool and TEG mode and it seems to function in either mode, though I haven't put them in the Independence yet for a full temperature test. My feeling is that they will also disintegrate as I've searched the data sheets and have seen different claims of operating temperature, but mostly 138C. Then I found this datasheet:



I guess many folks make this device and it really is a case of you get what you pay for. This particular manufacturer appears to offer different solder, sealant and ceramic optics. Mine has silicone sealant and for less than $9, I have no illusions that it has the 227 C CuSn solder option. The OEM versions don't have any sealant on the sides.

So, what to do.. I have no other purpose for these units, so I guess I will just install them and hope for the best. I will try and measure their performance before they get destroyed but I suspect they won't approach the OEM units. Yes, at this point, I have spent almost enough to buy the more expensive higher temperature units, but where's the fun in that? :)

I ran my standard test on these TEGs, which is basically measuring the battery consumption as they come up to temperature. In this case, I am running it on a unit that is powered by a PSU, but same difference:

- Maximum steady state power consumption on this unit was 750ma @ 5.85V. This is a little less than my other battery unit, which was 890ma @ 5.1V. Without actually measuring it, I can only surmise it's due to differences in the fan wear.

- The TEGs come up to voltage at about the same rate as the OEM units, even though these are supposed to be Peltier cooling devices. See chart below:



- The difference was that the maximum current from them maxed out at about 500 ma. Even at operating temperature, the 'battery' was still supplying 250ma.

- Disconnecting the 'battery' resulted in the fan slowing down considerably and the LED dimming, although the unit kept running. The ADC isn't accurate at this low voltage and I didn't have a meter on it, but from bench experience, I surmise that the TEG voltage is probably around 4V, which coupled with the 0.5V diode and regulator drop, results in an atmega voltage of only about 3.3V. The voltage booster is still producing about 9.5v at this point, but the atmega drive to the fan FET at 3.3v isn't enough to saturate it and thus the fan slows down. I don't have an original OEM controller anymore, so I'm not sure if the same thing will happen with it, or perhaps with the lower power consumption, these TEGs may actually work with the OEM controller (assuming they don't disintegrate in the heat)

- I will run these for a few days to see if I see any signs of degradation or differences in performance due to over heating.

- So I'm not sure. These TEC1-12710 devices may produce enough power to run an OEM controller, assuming you can get them with the higher temperature rating. I'm not sure how much performance varies between manufacturers, so as they say... YMMV

August 04, 2020: ..Here's some data I have on the Independence OEM thermistor.

Your very interesting chart shows three thermistors. The first is the 4365 Amphenol Thermometrics AL03006-111.3K-123-G1 replacement specified on this site, which came from the now defunct Mosquito Magnet Forum (available on the Wayback Machine). The second "4561" labeled, what is that? The third, labeled "OEM," has a much higher B25/85 value of 4623. Is this the device in your Independence? Did you measure this value yourself? What is the closest current part to the OEM device you have found?

I suppose I will have to re-calibrate my Defender. I used the specs for the Thermometrics part to calculate the display temperature. If a higher sensitivity thermistor is being used, then the exhaust temperature would be higher. I should also mention that the thermistor is not itself in the exhaust stream, but rather attached to a bracket on the outside of the metal case with a metal rod that goes inside. The rod and bracket are insulated from the case, but there must be a significant thermal impedance and temperature loss involved, which makes the indicated thermistor temperature lower. Perhaps these factors balance out...

On another note, the Wayback Machine provided this spec for the Liberty Plus TEG: Thermal Enterprises HT1-12710, $14.80 each way back then, which seems is in business on ebay, and they now sell "high temperature" 225°C modules for $30, "Only 7 Left." The ebay page has a paucity of specs, but perhaps these are the same units. Any why are they failing, anyway? Too many thermal on/off cycles? It may be more economical in the end to run the trap 24/7.

Also, looking at the picture of your destroyed units, did they come apart because of melted solder? Could the solder be replaced? They were made once, perhaps you could re-furbish them. Another project!

I was thinking of using the MM Executive net in my Independence, since it looks like it's been redesigned in a more rigid form, and fits better. In looking into it, I found this in the MM Executive manual:



The MM Executive has a fuel saver mode that can turn off the propane for up to 8 hours a day. The fan runs at a lower speed in this mode. However, the battery on this unit is still only 3000-3300 maH, and if it is running on battery for 8 hours a day, it is going to need to be recharged over the other 16 hours of running. Assuming the unit consumes about 300 ma in low speed fan mode, it will have consumed 2400 maH over this 8 hour period. Thus, it will have to charge that back in the remaining 16 hours of running time. To do that, you would need at least 150 ma of excess current from the TEGs. Based on what I measured from the existing OEM TEGs, that is borderline at full operating speed, so I wonder if they are now using higher power TEGs, or have sourced a more efficient fan.....

Another interesting item from the manual is that they state the middle of the day AND the middle of the night is the time of lowest mosquito presence. So maybe I should only run it during the extended dusk and dawn hours? We're just wasting propane in the middle of the day and middle of the night?



Edited 1 time(s). Last edit at 08/07/2020 05:49PM by Independence.

> The second "4561" labeled, what is that?

This is a thermistor I found on Digikey that is closer to the OEM characteristics than the KC-014G-ND. However, it's in a bead-like form factor and won't fit in the OEM thermistor housing.

> The third, labeled "OEM," has a much higher B25/85 value of 4623. Is this the device in your Independence?
> Did you measure this value yourself? What is the closest current part to the OEM device you have found?

The OEM B25/85 4623 is what I believe to be the device spec for the Independence. I can't remember where exactly, but I think I got that spec from some forum somewhere. I know I didn't come up with it myself. I DID verify it for myself though, as the last column on my sheet shows my bench measured values and they matched exactly for the lower temperatures. The problem is I could never find an exact part with those specifications. One of my two Independence has the original, while the other has the KC-014G-ND part. It occurs to me that I should modify my temperature limits (primarily the over-temperature limit) on that unit based on the thermistor differences, which I have not done yet....


> they now sell "high temperature" 225°C modules for $30, "Only 7 Left." The ebay page has a paucity of specs,
> but perhaps these are the same units. Any why are they failing, anyway? Too many thermal on/off cycles?

It appears that there is some reference design for these 12710 devices but folks manufacture them differently, using different temp solder, sealants, etc. The SP1848-27145 units were failing because of the low temperature solder. It is causing the leads to detach from the first element and also causing the top piece to seperate from the bottom piece. Now there is a chance that if you leave everything anchored in place, they may continue to work, since they are effectively compressed together and as long as the leads don't come off completely, you may be alright. I will be testing this on my current set of TEC1-12710 TEGs, even though they don't generate enough power...

> Also, looking at the picture of your destroyed units, did they come apart because of melted solder?

That would be my guess.

> Could the solder be replaced? They were made once, perhaps you could re-furbish them. Another project!

I did consider that as I have some high temperature CuSn solder. But some of the elements have completely detached and for $8-9 each, it doesn't seem worth the effort. I am mostly doing this TEG experiment as an exercise anyway as I run my main Independence under corded power. I have a spare one with OEM TEGs that will run (mostly) on battery, but the propane saver and wifi mode more important to me than the portabilty aspect.

And BTW, in installing the TEC1-12710 TEGs, I properly applied thermal grease and that unit is now back to a normal operating temperature of around 170-185C, based on the thermistor reading.

Update on the TEC1-12710 TEG devices:

> The difference was that the maximum current from them maxed out at about 500 ma. Even at operating
> temperature, the 'battery' was still supplying 250 ma.

After leaving them running for 5 days, while they are still running, they are generating 40 ma less than when I first started. The 'battery' is now supply 290 ma during normal operating temperatures, so they seem to have degenerated a little. I have not taken it apart to see if anything has melted, but will monitor it to see if any further degeneration occurs...

@Ziggy, what has been your experience with YOUR TEGs?

Quick question: are these TEGs running at their max power output for the temperature difference? I understand that these devices have a voltage vs current characteristic, and that an optimum load yields maximum power. It might be beneficial to place a high efficiency step up or step down converter between the TEGs and the load, and adjust the output for maximum power at operating temperatures.

Unfortunately I've been swamped at work and not had any time to do further testing. Both sets of devices have arrived though, so it's a matter of getting through this busy period and having enough free time to get start tinkering again.

TEGs do have a operating envelope and definitely have an optimal operating point. It is governed by three things: Hot side operating temperature, hot side/cold side temperature difference (delta T), and load resistance. To get maximum output it is necessary to get all three right. This info is usually in the datasheet. But you'll find that while this info is provided for TEGs, it is not provided for Peltiers - so identifying the proper operating point for a Peltier operated as TEG may not be easy.

> Quick question: are these TEGs running at their max power output for the temperature difference? I understand that these
> devices have a voltage vs current characteristic, and that an optimum load yields maximum power.

I don't have a data sheet for these TEGs that shows power generation characteristics. I do believe you are correct in that there is an optimum load that yields maximum power, much like the MPTT function for solar panels. However, given the degradation that I've seen so far, I'm not motivated to do much work on these cheap units, especially since I don't really need the function on my main MM. I'll just leave them in as part of the thermal regulation design of the unit and see what happens over time.I suspect that the next time I take it apart, the results won't be good.

> It might be beneficial to place a high efficiency step up or step down converter between the TEGs and the load, and adjust the
> output for maximum power at operating temperatures.

I did go so far as to gradually reduce the 'battery' voltage on the unit to see what happens to the voltage and current on the TEGs, and if I remember correctly, the current remained about the same (about 500ma at that time) but the voltage kept falling till it reached a low level that I can't remember. At that point, the fan had slowed down so much that there was not enough air in the combustion mixture and the temperature of the cat converter started to drop, resulting in a downwards vicious cycle. Again, this is with my controller that consumes about 100ma more than the OEM, but just comparing to the OEM TEGs, these TEC1-1270 devices still come up short, as least as a drop in replacement...

To get maximum output it is necessary to get all three right. This info is usually in the datasheet. But you'll find that while this info is provided for TEGs, it is not provided for Peltiers - so identifying the proper operating point for a Peltier operated as TEG may not be easy.

Easy, no. But you might be successful finding an optimum step up/step down converter "one knob" output voltage adjustment for an arbitrary TEG with the system at operating temperature running into an average static dummy load, or the real system. Of course, the real system seems to have a lot of constant current pieces, the PIC, the step up converter, etc., in parallel with that diode and battery with an arbitrary charge and charge capacity! Still, if you are trying to get close to the original TEGs, which have a certain number of elements in series, with another device that has a different number or type of elements, then perhaps some sort of conversion might be helpful.

> wonder if that would be a good replacement for the hot surface igniters in other models. Encapsulated 12V
> generators are very inexpensive..

It appears that all battery powered MMs (Independence, Liberty Plus, Executive) utilize spark igniters while all tethered MMs (Defender, Liberty, Patriot) utilize the cheaper (but less reliable) hot surface igniters. However, in reading Dev's blog on all the issues with hot surface igniters, it certainly seems worthwhile for someone to start a project to convert those igniters to spark ignition. They are certainly inexpensive enough:



The above spark igniter is powered by just a 1.5v battery and all you would need to do is to modify the Defender/Liberty/Patriot igniter driver to switch power to the unit instead of driving that MOSFET. It could be as simple as driving a 24V (?) relay with that MOSFET instead of the hot-surface igniter. In looking at the schematic, you may also have to spoof the IGNL input to make the PIC happy. As for the sparking point, on the Independence, it sparks from the tip to the inside of the cat converter, which is connected to the return terminal of the igniter. Not sure if that setup would work for the Defender/Liberty/Patriot but you should be able to cobble something up to allow it to spark inside the chamber.

I don't have a Defender/Liberty/Patriot otherwise I would try it, but perhaps someone else can give it a go...

someone to start a project to convert those igniters to spark ignition

I am sorely tempted to try this out, since I am now out of the expensive igniters. One igniter costs the same as about 3 tanks of propane. However,

I am concerned about running those devices continuously for a period of time. I don't have the specs for this device, but gas grills run the ignition for only a second or so. The NodeMCU design keeps the igniter on until the thermistor (exhaust) temperature rises 5°C, which takes about 60 seconds, perhaps more. But it needs to stay on until the system is sure that combustion is viable, then it waits an additional period after turning the igniter off for the thermistor to reach the minimum operating temperature. I forgot how long the original software keeps it on, but it might be this entire time, or somewhere in-between. Perhaps three 1-3 sparks on followed by 2-3 seconds off would work and not overheat the module. This would require a modification to existing DIY software, or a separate custom driver for stock traps.

The unit is a bit large, and requires a battery substitute to supply 1.5v (which could implement the duty cycle control) given the 12 volt input.

If the high voltage connections "come loose," it could send an arc into the controller, or me, causing problems.

The case igniter hole is just under the igniter Molex connector. The low-profile silicon nitride igniter cable just misses the Molex as it is. Longer ceramic insulator posts would hit the connector.

I am unsure how to bend a single electrode (the other lead attaching to the case) so that it arcs to the catalytic converter case, which is sort of attached to the metal case. I don't want it arcing to the mesh catalyst, possibly punching holes in it, but there is no way to tell where the arc is going.

Perhaps a dual electrode in one shaft constructed so that the arc is confined to between the two electrodes fed from this device's balanced output of would be OK, but I haven't found such an electrode, and don't know how to make one. Perhaps a miniature spark plug? Here is a link to a fellow with directions on how to make a custom miniature spark plug.

dev Wrote:
-------------------------------------------------------
> I am concerned about running those devices continuously for a period of time.

I found these pictures and schematic of a similar handheld gas igniter on the Internet. It was being converted to use as a spark generator for a homemade engine.



Looking at the board and schematic, it looks very similar in design to the one in the MM. Two transformers, a transistor or two, and some diodes. I believe that bulge in the original Amazon picture is the large HV transformer, so I think it is basically the same design.

For the Independence, IIRC, this circuit is turned on for up to 45 seconds, or until it detects a rise in temperature via the thermistor. In my experience with the original controller, it typically sparks for about 15-20 seconds before the detection of ignition. In my implementation, after the solenoid is turned on, I leave the igniter enabled for 15 seconds and then I turn it off and wait up to another 45 seconds for the thermistor to drop 0.2v, indicating that ignition has taken place. So far, I've never had it NOT ignite everyday, other than stupid things I did like forgetting to turn on the gas, or running the fan too fast.

I understand that the hot-surface igniter is left on for longer periods to guarantee ignition, but I think the spark ignition is much more reliable and doesn't need to stay on that long. But having said that, I don't think there is any downside in leaving it on longer. It's high voltage, small current and I don't think it gets warm, though the best way to find out is to invest a few bucks and buy one to experiment with.

> The unit is a bit large, and requires a battery substitute to supply 1.5v (which could implement the
> duty cycle control) given the 12 volt input.

I think when you remove the battery/switch mechanicals, you end up with something the size of the unit in the picture above, which I don't think is that big. I don't know how much current it takes when sparking, but it might be possible to do it via resistor dividers or a linear regulator. It might also be possible for the circuit to work at a higher voltage with just minor tweaks.

> If the high voltage connections "come loose," it could send an arc into the controller, or me, causing problems.

As mentioned, it's very low current and the arc has limited range. At least, that's the case for the Independence, and I expect that these gas lighters are probably similar or lower current. I did get zapped when my insulation was broken and it wasn't much more than a sting.


> The case igniter hole is just under the igniter Molex connector. The low-profile silicon nitride
> igniter cable just misses the Molex as it is. Longer ceramic insulator posts would hit the connector.

That might be the biggest challenge. On the Independence, the igniter electrode is just a small rod on a ceramic insulator. I'm not sure if it is a special material, but the wire from the igniter doesn't appear to be even soldered to it. Just being in close proximity is enough for it to spark. The igniter doesn't stick in that deep into the case, so I don't think there is any danger of it arcing to the platinum wires inside, which I don't think is connected to the case anyway. The path of least resistance would be the aluminium case, so I think you only need to insert an insulated conductor of some sort into the cat case, deep enough so it can light the gas, but not so deep that it can't spark to the case.

So, for completeness, I decided to trace out the schematic of the igniter board on the Independence. It is shown below:



Although there are slight differences in actual circuitry and implementation, they all appear to work the same basic way. I'm no expert in analog circuitry, but I believe the basic premise is as follows:

The first stage is an oscillator circuit that works in conjunction with the first transformer to generate a stepped-up voltage to charge the 1.5uF capacitor in the 2nd stage. When the capacitor reaches a certain voltage, the SIDAC (K1500G) device turns on and discharges it, resulting in the generation of high voltage on the secondary side of the 2nd transformer. This cycle then repeats as long as the IGNITER pin is pulled to ground by the transistor on the controller board.

This circuit generates a pretty decent spark and can bridge about a 1/2" gap. It generates about 4-5 sparks per second with an input voltage of 5 Volts. I ran this on the bench for several minutes and it consumed an average of about 20 ma while working and nothing got warm. Given that the gas grill igniter looks to be a similar design, I think there should be no problem using them as a replacement, assuming that the spark is sufficient to jump across whatever gap your igniter tip has to the case. The sample gas igniter circuit uses a 0.47uf capacitor compared to the 1.5uf capacitor on the Independence, so it's spark is likely less, but hopefully still sufficient to light the gas...

Note that the pin-out of the 4 pin connector on the igniter board and the controller board are not the same! On the igniter board, V+ and IGNITER are swapped and the connecting cable takes this into account.



Edited 1 time(s). Last edit at 08/20/2020 07:03PM by Independence.

Your efforts made me think of a potential spark electrode: the original igniter!

Just cut the top off. It has to arc somewhere. And, using the balanced module you found, shouldn't cause too much trouble with RFI (to mess up the wifi) or nasty ground loop induced glitches that could disrupt the CPU.

Your circuit description was very good. I am glad to hear the circuit didn't get warm, and drew only 20ma. Powering it from 12 volts will cause quicker charging and sparks (and stress on the 1N4007), but a series resistor should work to drop the voltage, and adjust the spark repetition rate. The spark energy is fixed by the K1500G breakdown voltage.

dev Wrote:
-------------------------------------------------------
> Your efforts made me think of a potential spark electrode: the original igniter!

You know, I was going to suggest that, but all the igniters I saw for the MMs on eBay had a single combined rod. That OEM one is perfect. Question is whether the resistance makes a difference. I don't think it will and think there is a pretty good chance it will work fine.


> And using the balanced module you found, shouldn't cause too much trouble with RFI

On the Independence, the control wire that runs from the controller to the igniter board is wrapped around a ferrite. Probably a good idea to do the same from your controller board to the igniter board to minimize the RFI coming back.

> The spark energy is fixed by the K1500G breakdown voltage.

Which is 150V, but I would have thought the size of the spark would be determined by the size of the cap, since that's where the energy is stored. Larger cap discharging = larger spark on the secondary side?

I guess this means you're getting one of those igniters and proceeding?

size of the spark would be determined by the size of the cap

The energy stored in the cap is the product of the voltage squared and the capacitance.

The "size" of the arc (the longest air gap it can jump) is determined by the output voltage, which is the 2nd stage voltage times the transformer step-up ratio, less losses. The energy into the transformer comes from the cap charged to a certain, fixed voltage, less losses in the semiconductor switch. A larger cap means more energy into the spark as the spark conducts. The voltage is fixed because, at a certain voltage, the switch suddenly triggers and conducts, discharging the cap with that fixed voltage into the coil. At that instant, the voltage across the transformer primary equals the cap voltage less losses, so the output voltage is also fixed. The cap and transformer are pretty ideal components, with stable losses, and their characteristics don't change much with temperature, etc. I expect the semiconductor characteristics change with temperature, but since it doesn't heat up vary much, it doesn't change a lot.

Assuming the output transformer has reasonable characteristics, a larger cap gives a higher energy, longer lasting arc, and more ignition heat. This may be important to ignite the propane. On the other hand, the transformer itself may be a major limiting factor. I don't know how much arc heat is required to ignite the gas flow. My guess is that it is the same as in the Independence.

...to minimize the RFI coming back.

Can't tolerate much RFI at all. It has an on-board wifi chip that is pretty good, but I don't want to mess it up, since resetting it means a trip outside and a few bites. This thing may require shielding, and insulation from the shield. Steady wifi is most important during startup to diagnose which of the many, many reasons the trap fails to start.

I guess this means you're getting one of those igniters and proceeding?

I don't know. I would have to write or think about a requirements spec, hazard analysis, etc., considering about what might happen if others try this at home, and other things. For example, it has to work when it is very damp or wet, clogged with mosquito parts, and hot or cold. Quite a project!

dev Wrote:
-------------------------------------------------------
> The energy into the transformer comes from the cap charged to a certain, fixed voltage

Gotcha... I guess I meant more energy, but as you said, that manifests itself in a longer lasting arc. Thanks for the explanation...

> Can't tolerate much RFI at all. It has an on-board wifi chip that is pretty good, but I don't want to
> mess it up, since resetting it means a trip outside and a few bites.

I'm surprised to hear that. I have an atmega controller AND the esp8266 wifi module all connected to the igniter board and have not seen any interference. I haven't even bothered to turn on the watchdog timer on them since I've not had any lockups yet.

> I don't know. I would have to write or think about a requirements spec, hazard analysis, etc.,
> considering about what might happen if others try this at home, and other things.

I can't decide if you're being facetious, or if you have a NASA type job that forces you to consider these things, even for a home project :)

> For example, it has to work when it is very damp or wet, clogged with mosquito parts, and hot or cold. Quite a project!

But well worth it to get rid of those unreliable hot surface igniters... it's like going from points ignition on your car to electronic ignition. I can tell you that gas ignition has never been a problem on my Independence, through thunderstorms, heat wave, cold waves, etc.

If I see a free Defender or Patriot on craigslist, I may try this myself...maybe I've been spoilt by the Independence....

>I don't know. I would have to write or think about a requirements spec, hazard analysis, etc.,
> considering about what might happen if others try this at home, and other things.

I can't decide if you're being facetious, or if you have a NASA type job that forces you to consider these things, even for a home project :)

I do. A home project, yes. But it involves some hazards including the propane, CO2, CO, and, now, high voltage. The metal casting can withstand the propane ignition explosion, but the plastic case is not as robust (my Defender case top is no longer even screwed to the base). A propane leak into the case, a (cracked high voltage lead), and then what? Might be OK, or not.

I don't want anyone to suffer injury or damage connected to this site. This site's visitors are not experts, nor am I, with lots of unknowns, known and unknown. Indeed, my trap has failed so many times, for so many reasons, including component fatigue from repeated disassembly and reassembly, weather, errors on my part, etc.

I did publish a design and construction details for the Defender on the wiki. There is a new design with more functionality, but I haven't had the time to update the docs, and the docs for the previous design were very daunting. See the forum post and the blog entry.

If, after reading through the above links, you can offer suggestions for how to deal with these issues, that could be a great benefit to the DIY community. This should be discussed in a new topic, however. Thanks for your interest.

That said, with this much time into it already, it seems silly to not now order the part. I may not do much with it. The first experiment would be to set it sparking and look at the RFI using the spectrum analyzers. I don't want to disrupt my fringe TV reception. Then to on top of the trap, and see what happens.

dev Wrote:
-------------------------------------------------------
> I don't want anyone to suffer injury or damage connected to this site. This site's visitors are not
> experts, nor am I, with lots of unknowns, known and unknown.

For sure, but I see this place as just a forum for folks to share experiences and ideas, not a de facto reference site on the subject. Your Wiki probably falls in a grey area. My goal on this forum has always been to share what I've experienced and learned (and also learn from others), not for someone to blindly take it at face value and run with it. That is why I won't post board layouts or code. I WILL share it privately with folks, but with the understanding that they are using it at their own risk.

> If, after reading through the above links, you can offer suggestions for how to deal with these issue

I'm not sure you can cover every possible occurrence. I do think that the risks involved in controlling the MM is small as long as you factor in the safeguards. But as you said, that is a subject for another topic, not here.....