A new ignition circuit

[jgedde]

John:

A small point but back in post #33 you should edit to correct your statement that the capacitor goes across the COIL (I think that was a typo). As you know the cap goes across the POINTS (and you've said that all along previously) The cap across the points makes the cap and coil a series resonant circuit of L1(coil) and C1(cap) when the points open (or the transistor turns off). From other posts I think that's what you meant to write. Certainly a cap ACROSS the coil would suppress any oscillations not enhance them.

It would be good to clarify this for those reading in the future looking for information.

As Lakc has pointed out a scope connection to the primary side is just as effective at looking at the secondary directly due to the transformer effect. In fact I believe your big ignition system analyzer connects to the primary side. As I remember them you could see all 8 cylinders at once as the same coil supplied sparks to all 8 in turn.
I always wish I had one of those. I once saw plans to convert a TV set to an ignition analyzer.

Sage

Sage,

Thanks for pointing that out. I went back and fixed it.

As far as the ignition scope is concerned, I got mine from a coworker for $50 which is what he paid for it at a flea market. It's still in its original box (including the corrugatred overbox) and is in perfect condition. We all get lucky sometimes, right? I've used it a few times, but not for several years.

John

 

[dsage]

RWO:

I forgot about that chip. I might have one of those chips in my ancient stock from "back in the day". Nice chip but I think they are obsolete. No matches on a search at Digikey anyway. Something like that for a hall sensor input would make things really simple. To do it any justice you'd need a trigger device with a differential output like the reluctor it's designed for though. Nice that it automatically controls dwell as well.

Sage

 

[mayhugh1]

Hi,
I might be able to shed some light on why a capacitor is needed for a mechanical points (Kettering) ignition system but not for a transistorized (non-CDI) ignition. When the mechanical points open in order to interrupt the current flow to the ignition coil and generate the flyback pulse, they are not ideal and do not switch instantaneously from zero ohms to infinity. They spend a small amount of time during their initial opening (as the spring pressure holding them closed starts to release) in the range of a few ohms to several ohms on their way to being an open circuit. While the points are in this low (but not zero) resistance area, the coil current continues to flow; and the IR drop across the points starts to disssipate the energy that was intended to generate the flyback voltage and produce the nice big fat spark at the plug. A lot of heat is generated across the points while this is happening and a 'plasma' which is conductive is generated and allows the current to continue flowing for a time even after the points have mechanically separated. This dissipation drains off the flyback energy and eventually burns up the points, and is the source of the visible arcing you would see if the distributor was transparent. If you put a capacitor across the points this capacitor will keep the voltage across the points at zero during the initial opening time of the points and even extend the current flow to the coil for a little while longer. The capacitor one would put here is for all practical purposes ideal and it does not dissipate any significant portion of the flyback energy. As the current continues to flow through this capacitor its voltage starts to rise after the points actually separate. If the capacitor is large enough there will be little or no arcing at the points to kill the flyback voltage or the points themselves. In fact it turns out that the current charging the cap is the same current charging the coil and the voltages across each of them are almost 180 deg out of phase; and so the energy oscillates between them for a short period of time extending the effective width of the flyback pulse. If you get to look at the flyback pulse with a scope you will see the remnants of the damped sinusoid on the back edge of the waveform after the plug fires and the bulk of the flyback energy is dissipated where it is wanted - across the plug gap. In a transistorized ignition (Kettering system with the points replaced by a solid state switch) the switch in this case is nearly ideal. It switches very quickly to interrupt the coil current and you can get an even higher flyback pulse without a lot of dissipation during the switching time. Of course the transistor, coil, and wiring all have stray capacitances and these and the non-zero switching time of the limit the flyback voltage to something other than infinite spikes. But none of these components are dissipative as are the mechanical points and so a big capacitor is not required.
I think your project goal of coming up with a universal bullet-proof ignition for model engines is a great idea. But, I'd like to offer one suggestion. The coil is an integral part of the design. I think the problem many of us have with model IC ignitions is that we mix and match ignition circuits with various coils and we get mixed results. There doesn't seem to be consistent source for universal model engine coils. And the primary resistance, inductance, and saturation characteristics usually aren't even spec'd when we do find one. (Most of us wouldn't know what to do with them if we did have them.) And they are very expensive. Ignition coils are difficult to reliably and consistently make because of the many turns of fine gage wire needed for the secondary and the vacuum potting that must be properly done so the cumulative effects of corona don't shorten their lives. The design parameters of the coil are intimately connected with the design of the engine. It would seem like designing an ignition circuit around a commonly available COP would be just the ticket for having an ignition that would be cheap, reliable, readily available, and be to a reasonable scale mode for model engine building. - Terry

 

[123RWO]

The MC3334 chip is $2.67 from http://www.aliexpress.com/wholesale/wholesale-motorola-mc3334.html

RWO

 

[jgedde]

Hi,
I might be able to shed some light on why a capacitor is needed for a mechanical points (Kettering) ignition system but not for a transistorized (non-CDI) ignition. When the mechanical points open in order to interrupt the current flow to the ignition coil and generate the flyback pulse, they are not ideal and do not switch instantaneously from zero ohms to infinity. They spend a small amount of time during their initial opening (as the spring pressure holding them closed starts to release) in the range of a few ohms to several ohms on their way to being an open circuit. While the points are in this low (but not zero) resistance area, the coil current continues to flow; and the IR drop across the points starts to disssipate the energy that was intended to generate the flyback voltage and produce the nice big fat spark at the plug. A lot of heat is generated across the points while this is happening and a 'plasma' which is conductive is generated and allows the current to continue flowing for a time even after the points have mechanically separated. This dissipation drains off the flyback energy and eventually burns up the points, and is the source of the visible arcing you would see if the distributor was transparent. If you put a capacitor across the points this capacitor will keep the voltage across the points at zero during the initial opening time of the points and even extend the current flow to the coil for a little while longer. The capacitor one would put here is for all practical purposes ideal and it does not dissipate any significant portion of the flyback energy. As the current continues to flow through this capacitor its voltage starts to rise after the points actually separate. If the capacitor is large enough there will be little or no arcing at the points to kill the flyback voltage or the points themselves. In fact it turns out that the current charging the cap is the same current charging the coil and the voltages across each of them are almost 180 deg out of phase; and so the energy oscillates between them for a short period of time extending the effective width of the flyback pulse. If you get to look at the flyback pulse with a scope you will see the remnants of the damped sinusoid on the back edge of the waveform after the plug fires and the bulk of the flyback energy is dissipated where it is wanted - across the plug gap. In a transistorized ignition (Kettering system with the points replaced by a solid state switch) the switch in this case is nearly ideal. It switches very quickly to interrupt the coil current and you can get an even higher flyback pulse without a lot of dissipation during the switching time. Of course the transistor, coil, and wiring all have stray capacitances and these and the non-zero switching time of the limit the flyback voltage to something other than infinite spikes. But none of these components are dissipative as are the mechanical points and so a big capacitor is not required.
I think your project goal of coming up with a universal bullet-proof ignition for model engines is a great idea. But, I'd like to offer one suggestion. The coil is an integral part of the design. I think the problem many of us have with model IC ignitions is that we mix and match ignition circuits with various coils and we get mixed results. There doesn't seem to be consistent source for universal model engine coils. And the primary resistance, inductance, and saturation characteristics usually aren't even spec'd when we do find one. (Most of us wouldn't know what to do with them if we did have them.) And they are very expensive. Ignition coils are difficult to reliably and consistently make because of the many turns of fine gage wire needed for the secondary and the vacuum potting that must be properly done so the cumulative effects of corona don't shorten their lives. The design parameters of the coil are intimately connected with the design of the engine. It would seem like designing an ignition circuit around a commonly available COP would be just the ticket for having an ignition that would be cheap, reliable, readily available, and be to a reasonable scale mode for model engine building. - Terry

Terry, I think you've done a terrific job of describing the situation and hit the nail on the head to be sure. That said, I think the dream of a more or less universal ignition coil circuit is nonetheless attainable. However, the coil characteristics and the engine dynamics are indeed unavoidable gotchas.

From what I'm finding both experimentally and in practice, the coil's leakage inductance is all important in predicting what a particular coil will do. http://en.wikipedia.org/wiki/Leakage_inductance

All that said, I am in hopes that a common circuit with some "cut and try", with regard to only the capacitor across the points (or the output transistor), may be necessary to achieve an optimum solution with a given coil and a given engine.

All in all, if the circuit can successfully drive any coil with only a capacitor experiment necessary to match the circuit with the coil, this leaves only the coil matching to the engine and the power source up to the builder.

I think what you were describing in practical, non-technical terms was the characteristic of an arc called negative resistance. http://en.wikipedia.org/wiki/Negative_resistance
Even after 25 years as an electronics engineer, the concept is non-intuitive for me, especially when it rears its head...

Cheers!
John

 

[jgedde]

The MC3334 chip is $2.67 from http://www.aliexpress.com/wholesale/wholesale-motorola-mc3334.html

RWO

RWO, I been meaning to experiment with that part just for the fun of it. I wasn't aware that was the basis for the GM HEI circuit. I wonder, though, how applicable the reluctor input is to us? On the other hand, the input circuit would be easily adapted to a hall sensor, at least from what the datasheet shows.

The only gotcha would be the fact that a reluctor's output voltage drops to zero inately when it stops turning where the hall sensor and magnet could be positioned in such a way as to cause the coil to be "stuck on."

That's not a good thing for both the coil, the output transistor, and the batteries. That's where dsage's as well as my circuit have an advantage - maybe... It looks like the IC chip has some sort of dwell time enforcement circuit in it that may come to our aid. I honestly haven't really dug to deep into it to say whether it fits the bill.

I can't help but wonder, though, whether a hybrid of the jgedde/dsage input circuit with the chip yield a kick@#$ solution.

John

 

[jgedde]

The circuit is built and working.

I needed to increase the value of the AC coupling capacitor to make it easier to test. I am getting a strong spark with no cap across the transistor. This lends a lot of creedence to Mayhugh's explanation since my coil with breaker points gave no spark without a cap.

Here is the final version... (see below)

Quiescent current draw (except for hall sensors) was 0.2 microamps when the coil is on. So, you could say I was successful in meeting my goal of low power consumption in standby mode. If it weren't for the hall sensors, I could even get away without a power switch!

Now the ugly... My engine has two Allegro A1102 sensors. One is used as the spark timing and another is used to inhibit the spark when the governor engages the exhaust pushrod. The current draw from both of those is almost 10 mA!

So now I need to ask the question: what good is a circuit that takes virtually no power when I have Halls that need to be powered and take vastly more supply current?

OK, it's elegant, and seems to be bulletproof, but simplicity is often better.

If you want to build it and use it, it's a working circuit. That said, it has little advantage over Dsage's circuit except for exceptionally low power consumption and a lot more circuit protections. His will undoubtedly work better from low voltage supplies (say 6V) due to no diode drop driving the IGBT.

Here's a link to Dsage's circuit:

http://www.homemodelenginemachinist.com/f26/ignition-circuit-help-19673/index2.html#post210702

John

View attachment Ignition Coil Driver.pdf

 

[jixxerbill]

Great work John. many thanks to you and Sage both.. Now whats the chances you could show me how to make a small coil ? lol... Something that could be put into a small base under an engine ? I know im a bum.. You work hard and finish a great looking project and all I want is more lololol.... Reminds me of ancient proverb ..Doing something good around here is like wetting your pants while wearing dark colored slacks!! You get a warm feeling all over but nobody notices..lol...Bill

 

[jgedde]

Great work John. many thanks to you and Sage both.. Now whats the chances you could show me how to make a small coil ? lol... Something that could be put into a small base under an engine ? I know im a bum.. You work hard and finish a great looking project and all I want is more lololol.... Reminds me of ancient proverb ..Doing something good around here is like wetting your pants while wearing dark colored slacks!! You get a warm feeling all over but nobody notices..lol...Bill

No worries Bill! I love that ancient proverb. I will remember that one!

The book "Ignition Coils and Magnetos in Miniature" by the late Bob Shores is a good source of info. The book is still available from Mrs. Shores at a very reasonable price.

http://www.bobshores.com/home.html
Scroll down to the bottom.

I haven't yet wound my own, but plan to. When I do, I'll photo-document the process. I got my circuit up and running with a small (1" dia, 1.25" long) 6V coil from J.E. Howell (available from his son Allen's website).

I also tried it with an MSD automotive coil: http://www.msdignition.com/product.aspx?id=5073&terms=8231 That one was just scary!

John

 

[dsage]

jixxerbill:

Have look for Bob Shores book on coils and magnetos. All the info you need to make a coil of your own is in there. I made a couple from his book but it's almost not worth the effort. Winding 10,000+ turns of #44 wire is a really difficult. Then there is the wax potting and vacuum source required.
There is a way to use the magneto coil from an old lawnmower. Apparently they work ok and are pretty small. There is a bit hack sawing and soldering required. I read an article on it but didn't save it so I haven't tried one. Maybe someone can enlighten both of us on their use. I'll try that next time I need a coil.

BTW you don't need extremely high voltage. An old school car coil can produce at least 25Kv if not a lot more. You'll probably only need about 3-4K - probably less for a model sized spark plug. In fact having 25kv is just asking for trouble and should be avoided. (Hence the problems you've already experienced).

Sage

 

[dsage]

John:

I guess we must have been typing at the same time but you won. - and as usual great minds think alike. Our recommendations are the same.

Re: your circuit. I wouldn't have thought the hall sensors would consume so much power just sitting there, but a quick look at the spec sheet confirms your observations. Surprising.

I guess it's a challenge to try to reduce power consumption but lets face it, your control circuit would run for at least 350 hours on a 7ah battery. I think it's about as good as it going to get.

BTW for those of you out there looking for a circuit to build, I would stick with building John's circuit over mine. Given that there is little control over the installation of the circuit in a users engine (proper grounding procedures, lead length etc.) the additional protection items in John's circuit will add a measure of reliability.

Sage

 

[dsage]

John:

One more small thing. The suggested values for LED resistors a IMHO are too low.
For 6v your 270 ohm resistor would set about 15ma for the LED. Very roughly (6-2)/270 = 15ma.

I have found that if you purchase the newer "High Bright" red LEDS with the clear 3mm package they CAN be blindingly bright at as low as 0.5ma - 1ma.
Sorry I don't have a part number but if you go on Digikey and look at the highest milli-candle (mcd) leds you can find in a 3mm package they work well.
Typically they are at least 2000mcd some are even higher. Check that their brightness ratings are at the same test current. They try to cheat a bit in the specs.
Red because they are the most efficient and 3mm because the die is smaller (less power) and the lens has a tighter pattern ( appears brighter).
If you use one of those you could probably get away with about 4k resistor for the 6v operation. (subject to the led chosen).

Sage

 

[jgedde]

Thanks for the advice. I reduced the values of the current limiting resistors to account for the additional drop introduced by the 100 ohm series resistor on the output of the hall. It's only there to avoid "spiking" the hall if there is a breakdown somewhere else in the circuit.

So, the 270 ohm is really 370 ohms, the 910 ohms is really 1.01k.

The LED has a Vf of about 2.1V (green) so this gives about 10 mA in the LED. If you like a red LED instead, the resistor values should be larger.

All that said, you're spot on regarding the new LEDs and the user could likely use much larger resistors if he/she didn't buy their LEDs from Radio Shack.

Being colorblind to red, I usually shy away from red LEDs since what looks bright to others isn't to me. Many years ago I was dating an X-ray tech. She took me into the developing room because I was interested in the process. She could see fine and was showing me how it works. I couldn't see a damn thing. At first I thought she was pulling a joke on me! As far as I could tell it was pitch black, but I was making funny faces at her and she knew exactly what I was doing!

John

 

[jgedde]

John:

I guess we must have been typing at the same time but you won. - and as usual great minds think alike. Our recommendations are the same.

Re: your circuit. I wouldn't have thought the hall sensors would consume so much power just sitting there, but a quick look at the spec sheet confirms your observations. Surprising.

I guess it's a challenge to try to reduce power consumption but lets face it, your control circuit would run for at least 350 hours on a 7ah battery. I think it's about as good as it going to get.

BTW for those of you out there looking for a circuit to build, I would stick with building John's circuit over mine. Given that there is little control over the installation of the circuit in a users engine (proper grounding procedures, lead length etc.) the additional protection items in John's circuit will add a measure of reliability.

Sage

How about a hybrid taking the best of both designs? In other words, add my protections to your circuit and call it a day. OK, so my circuit takes much less power. Does that even matter if we have to supply juice to the halls?

In addition, I have a diode drop in the gate drive to the IGBT - you don't. That means you'll have a somewhat beefier spark at low supply voltages

I'm running my circuit and coil right now from 4 AA cells. It works fine, but what happens as the batteries start to drop out? I'll need all the gate drive I can get.

I was considering a bulk storage cap to help lower the supply impedance, but a rough computation showed something like 47000 uF to be necessary to keep the voltage within 1V through a coil charge and spark cycle. That bloody cap is as big as my coil!

I am feeling funny about having my circuit recommended when I'll likely use a variation of your circuit in my engine!

Cheers,
John

 

[dsage]

10-4 on the 100 ohms you added. I forgot about that being in the LED circuit as well.

Re: the combined circuit:
Well, the way I look at it, they are variations of each other. I'm happy to recommend yours considering all the benefits you've added. There is a lot of merit in the additional protection seeing as you/we have little control over how people wire it up. Messy or long wiring, improper grounding etc. is going to be a big factor in how well it works or even if it will self destruct. The extra protection is good. Let the masses choose.

I'm not so interested in saving power. But it never hurts. I have a whole stack of NiMH and Lithiium Ion batteries from old cell phones and a lot of gel cells as well. Both my Atkinson engine and the Sage/Howell V8 have flat cell phone batteries at 1600mah hidden underneath. The Ford engine uses a 4Ah gell cell.
I could probably do better with some of the newer 2500mah+ Lion or NiMh AA batteries out there today. I assume that's what you are using.
I use what I've got around free.

BTW I assume you know about Super caps at a few farads. But they have really low voltage ratings.


http://davesage.ca

PS> I notice the spam generator has turned my word "cellphones" above into a link for cell phones.
Not my doing.

Sage

 

[metalmad]

Hi Guys
I hate replying to posts like this as I simply don't know anything about it, but this time I have a question!
Can a Hall effect circuit be made at home in the shed that has automatic timing advance?
Based on RPM for example??
Even I might be tempted with something like that:hDe:
Pete

 

[Lakc]

Hi Guys
I hate replying to posts like this as I simply don't know anything about it, but this time I have a question!
Can a Hall effect circuit be made at home in the shed that has automatic timing advance?
Based on RPM for example??
Even I might be tempted with something like that:hDe:
Pete

Patience young padwan. ;D
There are several designs out there already. Most appear to be CDI type with a microcontroller running the show. Eventually, I will have a design good enough to release here.
As several recent other threads have pointed out, its really impossible to have the "perfect" ignition circuit where the coil is not specified, its electrical characteristics are too important. Not to discount the work done here and elsewhere, this is a very fine circuit and should serve people well, but to design a modern bulletproof ignition system, you need the whole ball of wax in front of you. I applaud the work of John and Sage, both here and elsewhere, and hope to be able to make my own contribution sometime in the future.

 

[dsage]

Lakc:

Yes and we are patiently waiting for that CDI unit. I'd like to build one.

But holy smoke - what a can of worms to release that to the masses - unless you plan on selling pre-programmed PIC's and boards.

I have a nice circuit built but I'm not happy with the size and don't know enough about winding transformers to reduce it down. It's way overpowered for a model so it's on the shelf for now.

Sage

 

[tornitore45]

Late response, I have been offline for a while.
The CDI kit I mention was purchased at www.cncengines.com [email protected]
I have not utilized yet since the engine is not finished, but tested with a primary load rougly equivalent to the primary L+R, and got some waveform.
The design and modeling of a miniature ignition is quite a challenge, there are several parameters hard to define:
Behavior of the coil with a high turn ratio and high insulation lead to high leakage and the high frequency of a CDI oscillation give core losses hard to predict. Coil + Leads Capacitance is an integral part of the circuit and the circuit has two distinct modes of operations: initial rise to spark is "open circuit" (not really open since is loaded by the sec capacitance) after spark is practically a shorted load oscillating between leakage inductanca and capacitance.

 

[jgedde]

In case any of you don't know... Dave Sage and I are collaborating on a ignition circuit combining the best of his design and mine...

Here's where we're at... We have almost completed the documentation (schematics, wiring diagrams, etc). In addition, I've prototyped the improved circuit (two transistor version).

Here's a sneak preview: a video of it firing a full size spark plug... The coil is from J.E. Howell. Power is 6V from 4 AA alkaline cells.

[ame="http://www.youtube.com/watch?v=b8vc_2LUddE"]http://www.youtube.com/watch?v=b8vc_2LUddE[/ame]

John