flyback transformer ignition
 

Has anyone tried to develop a better ignition system based on flyback transformers? It can be designed for a higher voltage than a standard ignition system. The spark plug gap is then widened, so the voltage and therefore the energy stored in capacitance is increased.

I've heard that someone tried to install high voltage capacitors in parallel with the spark plugs, except the timing got thrown off by the charging delay and the engine ran very poorly. But if a flyback transformer with a higher current output is used and specialized control circuits installed, couldn't that problem be solved?

Is the goal to try to get a larger spark? What advantage does this have?

It would become apparent when you do a simple experiment with a flyback transformer and a small high voltage capacitor. Without the capacitor, the spark is weak. With the capacitor, the spark becomes brighter and louder. And the longer the spark, the more energetic the spark becomes provided the flyback transformer can supply enough voltage.

I built a high energy system with a follow-on energy supplied by a capacitor. The capacitor discharged to prolong the spark, no initial delay. The plug made very loud pops and bright light. The follow-on supply that charged the capacitor to 300V could be switched off to test, with and without. The test car was a VW Beetle air-cooled engine. The stronger spark, resulted in a better idle during warm-up. The negative - plug wear was very significant!

I went on to build an Engine management system and used waste fire for simplicity.

KitCarlsonEMS Homepage

More info another post in this forum.

Here a picture of the spark system used to test on lawn mower engine. Note HV supply not shown. See small fly-back coil (the type used for flash ignition).

Did you consider using platinum plugs to offset the increased plug wear? Many seem to enjoy 100,000 miles without problem.

A conventional plug lasted about 1K miles. The electrode was rounded and the gap increased about 0.01". I think the energy is sufficient to vaporize a Bosch fine wire platinum in a few sparks. I did this over 10 years ago. This was before other platinum plugs. The spark sounds like a small fire cracker going off. Visually the spark is a white ball of light about 0.25" in diameter. I could have reduced the capacitor value for less energy. The seat of the pants dyno showed slight improvement only under condition such as too rich or lean of fuel mixture. The complexity of the system made me think, reliability could be poor. I also worried about burning a hole in a piston.

Try a ball nose on the plug. Plugs seem to to want to wear to this shape so you could just form te tip to that shape.
I wouldn't worry about burning a hole in the piston. Detonation will cause that. You may find yourself having to adjust the timing to take care of the fact that the fuel is igniting that little bit quicker.

ollie

If the spark sounds like firecrackers, it is probably too powerful.

I agree, more is not always better. If I ever get a dyno, I will know for sure.

I've thought about this for some time, but never got around to trying it. My concept was to use a low frequency RF/high frequency audio signal to produce an AC spark instead of the single pulse that is produced by conventional ignition coils.

The proposed benefit was to have a spark that would start at the same time as a conventional one, i.e, around 8 degrees before TDC, but which would then persist until the exhaust valves open. This continuous spark would insure that any combustible mixture that passed the spark plug's electrode region would be ignited.

The frequency would be something like 100kHz or so, or whatever was most efficient for the transformer.

As far as putting capacitors in parallel with the sparkplug gap, there is a product called Pulstar (sp?) plugs that do exactly that. The key to avoiding ignition timing problems is to limit the capacitance to 80-100 pF and using non-resistor wires.

I think you might need a big amplifier.:)

Spark advance is often more than just 8 degrees. Maximum advance of 30-50 degrees is typical. It depends on many factors, including RPM, load and temperature.

It doesn't take much voltage or power to strike and maintain a spark across a gap of .04-0.1 inches. 500 watts is easily obtainable with a standard H-bridge driver circuit, and should be more than plenty.

The advance I mentioned was just off the top of my head. You get the idea, anyway... :D

At 100KHz? I would first try 10KHz, that might be the near spark resonant frequency. You are correct, once the spark ionizes the gap, the voltage requirement is low.

...And they don't appear to have any real effect on fuel efficiency, as tested by an EcoModder forum member... (Perhaps they would help ignite a very lean mixture? But that would require more mods than just swapping spark plugs.)

-soD

100 kHz was off the top of my head, as well. That's not critical by any means, but you'd want to avoid frequencies that could interfere with radio communications. LIke submultiples of 455 kHz, for example.

One factor would be to choose a frequency that would have at least several full cycles during the spark 'ON' time. For example, an engine at 6000rpm turns at 100 revolutions per second. If the spark was designed to operate for just over half a revolution, that's 1/200th of a second (5 milliseconds). I guess at 10kHz that would be 50 cycles, well over what I'd think would be a minimum, so that would work. The transformers are pretty efficient over a wide frequency range, so you'd have a lot of latitude.

I would definitely use coaxial cables for the spark plug wires in such a system, to prevent RFI and to transfer energy as efficiently as possible.

As far as 'spark resonant frequency' goes, I can't imagine the system would have a high enough Q to make resonance effects significant. The spark plasma itself is a fairly lossy high power resistor, and that would load the circuit enough so that resonance would be pretty broad.

Unless you have something else in mind by that term?

The Pulstar plugs evidently cause a faster growth of the flame kernel due to the intensity and speed (risetime?) of the spark. That's what the pictures at the manufacturer's website seem to show.

Supposedly the faster flame front means more complete combustion. If your engine is already burning the mix thoroughly due to good turbulence, homogenous mixture, and so on, this plug might not make a noticeable difference. I bet my old iron head Harley would pick up some benefits...:turtle:

I love spending time in the lab, bringing reality to ideas.

Me too! :D

It would be fun to work up this idea to a testable unit. I can make a MOSFET H bridge driver for a flyback transformer. I made a similar unit for an experimental induction heater a year or two back. A timing/triggering unit shouldn't be too hard to design, either.

Maybe a signal tapped from the cam position sensor, and some minor computing to derive an adjustable spark timing? How much overlap between cylinders (for the spark cycle) is there? Would you be able to do it with two transformers, or would you need one for each cylinder?

( I guess I could figure the answers to these questions out myself, but I'm feeling lazy this morning... :D)

Actually a lawn mower might be a better test bed, so the engine that blows up on version one isn't that expensive to replace. :eek:


Kit- e-mail me at engr@coloradoradio.com if you want to discuss this in a side conversation. Use the subject 'flyback transformer' so I can spot it through my spam filter.

I don't think they heard you some_other_dave :D

Maybe if we replace the fuel pump with a 10,000 psi pump? :rolleyes:

Hey, the transformer/lots of sparks thing sounds interesting and different. All I was commenting on was the specific "Pulstar plugs" product, not the whole notion of improving the ignition.

I suspect that any gains would be very minimal, but definitely worth (someone who is not me) checking into! ;)

-soD

A H bridge is overkill for this application. Use push-pull for direct off 12v or half bridge for operation from a boosted rail (100v and above).
I'm not aware of any radio that operates at 455kHz.
Maybe use copper core ignition wire in a plastic tube for additional insulation?

I think the whole idea would be most effective at high compression ratios (like in an engine designed for ethanol) and lean mixtures.

I heard him. :D That's what I was talking about in the quote below from my earlier post.

I'd use an H bridge for reliability's sake. The load on each mosfet is half what it is in a half bridge. There are dozens of circuits that would work, but reliability on a timescale of thousands of hours operation is something else.

455 kHz is the intermediate frequency (en.wikipedia.org/wiki/Intermediate_frequency) of nearly all AM radios. A signal near that frequency will bleed through and hash up reception over the entire AM band.

Ordinary RG-8 will stand off pretty high voltages, if you use the solid polyethylene dielectric stuff. You need the outer conductive braid of coax cable to provide a return path for the RF, to prevent your wires from becoming antennas broadcasting RF noise. The RF thing is why resistor wire is used for most spark plugs, if I understand correctly.

You could take plain copper braid (recycled from coax) and insert spark plug wires inside to get a fairly good system. I've considered doing this to improve my AM radio's DX capability, but I think the injectors and fuel pump contribute as much if not more hash to the band.

RFI (radio frequency interference) is a very big deal.

A H bridge ("full bridge") has more parts than a half bridge and is actually more likely to fail. We're dealing with peak powers of a few hundred watts at most at a low duty cycle. The average power will be quite low, probably 80w or so at most for all the cylinders.

I don't think there'll be too many issues with EMI. The metal of the car would make a Faraday cage. Remember that arc welding equipment is orders of magnitude more powerful than what we're dealing with, is (for the most part) unshielded, and yet it still isn't as much of an EMI problem as you would think. At one site (a metal shop classroom), the simultaneous operation of several arc welders in nearby welding booths (which are contained only with fireproof fabric, no EMI reduction at all) had no effects on the wireless network connection for a laser CNC machine just 15 feet or so away.

Sorry, but I think we are kicking a dead horse. I have already done this. I will reinterate. One of my prior posts shows the ignition system, less the 300V DC and the 12V used in conjunction with the point signal. It ran a lawn mower engine. It used the flyback principle desired in the first post. It made a very compact Coil On Plug ignition. No coax required!

I then used a conventional ignition to trigger ignition with a local capacitor follow-on for a VW engine. The follow-on supply could be switched on and off for test purposes. There was not a noticeable improvement in engine performance, unless the engine was running richer than desired. There was no delay in spark due to the means of injecting the follow-on current in either case. The current delivered was at least 10 times that predicted by the Pulstar plugs. This was an attempt to push the limit looking for performance improvement. The negative part was the advanced plug wear. It may be possible to ignite leaner mixtures, however an air cooled engine is not a good test bed for that. There are also numerous dangers in running lean with water cooled engines, without modification.

Modern conventional waste-fire or Coil-On-Plug ignitions are acceptable and reliable means of ignition. Proper dwell control, minimizes energy in, and maximizes energy out. The energy in, to run the ignition comes from the battery, supplied by the alternator via the engine. Optimal spark energy with minimal energy in, yields FE. Both systems are also relatively simple, light weight and reliable.

I highly doubt current ignition systems are "optimal". First off, the high resistance wiring would waste power. Just replacing that wiring with low resistance wiring would reduce those losses, although I'm not sure if it would overload the transformer or other circuits. The ricers have no problem when they do it for horsepower, however.

Have you tried to see what happens if the gap is widened and the voltage increased?

Not sure about the high resistance wiring. I see the grounding kits for ricers. I do not think they do much. The spark plugs ground at the block, the battery cable ground is at the block, not much lower resistance than that. My measurements suggest spark current is under 0.5A. On the coil primary side dwell time sufficient enough to fully charge the coil prior to fire is all that is necessary. Some ODBII systems have miss fire detection, and most monitor coil charge and adapt dwell for battery voltage and other factors. Some resistance is good for helping reduce EMI.

Increasing spark gap increases voltage requirement and energy. Too high a gap leads to failure of the ignition system components. Failures due to carbon tracking of high voltage components, and of semiconductor used to drive coil(s).

Kit- it sounded to me like your system was using a capacitor to peak a pulse, not to create a sustained AC arc over 90 degrees of crankshaft rotation.

It isn't necessary to send a huge amount of energy or power through the sparkplug; as you've mentioned, that causes extreme wear. The idea of using a flyback transformer (at least in my idea) is to create a sustained spark to make sure that the fuel mixture has been exposed to conditions that can burn any burnable mixture that is in the cylinder.

As far as H-bridge vs fewer components, it's a wash in my opinion. The minimum component solution isn't always the most reliable one.

RF interference is very relevant to me, as I'm both a ham radio operator and a broadcast engineer. Your example of arc welders not interfering with the wireless connection in a shop is valid, but wireless networks are pretty robust with respect to random RF, particularly as arc welders deposit most of their RF energy quite a bit lower in the spectrum.

Really, though, RFI isn't an issue unless someone was to attempt to commercialize the technology.

jim-frank,
The capacitor supplied enough energy to sustain arc over the combustion cycle. By viewing the current signature with a current transformer, I verified this. I also later found information that Saab ignition, does something similar. The sparkplug is used as a sensor.

I do not mean to insult you, however reading your posts suggest there is more for you to learn about sparks. I fail to completely understand your means and theories. I do not consider myself an expert, I am always learning and experimenting. I do have a BSEE degree, to do so I passed courses in electromagnetic field theory (courses that kept about 50% of students from becoming EE's). I also have 40+ years of design and R&D experience.

Go ahead and experiment! Perhaps you are on to something. Perhaps you are the one to make a great invention.

How is this different from msd products w/larger coil that have spark retard, to adjust timing on the fly? A coil that has been underhood for 100k miles definately doesnt have the intensity of a new oem one. i replaced a working coil and it felt like I did a tuneup.

The MSD is capacitive discharge into coil. The follow-on is direct discharge into plug. The follow-on yields a greater spark current for a longer time. The MSD is short duration multiple events.

As stated by others here increased power through the plugs increases wear on them. I do know you can get a short term power gain at least until the spark plugs degrade but is it worth it in the long run? I don't think its economically viable. The only thing that would really work is a double cored plug with two coils feeding it and a separate electrode for each core.


Aircraft have two plugs per cylinder. The also use magnetos that produce a longer and hotter spark. This does give them more power but it does wear the plugs out quicker. To deal with this they go through a much more rigorous inspection and maintenance schedule. I think they are checked every 100 hours of operation.

Kit- I'm a broadcast engineer and electronics hobbyist. I've been making high voltage and arc discharge devices for around 30 years also.

I think we're probably talking at cross purposes here. The idea I had is to create an AC arc that lasts for around half an engine revolution, from the start of the combustion cycle until the exhaust valves open.

I don't see how a small capacitor can sustain an arc through a sparkplug for tens of milliseconds, given that the impedance of the arc is very, very low.

Do you still have the trace of the current sensor? I'd be interested in seeing it, if you do.

A low frequency (tens of kHz) AC signal at 10-40kV delivered by a flyback type transformer is an easy way to get an arc of arbitrary duration. I've used a similar circuit to keep a lean-burn forge burner lit, although this was at a very low power and voltage.

It's probably not even relevant, as there have been so many attempts to 'improve' the classic pulse transformer ignition system.

I've been enjoying discussing it with fellow enthusiasts, though. :D

jim-frank,
I fail to follow your thinking. Once the spark is initiated, the impedance drops, a circuit component switches in a capacitor that delivers current to prolong the spark. The capacitor initial charge of 300-400V and 1uF provides the energy. I do not have a trace however visualize the side profile of a Dodge Intrepid. The back of the car is the initial spark the front bumper is the end of event.

In the flyback case a circuit designed for 10-40KV at low current, at spark, sees a drop in impedance and a voltage less than 50V. Now what! There is a huge mismatch in the circuit impedance.

ConnClarke brings up a good point multiple sparks help. They do this because the flame starts in two places. Speeding flame burn enables improved timing and centering of the energy delivered by the power stroke. If the flame is too slow and timing retarded, some combustion happens in the exhaust wasting energy. It is difficult to find the room in head for two plugs in multi valve (4) engine. Instead combustion chamber design is used to speed flame travel.

A correctly designed combustion chamber with correct mixture, and modern (electronic direct fire or COP) ignition system works quite well. A long duration spark that erodes plugs is not necessary to keep the fire burning!

The time constant of a 1 uF capacitor is one microsecond per ohm of spark gap resistance. If we assume an arc resistance of say 10 ohms, in fifty microseconds the capacitor will drop from 400 volts to nearly zero. That's not very long, and won't prolong the spark much, if at all.

The spark will dump whatever energy is stored in the secondary winding of the flyback, and when the voltage drops below what it takes to sustain the arc, it will extinguish. It will then restrike on the next cycle of the waveform.

Multiple sparks in the time domain help as well, as the Mallory MSD shows.

It's really a series of short duration sparks. Plug erosion will not occur if the energy and power per spark is below what it takes to erode metal from the electrodes.

Do it. You have all the answers. Don't let me stop you.

I'm just an old fart with experience.

Please post you results, I always like to learn.

You're not alone in the 'old fart' club. :D

I might make a model one of these days, but I really doubt that the end result will be sufficiently better than existing designs to make it commercially feasible. It would be interesting from a scientific viewpoint, though.