The purposes of this article is to design and build a simple system to trigger off the shelf dual or quad MSD ignition units.


Rotary engines fire four times as often as a four cylinder AC engine so the life of the plugs is shorter to begin with. The engine cruises at 5000 RPM instead of 2500 RPM and a four cylinder engine has four combustion chambers and not just two like the rotary. This places a lot of wear and tear on rotary engine spark plugs. When they are used up there will be a very noticeable loss of power so change them out every 50 hours.

My hangar neighbor has done tons of R&D on lean burn engines and he claims the more powerful the ignition the leaner you can run the engine and the lower the fuel burn up to a point. The hot tip is capacitor discharge ignition systems. Capacitor discharge systems have been around since the days of the vacuum tube. There is a switching power supply in there that jacks up the 12 volts from the battery to about 300 volts or more. That charges a large capacitor. When that capacitor is discharged through the coil a high voltage spark is generated.

Another reason we should be interested in capacitor discharge ignition systems is the availability of auto fuels at airports is not as wide spread as we would like. 100LL (lots of lead) fouls rotary engine spark plugs rather rapidly.

It is well known fact that the CD systems have a very fast rise time and that is what is required to punch through the lead deposits on a spark plug. This has been widely known for 30 years or more. One must plan one's cross country flights carefully to land at airports that have auto fuel.

This chart from Heinze Heisler s book Advanced Engine Technology showing a cap discharge time or pulse width of 30 milliseconds. Most cap discharge systems use a 1 mike cap and 300 to 350 volts. The pulse width is a function of the size of the cap and the voltage.

A slow rise time low cost inductive ignition system works OK for no lead auto fuel but that is not what we really need. If we are going to be able to burn 100LL we are going to need a capacitor discharge ignition system that will punch through lead deposits.

Klaus Savier owner of Light Speed Engineering has exploited this situation to lean his engine out at high altitude and improve the BSFC to remarkable levels by increasing the spark voltage. He has achieved as much as 100 MPG at 130 MPH and 60 MPG at more reasonable cruising speeds. This is a piston engine where the spark plugs are at one end of the fire and the piston is at the other end unlike our situation. IMHO much of the piston engine ignition knowledge is not applicable to the rotary.

The consensus in the rotary racing industry is MSD short for Capacitor discharge Multi Spark. It is also a brand name. The new MSD ignitions fire over and over again so the longer burn time of an inductive system is less of an issue. Here is what MSD Corp has to say about it.

MSD makes a multiple spark capacitor discharge ignition system widely used by Mazda rotary engine racers but it only works up to 3000 RPM. After that it is "one strong spark" says MSD. Not enough IMHO."

there is another company called AEM that make MSD systems. I called them it too does not fire more than once past 3000 RPM according to the electronic tech guy. He kept saying there was not enough time but he was incorrect about that in regard to a rotary. I don't think they designed it for a rotary where the rotor is moving one third as fast as the shaft.

If you use one MSD per rotor it has more time to fire more than once at high RPM's. We don't have distributors any more so we need more than one. One each for the trailing plugs and one for the leading plugs on each rotor or one for each plug.

7000 RPM is 116.7 rev per second or 42,000 degrees per second or one e-shaft degree in .02 ms.

Here is an interesting observation. In any low cost inductive ignition system such as used by most car companies the coil is saturated by a DC current for a period of time. This is called the dwell time. When the current flow is interrupted by the points or an FET or transistors opening, a high voltage spark is generated from the secondary winding due to the collapse of the magnetic field in the coil. There is a lot of power put into the coil so it could over heat and fail. Certain Chevy V8 LSx coils are very prone to this type of failure when used with a rotary engine.

The SAME coil used in a CD system operates in an entirely different mode. It becomes a pulse transformer. There is no longer a need for the ballast resistor. These had a tendency to fail anyway when the Mazda engine is run at a high continuous RPM.

Therefore don't throw those old Mazda coils away folks. As Perry Mick points out they can be unbolted from the igniter frame, disconnected and used as pulse transformers with a CDI system. Here are some pics. In fact you might want to start buying them up on e-bay before word gets out :)

If you want to buy new coils MSD sells units specifically designed for their MSD systems.

If you want to start from scratch here is an idea I had a while back.

Shown are three capacitors charged to 300 volts switching a coil in sequence by a CPU. Any number of caps and switches can be used but the output impedance of the power supply must be low and the converter frequency higher than 20Khz to recharge the capacitors rapidly. The sequence can be rapid or somewhat less than rapid at lower RPMs. In other words a variable pulse rate multiple spark capacitor discharge ignition system. The limit is the CL time constant. High C voltage low L means a rapid voltage rise time with a narrow pulse width.

I was thinking of experimenting with a 15KHz CRT fly back transformer. 15KHz is a period of .01 ms if my arithmetic is correct.

I have come up with a very simple way to have four or more sparks one every 15 or 20 degrees or so of the e-shaft. I think this might reduce the BSFC by about .03 or .02 at least I hope it will. A .03 improvement in the rotary BSFC would be a 6% reduction in fuel burn bringing the rotary almost exactly in line with a Lycoming aircraft engine.

These are the range of rotor positions I intend repeatedly firing the spark plug. Mostly the leading plug (lower plug) as one does not want to risk igniting the mixture in the following chamber. In other words we are going to chase the fuel air mixture with the spark.

Patent applied for.

That BSFC reduction does not sound like much but it is about 5% reduction in fuel burn. Mazda completely redesigned the 16B rotary and retooled the factory to get a 5% reduction in fuel burn at car power levels. Cost? Hundreds of millions if not a billion dollars.

Most of this was optimizing the position of the spark plugs and the timing. Evidence that there is potential for lower rotary engine fuel burn with more R&D focused on the ignition system.

In our case we have a moving combustion chamber where the fire travels along side the spark plugs for quite some time. IMHO We want the spark plugs to continue to fire for as long as the unburned and burning mixture is still present moving past the plugs. That takes 8.3 ms at six thousand RPM cruise RPM.

A capacitor discharge ignition system is very good at firing spark plugs fouled with 100 LL but it only fires for .2 ms. Even an inductive system (Chevy LS1 coil) or stock Mazda ignition only fires for about .8 ms.


IMHO we need an ignition system to fire for many milliseconds or at least 60 e-shaft degrees.

This is why I am doing a m.s.d. ignition system with several pre-charged capacitors fired in rapid sequence by a by 3 or 4 inductive sensors per rotor mounted 15 or 20 degrees apart near the e-shaft ignition steel disk with one tooth. Extremely simple and no computer required.

I hope this is a break through in improving the BSFC of the rotary engine. I think it is worth a try.

Ignition triggers or crank angle sensors. If you are lucky enough you might get one or two with your engine. If you have to buy them they are $60 to well over $100 each. We need at least two for the TLC555 FI and maybe a couple more for the ignition.

I have decided to do away with the ABS and crank angle type magnetic inductive sensors as too expensive and too big. I would have trouble clearing the water pumps. Klaus has been using hall effect sensors for years so I am going to copy him. They are cheap at less than 2 bucks each and we need ten. It will only require one small rotating magnet to trigger the ten sensors. The sensors can be placed much closer together so the diameter of the trigger system can be reduced and fit below the water pump.

Here is a Hall Effect bench test set up using a Radio Shack DC motor activate by a varialble voltage power supply. This motor will rotate up to 10,000 RPM.

Here is what the trigger wave form coming from the Hall Effect sensor looks Like. The pulse width is about one millisecond wide.

In a 555 fuel injection system you need two sensors 180 degrees apart to trigger the fuel injection. Eight more for this ignition system and a magnet on a trigger wheel. The angular positions of the Hall Effect sensors control when the injectors are fired and when the spark plugs are fired. Simple!!! Move the sensors to change the angles and hence the timing. Fixed timing works fine for aircraft engines. Magneto timing is fixed at about 22 degrees BTDC. You would like to retard the ignition for starting and advance it at high altitude low power cruise. In the old days distributors with centrifugal and vacuum advance did this automatically.

Manual advance mechanism.


I have searched high and low for a long time for a LOW COST high power 12 volt to 300 volt DC to DC converter. Finally I found the solution. Here is what I have come up with.

I have located a higher quality, more expensive 400 watt inverter unit that puts out European 220 volts AC so the DC voltage should be around double this or about 290 volts. Just about right for a CD ignition system. More to come as we will need caps and switches.

This is a mass produced $35 automotive DC to 220 volt AC 400 watt fan cooled converter. It weighs about 2 pounds.

Fortunately Power Bright did not go to all the trouble of forming a perfect sine wave. This modified square wave is perfect for our purpose as I will invert the lower half with a bridge rectifier. When inverted and filtered it will be 290 volts DC. That is the big trick that allows us to use a mass produced off the shelf item as a powerful CD ignition system. The current drain on the 12 volt bench power supply was 8 amps. The load was a 60 watt incandescent light bulb. The unit does not like to run with no load.

All we need on the output is a full wave diode bridge circuit to convert it to about 250 volts DC for rapidly charging the capacitors. One for each spark. Current to the capacitors can be as much as 3 amps! At the full 750 watts current drain on the battery will be as much as 60 amps! I am not sure we need this much ignition power until I test the overall system at 7500 RPM but I have it if we need it.

You are looking at one of the most powerful ignition systems ever put on an engine and it only cost $35 :)

Paul Lamar