For those interested here is Ignition 101.
Aside from historical dinosaur there are two distinct method still viable for hobbyist: Ketterink is what you find on "Your father Oldsmobile" like the commercial says; and
Capacitive Discharge Ignition or CDI
Kettering - The energy for the spark is stored in the Coil Core's Magnetic field.
The points close for a fraction of the cam rotation applying battery voltage to the coil primary (low voltage low resistance winding).
When the points are close current build up rapidly.
Problem No1
If the engine stops where the point are closed the the current drain is high (several Amps) and coil overheating may results.
Cars have an ignition switch that hopefully is off when the engine is not running, plus for extra safety there is a bulky power
resistor limiting the current and the size of the coil can dissipate the heat if you were so unlucky to stall the engine and leave the
ignition ON.
When the points open, the magnetic field "disappears" and the energy contained in it must go somewhere.
Since there is a small capacitor (condenser) across the points, the circuit opening is not instantaneous but still very fast.
The voltage across both primary and secondary reverses and rise as fast as the capacitor can charge. Some of the energy in the
coil ends up in the capacitor but is only a small fraction of the enegy stored in the coil.
The coil's turn ratio is in the order of 1:100
When the voltage across the points/capacitor reach 150V the points have separated enough that no spark at the points is possible,
on the other hand the secondary voltage could be 15,000 V unless a spark has already occurred.
Keeping the point from arcing is essential to precise timing and long lasting points.
Problem No2
The coil size is dictated by the necessary energy storage, the high insulation necessary and the high temperature environment. In
theory a smaller coil could be built for model engines by relaxing the high temperature requirement, however the reduction is small and the small model market limit the incentive to make such smaller coils.
How do we reduce the size of the Ignition System?
Well the energy needed for each spark is a given, Industry standard sets the value at around 8 mJ.
How much is 8 mJ ? To get an idea 8 mJ of heat energy added to 1 milligram of water will raise the temperature less than 2 degrees Celsius.
8 mJ of energy deposited on your finger will not be felt thermally, however 8 mJ is much above the sensitivity of your nerve cells and a definit electrical stimulus is the result.
Energy stored in an inductor (or more precisely it the inductor magnetic field) is
W = 0.5 x L x I^2. Notice the term I square.
Any attempt to get high energy W by raising the current I requires a bigger wire.
There is a limit to the field magnetic strength in the core, known as saturation.
The inductance L is somewhat proportional to the volume so we have a barrier there.
All this means that the designer of ignition coils have already extracted the optimum design given the material and the constant of nature.
What about storing energy in a capacitor? Can we do it at higher density?
The energy stored in a capacitor is W = 0.5 x C x V^2
Raising V requires thicker insulators, but we have exceptionally better insulator that we have conductors.
Lets say we want to store 8 mJ at 400V, that requires less that 0.15 microfarad.
A 0.15 microfarad rated at 600V is no bigger than a walnut.
So what about charging a capacitor to 400V and dumping onto a 50:1 turn ratio transformer to elevate the voltage to 20,00V.
Well that is good but now we have a bulky transformer... not exactly, the transformer can be quite small.
First, transformer do not store energy therefore there are no fundamental limitation to size.
Second, transformer size is proportional to the pulse time length they have to handle.
It turns out that as long as we deliver the spark energy required to ignite the mixture, within limits, it is not very critical how long we take to deliver the energy.
In other words we can dump the capacitor energy in a much shorter time that the coil energy, therefore making the pulse transformer very small.
Luck has it that a short circuit is ideal to discharge a capacitor fast but is farthest from ideal to get the energy out of an inductor.
The spark is essentially a short.
There is still one problem: where do we get 400V ?
It turns out that semiconductor electronic can provide the necessary circuity in a very compact unit, exploiting the fact that the necessary transformer in the circuit operates at high frequency and is therefore minuscule.
In summary a bulky inductor can be replaced by one small capacitor, two small transformer and a thimble full of components at a fraction of the volume.
I have come to the conclusion that for scale model engine CDI is the only viable solution, if you can not hide an automotive coil.
