A good spark is crucial to let an IC engine run well and reliable. Usually more or less a headache for mostly mechanical orientated model builders for various reasons. Lack of sufficient knowledge and experience is one causes but also large dimensions, the obtain ability, the price, the complexity of the systems or components are obstacles for solving problems in this field. I myself belong also to this category of model builders and I have been wrestling with this problem as long as I make little IC model engines. I think I just understand enough to know where the shoe pinches in in what direction the ideal solution must be found.
My ideal system should meet the following conditions:
1. The spark energy must be sufficient to ignite the gas mix reliable at the prevailing compression pressure but also safe to touch it in all circumstances. That means that the high tension current may not exceed 2 mili-amperes.
2. The circuit must be able to make sparks with at least 25 Hz to make an engine speed possible of about 1500rpm (25x60) for 2-stroke engines; a 4-stroke engine can run 2x1500=3000rpm with that. A higher frequency is OK of course, but I personally don't like higher running speeds than 1000 to 1500 rpm for this kind of (1 cylinder) little IC engines.
3. The system must be relatively small so it can be built-in easily without dictating highly the size of the engine. Integrating the system in the engine is not only a cosmetic item but in fact it is necessary because connecting the spark plug to an external high voltage is very impractical.
4. The circuit must not be too vulnerable or susceptible for interferences and must have a good or, preferably, unlimited life.
5. The mechanical force to drive the switch for triggering the circuit must be relatively low so also very small engine with low power can do this.
6. The circuit or circuit parts must be easily obtainable at a reasonably low price.

1. The classic ignition coil circuit for auto cars or motor bikes
In fact this is a high tension transformer with a short and thick primary coil and a very long and thin secondary coil. The supply for the primary coil is mostly 6 or 12 volt DC with currents between 3 and 5 Amps. If the primary voltage is shortly interrupted an induction peak of about 400 volt on the primary coil is transformed to a very high voltage by the secondary coil: 15 KV or even more. Connected to the spark plug this high tension makes a firm spark over the electrodes from what one mostly is connected to the mass of the engine. Interrupting of the primary current is done by a switch ("points") that is driven by a cam on one of axles of the engine. Parallel to this switch there is a capacitor which has two functions:
- damping the sparks on the contacts of the switch to avoid burning-in;
- further amplification of the high voltage.

On the picture above two examples of classic motor bike high tension coils (bobbin) with diameters of about 40mm and 140mm length. Such coils for auto cars are considerably bigger: diam. 60mm and length 160mm.
The picture below shows a smaller coil used for modern small 2-stroke engines such as mopeds, mowing machines, etc. According to my experiences they are not suitable to be operated with only 6 or 12 volt DC as on the scheme above: the sparks are weak or not at all present then and they become very hot. As far as I know here a capacitor, loaded to some 100 volt by a generator on the crank shaft, is discharged cyclically in a split second over the primary coil so the load of the primary coil is much less; it is a kind of CDI (Capacitor Discharge Ignition) system. So I must advise against the use of these type of coils for model engines.

The advantages of the classic motor bike coil are:
- Strong sparks with almost unlimited switching frequency.
- Very robust and reliable.
The disadvantages are:
- Very bulky compared to the dimensions of most small IC model engines. For model engines with outlines of half a shoe box it is hardly or not acceptable to build-in such a big high tension coil..
- Rather large battery needed for the 6 or 12 volt supply with 3 to 5 Amps; most of the time not suitable to build-in.
- The switch that interrupts the inductive current is a source of interferences if one uses another switch than the original "points". These original points are more and more difficult to obtain and need rather high mechanical force because of the relative strong spring blade; small engines with rather low power can have problems to drive them. Mounting is difficult also sometimes because of the specific and diverse fixing parts. That's why I always use the well known micro switches; easy to obtain for little money, easy to mount and they need very low mechanical force. But they are not made to switch this kind of inductive high currents, reason for relative short lifetime due to burned-in silver contacts.
- Because these kind of coils are growing obsolete for industrial engines the obtainability is more and more difficult. Sometimes I can find a suitable coil at a demolition firm for some 15 Euro and for a new coil (if available) I pay 40 to 50 Euros.
I mostly use this kind of classic coils before, but the mentioned disadvantages encourage me more and more to find a much more suitable solution with the characteristics I mentioned in the introduction above.

2. A piezo ignition.
Piezo crystals produce high voltages when they are squeezed with rather high mechanical force. They exist with all kind of geometries and sizes and are used for e.g. phono cartridges, cigarette and gas lighters. To produce a strong spark they must be relatively large, but even then they are very much smaller than the classic motor bike bobbin. Apart from the small dimensions the advantage is that you don't need any external electrical supply unit. You can simply drive them mechanically with a pusher that is driven by a cam disc on an engine axle.
I applied such a piezo rather successfully with two of my 4-stroke engines: de Otto en de Atkinson. I used a piezo element from a gas lighter that I bought for about 3 Euro in a local warehouse; see the picture below.

Unfortunately, with all my other model engines it didn't work well or not at all. One of the reasons was that the needed mechanical force is too high for the smaller models. Another disadvantage is that piezo crystals are very hard and fragile. They are not made to survive high forces with such a high frequency so they regularly break down although the lifetime was not that bad sometimes. Furthermore the obtainability is not always assured and as far as I heard you can't find them in the USA and other countries; don't ask me why.
All together an incidental solution may be, but not very universal.

3. Electronic ignition (CDI) systems.
First of all I must confess that my knowledge of electronics is rather poor, so everything I state here could well be nonsense. As said before, model builders with "iron hands" normally don't like electronics, except when they can buy a simple "black box" with small dimensions, easy to connect and at low costs. Perhaps somewhat strenuous but I couldn't find anything like this on the whole internet. There are articles how to build electronic circuits yourself, but in general this will horrify the model builders with iron hands. So for them this kind of ignorance is a barrier in advance to think about electronic solutions.
Most ignition systems for industrial engines are electronic these days; you can read all about it on internet. But for the best you can understand a little bit of it but never enough to do something with it, at least not for me and for most builders of model engines I believe. Almost all circuits are or look complex and they often are not very small at all, presumably because they are meant for heavy duty tasks with industrial applications where the requirements are much higher than for simple model engines. As far as you find something that might be suitable for small models they are very expensive; I saw prices like 70 to 150 dollars! Furthermore there are stories on several forums about disputable performances and/or vulnerability.
So I never bought electronic circuits like this which also means that I have no experience with that whatsoever.

About a year ago I found a kind of "black box" spark igniter. In fact it is a small circuit in a gas lighter, making rather strong sparks with a single 1,5 volt penlite battery supply. I can buy this gas lighter for 4 Euro in a local warehouse called "Blokker". That's why I give it this name for this occasion, see pictures below:

So this also is electronics but from an entirely different order of magnitude than what I was talking about above in point 3, both with regard to the size, power and price. I thought it was worthwhile to see if it could be used for making the sparks for an IC model engine.
The principle is actually simple, at least when it is explained to a non electronics like me by some electronic expert. Initially my expert friends Ko Cruck and Rene Duyster and later also Oswald Moonen investigated this circuit and provided me the electrical data with what I could make the below wiring diagram of this original "Blokker" circuit:

In fact the principle is very simple:
- A 0.47μF capacitor is charged continuously with about 100 volt by a free running supply oscillator as long as the 1.5 volt battery is activated by a manual switch.
- An auto trigger circuit behind the oscillator discharges the capacitor over the primary coil of the little bobbin with a frequency determined by the electrical values of the diverse elements in the circuit, mainly the 1N4148 diode that is puncturing every time at about 70 volt causing the thyristor MCR100-6 to discharge the capacitor load over the primary coil of the bobbin. The secondary coil of the bobbin increases this 100 volt to about 6KV, high enough to make good sparks for the original intended aim to light the gas of a stove, for what the spark frequency is of no importance at all.
Below a picture of the dismantled original circuit:

The resistance of the primary coil of the bobbin is about 0.5Ω and I measure about 440Ω for the secondary coil. The diameter of the bobbin is 14.5mm and it length is 18mm. The dimensions of the total circuit are 60x20x20mm.

Adapting the "Blokker" circuit for use in model Internal Combustion engines.

To make this Blokker circuit suitable for modelling internal combustion engines in any case the the diode 1N4148 must be removed because the spark should only occur at the right time in the combustion process, namely at the time the gas mixture in the engine cylinder has been compressed to the maximum pressure at TDP. The control of the thyristor in the trigger section should therefore be made directly or indirectly by a switch on the engine.
After a long time of endless experimentations with much trials and errors I could develop three viable alternatives which what I got an awful lot and unwavering support from my expert friend Peter from Belgium. He gave me not only his expertise in the electronic field, but also the physical tools (including a home made oscillator to simulate the switch with adjustable frequency) and various electrical components to do the experiments. He also has done a large number of experiments and measurements by himself. Without all this work of Peter I would never have achieved the results as I can describe now below. So I am greatly indebted to him or all this work and help.

Some general comments in advance:

1. By removing diode 1N4148 the voltage on the 0.47μF capacitor is more than doubled from about 100 volts to about 220 volts. In itself this is advantageous in that the charge of the capacitor approximately doubles also, resulting in a stronger spark. But this means that the original capacitor which is made for 100 volts maximum must be replaced by a film capacitor that is suitable for 400 volts. To avoid destructive breakdown in the capacitor it must be of the type MKP (some manufactures use the designation MKS).
2. The higher voltage on the capacitor and hence its load increases the high-voltage of about 6KV to about 10KV volts. These are rough estimates based on the electrode distances at what still sparks occur in dry air for which the rule of thumb is that about 1000 volts per mm is required to make a spark. This 10KV volt is substantially lower than the high tension voltage of 15KV volt or more with car or motorcycle ignition coils. But the 10KV does make strong enough sparks for model motors as I make them. But to avoid spark extinguishing at the 3 to 4 Bar compression pressure or more in the engine cylinder the distance between the spark plug electrodes must be made rather small: 0.4 to 0.6 mm maximum.
Special for the Blokker circuit I designed a spark plug that can be simply being self-made:

To make a request for the drawing sheet of this spark plug click here.

3. In principle the spark energy can be made higher by increasing the 0.47μF capacitor to 0.68 μF and/or doubling the high voltage 27pF capacitor by adding a second one parallel over it. However, with a wide range of experiments in which I made use of the oscillator of Peter that simulates the switch on the engine I found that the spark frequency decreases in both cases. Apparently the supply part of the circuit is not strong enough to make sufficient frequency with these increased capacitor values. In particular, the doubling of the high voltage capacitor reduced the spark frequency by more than half, which proved to be fatal for a two-stroke engines that require twice as many sparks per time unit compared to 4-stroke engines. The 0.68 μF capacitor reduces the spark frequency by about 10%, which may be still be acceptable, especially for a 4-stroke engines that need half the spark frequency compared to a 2-stroke engine.
I discovered that the spark energy actually is less important than the spark frequency as well as preventing the extinction of the sparks due to a too a great distance of the spark plug electrodes. Even a short spark of 0.4 mm is igniting the gas mix very well according to my experience.
4. The little bobbin on the original circuit is not electrically isolated. Partly due to the increase of the high-voltage to approximately 10KV it is possible that some breakdown occurs somewhere on the coil or between the coil and the high voltage capacitor. The chance for that is especially great if there is no ignition spark for some reason, for example when the high tension cable is not connected to the spark plug. It is therefore advisable to isolate the bobbin and the high tension capacitor eg. with a 2-component (Bison) epoxy resin. If one uses the epoxy type which is no longer flowing out after about 5 minutes, you have also that 5 minutes to turn around the printed board over and over to achieve a nice and well closed epoxy film all over the bobbin and capacitor. Then you can lay down the circuit and after a few hours, the resin is tack-free; after one day it has hardened for 100%.

In short:
- Always remove the diode 1N4148;
- Always replace the 0,47μF/100 volt film capacitor by a foil capacitor 0.68μF / 400 volt (MKP or MKS type).
- Isolate the bobbin and the high voltage capacitor electrically with a 2-component epoxy resin.

This alternative is so simple that even model builders without any experience in the electronic field can make it I believe. The bottom line is that the MCR 100-6 thrystor simply is replaced by a micro switch on the engine that takes over the cyclically discharge of the capacitor over the primary coil of the bobbin. The wire diagram below shows this circuit:

The pictures below illustrate which components of the original circuit first must be removed (left picture) and, what should be added after that (right picture):

A cam disc on the crank-shaft of a 2-stroke engine or on the camshaft of a 4-stroke engine must press the micro switch as short as possible (about 10% of the cycle time or less ), during what the capacitor discharges instantaneously through the primary coil of the bobbin resulting in the spark on the spark plug. During the time that the micro switch is open (about 90% of the cycle time ore more) the capacitor is charged again.
The 2.7MΩ leak resistor ensures that the capacitor is relatively quickly discharged when the 1.5 volt battery supply for the circuit is turned off.

A small drawback of this very simple and robust solution is that there is a fairly sharp peak current flow every time the micro switch is closing causing sparks across the contacts thereof with possible burn-in effects on the long time as a result. But this is hardly the case with the very good Hartmann MAB1 type micro switch, see the picture below:

The pictures below illustrate which components of the original circuit first must be removed (left picture) and, what should be added after that (right picture):

The reed switch can be controlled by a small Neodymium magnet in the flywheel with a two-stroke engine or in a disc on the cam shaft in the case of a 4-stroke engine.
The spark frequency of the circuit is at least 25 Hz so that a two-stroke engine can run with some 1500 rpm or more with it and thus a 4-stroke with 3000 rpm.
According to the specifications a reed switch only need one millisecond to close and open again, so that's no problem. Incidentally, a good micro switch can follow a switching frequency of at least 30 Hz also according to my experience, so that the same engine speeds can be achieved with that also.

As previously stated, the electrode gap of the spark plug must then be relatively small (0.4 to 0.6 mm maximum) in order to prevent the sparks to extinguish at the compression pressure of the gas mix in the cylinder of 3 to 4 Bar or higher.

Summary

1. Alternative 1 is actually childishly simple and robust and therefore is suitable for modellers with little or no experience with soldering on electronic circuits.
The only condition is that there should be enough place for a micro switch on the engne, which usually will be the case. Furthermore the burning-in of the micro switch contacts on the more or less longer term must be taken for granted, replacing it when the micro switch appeared not to work properly any more. This burning-in effect is not so bad if you use a good quality micro switch such as the type Hartmann MAB1. Maybe the use of the so-called "points" as used in (classic) motorcycles can circumvent this problem if the engine has the power to drive these rather strong spring loaded points.
With this alternative a a 1μF / 400volt capacitor can also be applied to increase the spark energy if wanted. The somewhat lower spark frequency of the circuit will not be a serious problem, at least not for 4-stroke engines that operates with a twice as low spark frequencies than 2-stroke engines.

2. Alternative 2 is a more "neat" solution because it can be triggered with a reed switch (or similar switching element) and with very low current. In particular a reed switch is very small, can be easily installed and simply operated with a small magnet in the flywheel with a 2-stroke engine or in a disc on the camshaft in case of a 4 -stroke engine and the mechanical load on the engine is actually zero. The switching frequency of such a reed switch is very large and it will have "eternal life" due to the very low switching current.
This alternative is suitable for both 2- and 4-stroke model engines but the gap between the spark plug electrodes must be relatively small (0.4 to 0.6 mm maximum) in order to prevent the extinction of the sparks due to the 3 to 4 Bar compression pressure of the gas mix in the cylinder.

3. In both alternatives it is possible to use a microswitch instead of a reed switch with magnet in case one cannot or will not apply such a reed switch for some reason. The micro switch should be operated with a cam disc on the crank shaft with a 2-stroke engine or on the camshaft in the case of a 4-stroke engine.

The advantages of this Blokker spark system

1. Very small dimensions which makes it possible to build-it the circuit in small engines.
2. A very small power supply with a single 1.5 volt AA battery; I think it cannot be be simpler, safer and cheaper.
3. It enables the use of reed switches or similar elements with a very low currents, very high switching frequencies and no floating. They also cause a negligible mechanical load on the engine, which can be of interest for very small models with low power.
4. Extremely cheap: about 4 Euro for the gas lighter and some little change money for some components, a reed switch and a small Neodymium magnet.

The availability of this type of gas lighters:

This gas lighter was for sale at department store 'Blokker' in the Netherlands and Germany but unfortunately they recently have removed this kind of gas lighters from their stock range.

Other suppliers are (click):
- Amazon (UK)
- Kitchen Craft
- E-bay