Naar Nederlands

A Hit&Miss 2-stroke engine

1. The idea
After designing and building three 4-stroke and three 2-stroke engines I got the idea to build a Hit&Miss engine.
Characteristic for this 4-stroke hit&miss is that the exhaust valve is blocked with a governor system on the fly wheel at certain high revolution speed. During that time the exhaust valve is kept open and as a result there is no power because there is no compression and no under pressure for sucking-in the fresh gas mix in this "miss"-phase. The speed is going down until the moment where the exhaust valve is de-blocked again and the normal compression, the sucking-in and the ignition of the gas mix occurs again. In this "hit"-phase the speed is increasing and the hit&miss process is repeating. In fact it is only a very peculiar way to limit the speed of the engine.
The sound of these hit&miss engines is very characteristic: a distinctive "POP-whoosh-whoosh-whoosh-whoosh-POP" as the engine fires and then coasts until the speed decreases and needs to fire again to maintain its average speed. In fact it is a very peculiar way to regulate the speed of the engine or to keep it in check.
These hit&miss engines were in vogue in the years 1900 to about 1930, mainly for pumping machines in agriculture environments. At that time one didn't have disposal of the modern techniques to stabilize this kind of stationary engines so this typical hit&miss process was used for that. These engines were rather heavy and plumb with low speeds and mostly two heavy fly wheels to keep them going on during the miss-phase. It is the very characteristic "POP-whoosh-whoosh-whoosh-whoosh-POP" sound that makes these engines still popular for model builders who remember this walking through farmer lands. Look and listen to this video for a nice demonstration. The practical usefulness of this kind of engines is gone now but the nostalgia around it not at all and that keeps the building of hit&miss model engines very highly spirited.


Two-stroke IC engines don't have an exhaust valve at all, so it is not very obvious to think about a Hit&Hiss two stroke engine. It is even comprehensible to think that it is not possible at all to make a Hitt&Miss two-stroke engine. As far as I know there is one 2-stroke hit&miss engine (the "May Tag") but there only the ignition spark is disabled during the miss-phase but the compression and the sucking-in of the fresh gas mix is going on then. Not the real work if you ask me because the engine must wrestle with the compression during the miss-phase and the fresh gas mix is burning in the hot muffler causing a lot of smoke and pollution.
But I always like challenges and realized that a "real" 2-stroke hit&miss should be possible using the fuel by-pass system that I developed for my "Pressure Controlled two-stroke engine", see also this animation. The trick is to keep the ball in the (upper) intake check-valve lifted during the time between a certain high and a certain low revolution speed. In that lifted position the room above the piston is continuously connected with the room below the piston. The result is that there is no under- or over pressure in the cylinder during that miss-phase so the engine is turning around freely "as driven by the wind", driven by the absorbed energy of the fly wheel. Because the absence of an under pressure there is also no fuel consumption during that miss-phase. As the moment ball is falling back on its seat again the normal two-stroke process is recovered (hit-phase) and the engine is delivering power again.
The animation below is demonstrating this process.

2. The approximate design

The basis for this engine
Apart from some scale enlargement the point of departure for this engine was the design for my Pressure Controlled 2-stroke. I was pretty sure that this enlargement was necessary to drive the extra mechanisms for the system that should lift the ball in the check valve. I choose for a 24mm diameter and stroke for the piston instead of 18mm resulting in a a cylinder content of 11cc instead of 4.5 cc. The remaining part of the basic engine was a matter of enlarging all dimensions proportional. Because the Pressure Controlled 2-stroke engine is running perfectly I didn't need to worry about the behaviour of this engine in its normal (hit) phase.

The mechanism for lifting the ball in the valve.
The movement of the mechanism that have to lift the ball in the valve must be derived from the speed of the engine. With all 4-stroke engines the exhaust valve is blocked by means of a "governor" system with what weights in the fly wheel are moving caused by centrifugal forces. These weights are connected to a lever system ending in a mechanism that block the movement of the exhaust valve at a certain high speed to start the miss-phase. At a certain low speed the same mechanism unblocks the exhaust valve again causing the engine to restart in its hit-phase. For me this is a rather complex mechanical system so it brought me the challenge to design something that is easier-to-make.
In fact I had to solve two problems: how to lift the ball in the valve and how to make the speed dependent mechanism for it. I made several solutions on paper but with none of them I was quite satisfied. So it was time to brainstorm about this with my good friend Jos who always inspires me with smart technical solutions combined with a positive critical approach.
Just before that I had discovered that I could just as well use a steel ball in the valve instead of a neopene ball. This solved my first problem easily: lifting the ball with a Neodymium (NIB) magnet. This kind of magnets are alloys of the rare element Neodymium with iron and borium and are extremely strong. You can buy this magnets of all sizes everywhere and for very little money; see i.e. this web site. For lifting the steel ball I used such a magnet with diameter 6mm and length 8mm. It can lift about 1 kilogram mass which is far more than needed here.
The last challenge was to design a simple as possible mechanism to let the magnet move in front of the ball valve depending on the speed of the engine. After a brainstorm with a good friend we got the idea to apply an "Eddy Current" clutch that e.g. is used in speedometers for (old) auto mobiles and other motor vehicles. The spring loaded pointer is coupled contact less via such an eddy current clutch to the cable that rotates with the speed of the wheel axis of the vehicle, and that was exactly the system that I needed here.

The principle of the Eddy Current clutch.
If a magnet is moving along a metal plate it induces eddy currents in that plate causing counteracting magnetic fields. The strengths of these eddy currents and, with that the counteracting forces depends on the magnetic field in the plate and the speed of the magnet. If the plate can move freely it will move with about the same speed as the magnet below it. If the plate is loaded with some force this eddy current clutch will start slipping. The slipping force can be influenced in several ways, such as:
- Changing the strengths of the magnetic field by means of changing the electrical current in electro magnets. This is done e.g. for regulate the speed of machines staples. Changing the magnetic field in the plate can also be done mechanically by changing the distance of the magnet to the eddy current plate.
- Changing the speed of the magnet. On this effect the speedometers are based and also the "governor" system for this hit&miss 2-stroke engine.

My experiments with the Eddy Current clutch.
To find out if I could make an eddy current clutch with sufficient force and what the geometries should be I did extensive experiments. For that I made an aluminium disc in what I pressed Neodymium magnets and aluminium and copper discs in what the eddy current could be induced. I put the d
isc with the magnets in the head of my milling machine and the eddy current disc on a axis below it with a light spring on it so it only could make some angular rotation. In this way I could vary all kind of relevant parameters and determine the influences of them on the angle rotation of the eddy current disc. I will not list all the result in figures here but only my most important conclusions:

1. The material of the discs.
It will be clear that the disc with the magnets must be made from not magnetic material, so I choose for aluminium. For the eddy current discs I did experiments with aluminium and copper. The better the electrical conductivity is the stronger the eddy currents will be and that's what I found indeed: the induced torque force in the clutch was 2 to 8 times higher with copper compared to aluminium, depending on the distance of the magnets to the eddy current disc. During these experiments I varied this distance between 0.5 and 1 mm.

2. The distance of the magnets to the eddy current disc.
Remarkable was the big difference between aluminium and copper to this respect: per 0.1mm distance increase the torque force reduces with 6% for copper disc but about 20% with the aluminum disc.

3. The number of magnets.
With 4 magnets in the driven aluminium disc the torque forces were about 60% higher than with 2 magnets. With 3 magnets this difference was about 40%.


4.The circumference of the magnets.
With the magnets on a circumference of 44mm the torque forces were about 3,5 times higher than on 22mm; a big influence also. The choice of the circumferences was arbitrary with this tests.


5. The rotation speed of the magnets.

This influence was linear but with a magnet circumference of 44m it was about 1.5 times bigger than with a 22mm circumference.


The results of all these experiments learned me to choose for 4 Neodymium magnets in a aluminium disc with a circumference of 40mm and a copper eddy current disc. The optimal distance of the magnets to the copper disc should be determined on the engine itself. There are more forces playing a role in this system: the pulling force between the steel ball and the magnet and the force of the spring in the other direction. I replaced this spring by a contra weigh later on with the arguments you can find further on.


3. The elaboration of this design
As said, the basis for this engine is my "Pressure Controlled 2-stroke engine" except the scale enlargement; for the description I can refer to the page for the engine. So I will restrict myself here to the constructions that carry out the hit&miss behaviour and some other remaining peculiarities.

The system with the Eddy Current clutch.
I will clarify the final elaboration of this system with the help of the figures below, showing the most important elements for this mechanism.



1.The one way ball valve.
For construction reasons I choose to let the magnet move in front of the ball valve instead above it. On top of this valve there is an adjustable screw to limit the free vertical stroke of the ball to 0.5mm. Later I added there a construction to make the movement of the magnet more "digital" (see under 5) that should be in the way for the magnet. Now the magnet is pulling the ball sideways from its seat but functionally that make no difference.
The ball valve is mounted as short as possible to the cylinder head to make the dead volume between the piston and the ball as small as possible. I applied a bigger steel ball here with diameter 7mm instead of 4.8 because I assumed that this bigger engine would need a greater passing for the gas flows and also that a bigger ball will have less tendency for floating.
This valve must be screwed first in the loose cylinder head before mounting this assembly on the cylinder because the valve inserts the holder for the eddy current axis bearings. This holder must also first be screwed against the cylinder.
All screw fittings of this valve must be made air-tight.

2. The lifting magnet.
The Neodymium magnet with diameter 6mm and length 8mm is pressed (or glued) in a brass holder that is fixed on the axis op the copper disc. On the back side of this holder there is a little hook on what the contra weight (7) is connected with a thin sewing thread that is layed around a little wheel (not visible on this picture). The distance of this magnet to the ball valve housing is about 1mm in its ultimate miss position. This is not critical with this very strong magnet.

3. The copper eddy current disc (see also the picture on the right of this page).
The thickness of this disc is 1mm but is not critical at all. This disc has a short slot in what a detention pin inserts with what the angular rotation of the copper disc is limited. The magnet must be fixed on the axis of the copper disc such that the magnet is right in front of the ball valve in its utmost miss position; not very critical although.
The optimal distance of the magnets to the copper disc appears to be about 2mm. I don't exclude that this could be somewhat more or less i.e. in case of somewhat different magnets.



4. The detention pin.
This pin is a M3 screw in the aluminium holder for the bearings of the copper disc axis. The bottom of this pin is lathed back to about 1.5mm diameter to avoid the pin touching the side edges of the slot in the copper disc.

5. Jug with a spring loaded steel ball.
This brass jug contains a 7mm steel ball that lightly presses the magnet holder on the top of it. With an adjustable spring above this ball this pressure can be regulated. The effect is that the magnet makes a "toppling" (almost digital) movement under the ball with the result that the magnet keeps standing left and right for some time in its both extreme hit and miss positions. Without this "digitizing" the engine is only running restless instead of having a real hit&miss behaviour because at the moment the ball is lifted the engine slows down and immediately the magnet is pulled back letting the ball fall back on its seat. With such an "analoge"system the engine will oscillate around the point where the ball is lifted. The addition of this "digital" mechanism was the most important contribution to give this engine the aiming hit&miss behaviour.

6. The aluminium disc with the 4 magnets ( see also the picture on the right of this page).
This disc must be made from not magnetic material to avoid the disturbing of the magnetic fields. I pressed the 4 magnets with diameter 8mm and lengths 4mm in this aluminum disc. The axis for this disc is rotating in two ball bearings, one in the upper and the other in the bottom mounting plate. The revolution speed of this disc depends on the speed of the engine and on the place where the rubber driving wheel (8) touches the fly wheel. Determining the optimal place for this rubber wheel was a matter of trial and error. Completely at the top of the fly wheel the speed of the disc is maximal but also the counteracting force that the engine must overcome and I had the impression that the engine had some troubles to do that. Finally the best place for the rubber wheel appeared to be at the bottom end of the axis where the wheel is touching the fly wheel at a circumference of about 45mm; so the lowest speed but the system was most stable with that.

7. The contra weight.
When the engine slows down in the miss phase the magnet must be pulled back from the ball valve. At first I used a little pulling spring for that but the adjustment for it appeared not to be very easy and/or distinct. The pulling force of such a spring depends on its thread stiffness, the diameter, the number of windings but also of its stretching. Finally I replaced this spring with a contra weight that is connected to the magnet holder with a sewing thread. This thread is turned around a little wheel so the vertical pulling force of the contra weight is converted to a horizontal one on the magnet holder; extremely simple but 100% distinct. With experiments I determined the optimal mass of this contra weight to about 35 grams in co-operation with all the other forces on the moving magnet.

8. The rubber driving wheel.
This little wheel must be made from soft rubber to make sure that it keeps touching the fly wheel even when this is wobbling a little. So by prefer use a thick and soft rubber O-ring around the metal kernel of this wheel.


The carburettor.
For this engine I again and successfully applied the "Petrol Vapour Carburettor". But in this case I experienced a very unexpected but learning phenomenon! I was already a little afraid that this 2-stroke engine would not easily restart at a very low speed after the miss phase and indeed that was the case at the beginning. Not that astonishing because most 2-stroke engines become rather unstable at very low speeds and after the miss phase this engine almost stands still! At first I thought the problem was that the engine wasn't running fast enough in the (normal) hit phase. It appeared that the reason for that was the expansion vessel that is a substitute for the crank box volume of a normal 2-stroke engine. I made this vessel with the same scale enlargement as for the rest of the engine but that appeared to be wrong afterwards. With experiments I determined that the optimal volume for this vessel had to be about 8 cc instead of the 17cc that I made it originally. After that correction the engine did run perfectly and much faster in the normal hit phase. But the problem with restarting after the miss phase remained.
I have two versions of the Petrol Vapour Carburettor that are entirely the same except for the air in-stream tube on the tank of the carburettor. With the original design this tube inserts the petrol and the air is bubbling through the petrol through little holes in the bottom of the tube. With the second version this tube ends about 1 cm above the petrol surface without any hole pattern. The in streaming air is striking over the petrol surface making a little dimple there. I made this version before because I had the impression that my little engines run even somewhat better with it and that the speed regulating of the engine is less sensible, at least starting with fresh petrol. To my pleasant surprise it appeared that the restart after the miss phase of this engine was no problem at all with this carburettor version, even when the speed of the engine is almost zero!
I have only one explanation for this: as a result of the hit&miss cycles the air bubbling through the petrol is irregular and with that the ratio air/petrol vapour in the gas mix. I guess that this ratio is not optimal just at the moment the engine must restart. It is reasonable to think that this ratio is much more stable when the air is streaming over the petrol surface than when it is bubbling through it. Whatever it is, the difference between this two carburettor versions is tremendous in case of this hit&miss 2-stroke.
Considering afterwards it is a little miracle that this 2-stroke engine restarts so well on this carburettor version. For the same money this project could have failed because of the fact that most 2-stroke engines are unstabel at very low speeds.

The spark ignition
More than with normal 2-strokes it is of big importance that the spark don't miss at any cycle. It could be fatal when this happens at the restart moment at the end of the miss phase. The small micro switches I often use for my small IC engines are not very suitable here. The contacts of this kind of switches tend to float sometimes and burn-in after some time, causing irregular sparks. For this reason I here use the points used for e.g. classic motor bikes and that is working well; they don't float and are heavy duty. The high tension coil is mounted in the wooden base of the engine. For the external 12 volt supply I use the rechargeable battery of my hand drilling engine.

The fly wheel
With my simple one sided crank shaft construction it is not possible to mount two fly wheels on both sides of it as you mostly see on hit&miss engines. So I made my steel fly wheel relatively heavy to keep the engine going in the miss phase. Although this engine is running very well with it this fly wheel it is advisable not to make a lighter than the one on the drawing plan; a somewhat heavier fly wheel could be somewhat better even.
Because the rubber driving wheel for the eddy current clutch is touching the side face of the fly wheel there is no real place for any spoke pattern. That's the reason why the inner side of the fly wheel is entirely flat.

4. Starting-up and adjustments
In good condition the engine can be start-up manually, but generally it is easier or necessary to do that with a loose belt around the pulley on the crank shaft and around similar pulley in the head of your hand drilling machine. It is useful to block the magnet movement at this start-up with e.g. with a piece of rubber between the magnet and the ball valve because the engine then keeps in the hit phase and can be best adjusted to the wanted speed with the throttle valve on the carburetor. When the engine is running well this rubber piece can be removed and the engine will start hitting and missing due to the speed depend movement of the magnet in front of the ball valve.
Staring-up with fresh petrol the throttle valve on the back of the carburettor must be adjusted so that very few fuel mix out of the carburettor tank is mixed with much extra air. Finding this first adjustment point can be somewhat sensible and one tends to turn the throttle valve in the direction of more fuel mix from the tank, but this is absolutely wrong! This initial sensibility is caused by a very volatile component in the fresh petrol but this is disappearing rather fast when the engine runs which is noticeable at a gradually slowing down of the engine speed. From that moment the throttle valve must be turned slowly in the direction of more fuel mix until the wanted speed is back again. From that moment the speed regulation is much less sensible.
The "teamwork" of the several forces causing the magnet movements for the hit&miss behaviour is somewhat subtle. The only parameter that can be readjusted to this respect with a running engine is the pressure of the steel ball on the magnet holder with the adjustable screw above the spring in the jug. All other parameters are fixed adjustments once they are set optimal such as the distance of the four magnets to the copper disc, the mass of the contra weight and the place where the rubber wheel touches the fly wheel. If everything is made and adjusted according to the drawing plan the best hit&miss behaviour can be adjusted with the pressure on the steel ball under what the magnet is flipping. If this is not working well one of the other parameters must be changed, thinking first on the distance of the magnets to the copper disc. In general this together with the presure on the ball will be sufficient to bring the engine nicely in its hit&miss behaviour.


5. Finally
The hit&miss sound of this 2-stroke differs from that of the classic 4-stroke hit&miss engines. One could find that a pity, but except from the fact that the sound of a 2-stroke always differs from the sound of a 4-stroke by nature this 2-stroke design is unique and stubborn so it is not surprising that it makes its own and stubborn sound. In any case this thing is "hitting &missing" undeniable.

I thank my good friend Jos for his significant contribution to this project and also Ernie Weinberg for his useful remarks and comments while building this engine according my drawing plan at the other side of the ocean.
























Nice replica made by Gerard Versluys: