By Jan Ridders

I always make my models to my own designs, and never follow existing drawings or concepts. I am forever striving to make model engines with one or more unique features, but mainly based on existing principles. It is actually more ‘playing’ with the basic ‘flat’ physics, a consequence of my working life as a process developer in the field of gas exchange, vacuum techniques and dosing systems. It is a kind of professional deformation of a developer trying to do everything different whether it is necessary or not.
Another feature is that my designs must be as simple as possible from a mechanical point of view, so that I can make the model myself with my fairly limited mechanical skills and tools.
I never look for practical use, high performance and/or efficiency. I leave that to industrial designers who possess the knowledge and experience, have the technical resources, and who are being paid for it.
The difficulty is again and again to find new ideas and to convert them into a working model. This leads more and more to rather bizarre models which I myself find kind of funny, but others probably will look at it with some amazement.
This project is a typical example of such a weird venture. A 2-stroke engine with only one piston in a glass cylinder in which the process takes place alternately on both sides of the piston that is fixed on one common axis.
Figure 1

Free running piston concept

As the piston is pushed away by the cyclical gas expansion of the gas combustion (be it left or be it right) it will cause compression of the fresh gas mix on the other side of the piston at the same time. It's like an engine with a free running piston for which there are several concepts, although I do not know how they are constructed exactly. I could not find much in the way of industrial application so I assume the success is not great.

Figures 2 and 3

Gas exchange process
I opted for the 2-stroke process because I think it is not possible actually with a 4-stroke process here because that process is not producing energy 3/4 of the time. Also a 4-stroke construction would be much more complex as a result of a required double valve system with a 1:2 distribution.

The glass cylinder has two outlet ports through which the exhaust gases escape, and which make it possible to flush the space under the piston again with fresh gas mixture from the carburettor (Fig 3). The piston axle has two constrictions that serve as gate valves. At the right moment the one or the other side of the cylinder is alternately flushed with the fresh gas mixture from the carburettor which is then compressed and ignited when the piston moves back in the opposite direction.
The problem is that the piston itself can not suck this fresh gas here because the process is always on the same side of the piston so the gas combustion and the intake can not occur simultaneously. For this preliminary design, I put a little aquarium air pump between the carburettor and the inlet of the cylinder that makes a constant low pressure fresh gas mixture from the carburettor but with sufficient flow (Fig 1). I tested this on two of my 2-stroke engines and it works fine by itself. If everything works properly, it is in principle always possible to make a little pump mechanically driven by the engine itself. This is possibly a continuation of the project.

Spark ignition
I had to to think of something special because here the one and then the other spark plug must get alternating the high voltage at the right moments, so I had to make a kind of spark distribution. I made the following solution which actually appears to be very simple:
I used the circuit of a Blokker gas lighter. It is constantly connected a 1.5 volt supply battery as normally is usual with a gas lighter. The flexible high tension cable is connected to the spark distributor as shown in figures 1 and 2. This spark distributor is mechanically coupled to the piston shaft and thus moves together with the piston from left to right. The little springs at the ends touch one of the spark plugs at the moment the spark must be made. The contacts on the spark plugs can be slightly rotated so that the timing of the sparks can be controlled accurately. The spark will automatically extinguish if the distributor is more then about 6 mm away again from the spark plug and this is well before the piston reaches the exhaust port so the fresh gas mixture cannot be ignited when it flows through the cylinder through the exhaust port. In this way also there is no need for two micro-switches or reed switches.

Engine behaviour
The gas ignition on the one side of the piston provides the compression on the other side of the piston and vice versa. On to the initial design the stroke of the piston is restricted by rubber pads on the outside of the piston axis. The limitation of the stoke will also be partly determined by the compression generated by the piston movement. Also the timing of the spark will have an impact because the piston will reverse at the moment the gas mixture is ignited.
All in all it was difficult to predict what the speed of movement of the piston is going to be, not in the least because there is hardly a load in the form of significant mass inertia or pressure springs in the preliminary design. But I like this kind of uncertainty because it keeps-up the suspense and interesting reaction when starting up the engine.

I made a test set-up, fixing the engine with the brass base plate in a loose machine vice and connected the carburettor via the small  pump and also the Blokker circuit on the spark distributor. When I manually pushed or drew the piston axis to the left or to the right the bright blue gas explosions immediately occurred on either side of the piston. So, in principle everything worked as I expected.

But something else happened: at the moment one of the exhaust ports is opened the expansion pressure vanishes straight away so that the piston is pressed insufficiently far to make the needed compression on the other side of the piston. To solve this problem, I increased the mass of the piston shaft with pretty heavy weights instead of brass buttons. I also reduced the length of the piston with few steps to give the piston axis some more time to accelerate due to the inertia of the weights. In this way the piston-shaft combination is moving like a bullet shot from a gun barrel but alternately in opposite directions. At one time the engine made a few turns but I was scared to see its extremely violent behaviour: the whole system including the quite heavy vice and workbench trembled; exciting to see, but not exactly the intention, of course!

Concept failed
I have to admit that the free piston concept, unfortunately, failed.

This adventure has again taught me that a principle can be good but not work in practice.

Three interrelated facts:

1. The expansion pressure on the piston, which also has to provide the compression on the other side, is cut off instantly when the piston opens the exhaust port. In order to enable the piston to make sufficient speed, quite a large load must be added to the piston axis.

2. The piston/mass combination is catapulted alternately and in opposite directions, causing an unacceptable rambunctious motor behaviour.

3. If there is only one ignition failure the engine will stop running immediately and irreversibly because there is no element that helps the engine to overcome such a dead moment in the process.

In short: here the effect of a flywheel is lacking which forces the stroke of the piston and helps the engine to survive the dead time in the process.

I should have known this , but I was seduced and deceived by the success of the Bounce 2-stroke which also has no flywheel. But that engine makes a single-stroke without building up any significant pressure to the other side of the piston, where the force of gravity acts on the weight hanging on the piston, which is not the case here.

I had decided to put the thing in the ‘failures’ gallery, but changed my mind. Why not apply a flywheel maintaining the other elements of this design such as the ‘dual’ 2-stroke process that is nicely visible on the left and right of the piston through the glass cylinder and the spark distribution system with the original and unaligned circuit of the Blokker gas lighter?

New Dual 2-stroke concept

So I added a flywheel added to the design and left the rest of the free piston concept virtually untouched. Because the piston is no longer running free I renamed the engine "Dual 2-stroke" to indicate that the two-stroke process is (still) playing on both sides of the same piston.

There was a problem with the spark system: with the free-running ‘Blokker’ circuit the sparks occurred 4 to 5 mm before the springs touched the contacts on the spark plugs. The sparks are thus slightly too early which sometimes caused back firing. This cannot be avoided because it is not possible for a spark only on the extreme positions of the piston or preferably even a little later, but never before. I actually expected this phenomenon but I was hoping that the ignition delay would solve this problem automatically, but this time the luck was not with the stupid.
I also had the impression that the energy of the unaligned Blokker circuit actually is just too low to always ignite the gas mixture reliably.
I have, therefore, again applied the modified Blokker circuit that is triggered by a reed switch, that is switched by two magnets in the flywheel. This modified circuit also has the advantage that the spark energy is a lot higher due to a larger capacitor, the larger voltage results the removal of a zener diode and the doubling of the high voltage capacitor.
Peter, who helped me earlier with the development of the ‘Blokker’ spark circuit, also told me that the timing of the spark with the free-running circuit almost certainly would be too inaccurate due to its own spark frequency in relation to the speed of the engine. I have not been able to check that because I had already left the system with the free-running circuit. But given the expertise of Peter he certainly will be right so this justifies the use of the modified circuit even more.

The engine now runs nicely with a speed of 600 to 1000 rpm which can be seen on the video.

Possible follow-up
The gas mixture from the carburettor with some low excess pressure is still provided by means of a small aquarium air pump. It works well, but it's really somewhat improper, and forced. It would be nicer to add a pump on the other side of the the flywheel that is mechanically driven by the the piston axle. Not that easy to make such a little pump with sufficient yield, but I might attempt to do that sometime, anyway.