A 15cc sidevalve opposed twin

[Peter Twissell]

Combining my hobbies of RC model flying and model engineering, I have designed an engine for my next aircraft model.
The aircraft is a well established design (Chris Foss Acro Wot) and is intended to use engines from 6cc 2 stroke to 15cc 4 stroke.
I don't hold out hope for impressive power output from my design, so I have gone for the upper end of the engine displacement range.
I have decided on a sidevalve layout, both for simplicity and to keep the overall width of the engine within the model's cowling.
My design uses a pressed together crankshaft assembly with one piece connecting rods.
The main structure of the engine consists of two identical 'blocks', machined from aluminium with iron liners for the cylinder bores.

 

[Peter Twissell]

The block half has the iron liner and bronze valve cages pressed into shouldered bores. The pair of block halves is bolted together and bored through for the crankshaft and camshaft.

 

[Peter Twissell]

The Crankshaft is supported by ballraces mounted into the end covers, which locate in the large bore through the blocks.
The camshaft is supported by ballraces which are captured directly in the smaller bore.

 

[Peter Twissell]

I have already started manufacture.
The photo shows one of the blocks with the liner installed and most of the machining completed. The liner bore is undersize and will be finished as the final operation, after the valve cages are pressed in and all other machining is complete.
Apologies for the poor quality of my photos!

 

[Peter Twissell]

A little more detail on the engine design; the bore and stroke are 21mm and 22mm respectively. The valves are 9mm diameter with 3mm lift. The combustion chamber is the well-established heart shape. The engine will run on methanol with glow plug ignition. Running on methanol significantly reduces the heat rejection issues, especially around the exhaust valve area in sidevalve designs.

The total weight of the engine 'firewall forward' i.e. including carburettor, exhaust system and mount, is 728g (calculated in CAD). My CAD model is created in Fusion 360 and I have included every nut, screw and clip, each represented in their correct materials.
For comparison, an OS90 4 stroke weighs 703g with exhaust but without mount, so I am happy with that.

The camshaft has only three cams. The middle cam operates both intake valves. The other two cams operate the exhaust valves.
Lubrication will be similar to the majority of 4 stroke model aircraft engines, using oil in the fuel and relying on blow-by to get it into the crankcase space. A vent in the timing gear case will allow excess oil to escape and crankcase pressure to be relieved.

 

[Peter Twissell]

I have yet to decide on how to produce the induction and exhaust pipes. Both are 8mm bore, 10mm OD and require bends at 1D (i.e. 10mm centreline bend radius).
For my larger engines, I have been able to purchase stainless steel 'dairy tube' mandrel bends at 1D, but I have not been able to find such bends in smaller sizes.
The options I am considering are:
1) Manufacture a suitable miniature pipe bender with a mandrel and make the pipes from copper.
2) Use the lost PLA method, 3D printing the parts and casting in bronze. I have no experience with lost PLA casting. From what I have read, bronze is among the easiest metals to cast.

Any suggestions welcome!

I am currently leaning towards making a mandrel bender, not least because I like to make tools

 

[Peter Twissell]

Part of the reason for building this engine is as a practise run before I start on a Whittle V8.
I have never made anything this small and my previous engines have used commercially available piston rings.
I intend to use the Trimble method to produce rings, as detailed in the Whittle V8 build articles. I will also use Eric Whittles instructions on manufacture and use of an iron cylinder hone.

 

[Ken I]

Pete - I'll be watching closely - always wanted to design a 4-stroke.
Blow by lubrication in the fuel ? Will crankcase vent be to waste or to inlet to "burn off" ?
Are you going to key the pressfit crank, square, round or doweled ?
Suggestion for exhausts - I worked for an electronic company where we made Copper wave guides with a tapered rectangular bore with razor sharp internal corners - if you looked closely you could see circular machining marks right into the internal corners which puzzled the hell out of most people.
We did this by making a male mandrel from cerrobend - we electrolytically copper plated it for 48 hours and then machined the outside from datum holes in the cerrobend - plugged with wax for later mounting points (After 48 hours of plating the coated mandrel looked like a brown Kiwifruit). We also attached perspex blocks to the mandrel to provide all the holes, windows and grilles required without further machining.
You can plate onto a heavier flange (pre-machined) - laquered where you don't want to plate.
After machining the outside - we simply melted out the cerrobend - the machining marks remained evident on the female inner surfaces - right up to the sharp corners - very puzzling unless you knew how it was done.
Wall thickness after machining was ±2.5mm - I don't think you are going near that but it might be an interesting method to explore.
Where the bore shape permitted (no rentrant angles) we used steel mandrels which were knocked out an reused - they were coated with some sort of conductive release agent - unfortunately I can't remember what that was.
You might get away with using a 3D printed PLA plastic former and premachined Copper flange - after plating the tube to thickness you just decompose the PLA in an oven.
The surface finish might cause poor plating to thickness - so maybe just a piece of heat formed rod.
These days, I think plating with what is termed "self-leveling" Copper plate - you should be able to plate to an appropriate thickness without having to touch the surface afterwards.
Most people don't get that Copper has a high melting point = 1085°C.
Regards, Ken

 

[Peter Twissell]

Hi Ken,
Thanks, I hadn't considered electro deposition. I'm not sure it is possible to build up 2.5mm thickness (typo?) But it's worth me trying some test parts. I would make the cores from white metal, which can be accurately cast in silicon rubber moulds, which in turn can be made from 3D prints.
Blow by lubrication is the norm in model aircraft 4 strokes. I'll start with some oil in the cases anyway and monitor the vent, which will simply exit to atmosphere.
The pressed crank assembly will not be keyed, following common practise in motorcycle 2 stroke engines.

 

[Ken I]

Pete, We built up way more than 2.5mm - that was the finished thickness - we actually built it up to about 4mm to ensure we had enough material everywhere otherwise it would be scrap.
The build-up was quite uneven (especially over sharp corners) but I know there are much better self-leveling copper plating solutions out there today.
I would start with a straight rod for simplicity, next add a flange - once you have all the bugs doped out, go to the actual parts. (There's always going to be bugs.)
The plating was done in our in-house facility - I had nothing to do with that - but they weren't terribly sophisticated in those days (early 70's) - we (in the model shop) referred to them as the "Chemical Milling Department" because of the number of precision parts they etched undersize.
I know about motorcycle cranks I used to mechanic for various racers for a number of years - we even assembled a 250 2-stroke twin so both cylinders fired together (the owners idea not mine - I didn't approve) and he called it a "twingle".
I was just curious as to how that works when scaled down.
Regards, Ken

 

[Peter Twissell]

I'm impressed that they could build up that much copper in such a relatively short time. I am used to plating build up in a few thousandths of an inch per hour.
The 'twingle' thing was and still is popular with some racers. I am told it gives them better feedback at the limit of grip.

 

[Ken I]

It was an electrolytic process which tends to deposit way faster than current electroless types. And yes - typically only 2 thou an hour - so maybe they pushed it a lot but sooner or later that causes trouble and you end up with a pineapple. I never actually watched the process I was only involved in the machining processes and the 48 hours was a "throwaway line" of dialogue and of course my memory might be faulty.
But we did make 'em this way.
Regards, Ken

 

[Peter Twissell]

Thanks Ken, definitely worth further investigation.

 

[awake]

Peter, I love the look of this - looking forward to seeing it take shape and run!

 

[propclock]

machine botherer. I like that ,
I often think of myself as a machine Racoon,
I can't "see" it if I can't touch it. Great project.

 

[Peter Twissell]

More swarf has been made.
The cylinders now have cooling fins. They're not very deep, but based on other glow ignition engines of similar size, should be sufficient.
Valve cage have been made and pressed into the cylinder blocks.

 

[Peter Twissell]

Next is the valves.
The photo shows the "blanks", with the valves head OD finished, the neck machined and the clip groove finished.
The valve stems are left 0.005" oversize and will be finished at the same setting as the seat angle, with the aid of a traveling steady. Final size and finish will be achieved with a lap.

 

[Peter Twissell]

More progress:
To finish the valve stems and seats, I set up a travelling steady, cut the seat with the cross slide rotated to 46 degree angle and then advanced the tool to finish the the stem diameter. With both machined at the same setting, they should be concentric.

 

[Peter Twissell]

To finish the valve faces after parting off, I had to make a split collet. The neck of the valve is 2.5mm diameter.
The collet grips the valve on the neck, but also provides support on the back face of the head.

 

[Peter Twissell]

Valve cages are a simple turning job. The OD was turned to a press fit in the bore in the block and the internals finished at the same setting.
As with the valves, this ensures concentricity of the critical features. The gas flow cavity was cut with an 8mm ballnose endmill.