Start on liquid-piston X-Engine build.

[Owen_N]

1) has anyone here built one of these?

2) the first step is: work out the gearing ratio and actual forms of the rotor and housing.

the rotor is an epitrochoid of parameters 2:1 ratio, d = 1/2 where the generating circle rolls twice round the main circle.
the housing is epitrochoid dim 3:1, 1/2.

The gear ratio of rotor to housing seems to be 3:2 (diagrammaticaly), This is similar to the Wankel gearing.

Possibly the drawings are wrong, and the rotor is the bigger gear, and the ratio is 2:1 ??

You would think it would be different, as the Wankel rotor advances 1/3 of a full turn for each full turn of the crankshaft
The Wankel has the bigger gear on the rotor and smaller gear on the housing.

The LP engine rotor seems to rotate once for every 2 full turns of the crank.

Another oddity is- is the crank at tdc when the rotor is horizontal, or at tdc within each chamber as as it reaches maximum compression?
How does the crank timing match the rotor orientation?
I could plot it carefully , I suppose, and work this out.
you would think maximum compression would be at tdc.

The rotor appears to turn at half crank speed in action,-from diagrams- so for every crank rotation, the rotor ends up in the same position but has only turned 1/2 a turn.

3) produce x-y formulas for the two shapes which can be passed to a CAM program. - these curves can be expressed as this kind of formula.

4) draw up the details in a CAD program. Can this use formulas to plot a shape??

5) buy a suitable mill.

6) buy a suitable cnc-cam conversion kit and install it.

7) buy a dedicated computer and controller adapter- the adapter is usually in the kit.

8) buy the appropriate software and install it.

9) buy or obtain suitable cad-cam software that can talk to the mill and pass in parameters.
10) start building.

I have a tutorial here to do rotary engine CAD drawings, but it doesn't cover this configuration.
I will look for more drawing tutorials on this engine.

I quite like this engine as a potential RC model aircraft power plant.
It can be built down to around 1500g, and make about 5 hp at 6500 rpm, I think, and has very low natural vibration.

The rotor probably needs to be plated with nikasil or similar, and reground.
I need grinding gear and stone trimming gear to go with the mill.

Tip seals could be adapted from a Wankel to be the housing wiper seals.
rotor side seals would be of similar layout to wankel ones, but would need to be hand-formed and ground.
Grinding the edges could be a problem.

Possibly a different form of packing could be used- strings of PTFE soft rod, maybe?
It needs enough stiffness to not extrude under combustion pressure.

This is similar to hydraulic ram pressure, and fairly soft nitrile o-ring material seals work there.
a cross-section of around 2mm would be desirable. O-ring seals are usually about 3 mm.

I will have a bit of a read on blogs to do with mill cnc conversion, and see if I can get up to speed on this.

There is a more mill-oriented post under the Tools forum.

 

[Owen_N]

Further details on the LP x-engine. There are technical PDFs easily available. Look on Google.

Here are some extract images.

At 1 hp at 8000 rpm for 69 cc, sizing will require some revision to get to 5 hp at 6500 rpm.

I don't want 8-10,000 rpm, as it would require a heavy and expensive gearbox, with an included prop drive.

It also seems rather heavy, at 1.8 kg for the 69cc version.
A part responsible for the weight could be the synchronising gear set.

That shouldn't be heavily loaded, and could be made of a much lighter material.
possibly polyamide??

There doesn't seem to be much of a trade in lumps of polyamide for machining.
- I think it is quite machinable.
The entire gear carrier and rotor centre could be made of polyamide.
Cutting the gears at home could be an issue. you would need an indexing fixture or table.
Are gears usually cut on a horizontal overhead tool spindle?

I have seem online claims for Wankel engines of 5 hp at 1500 g, but probably at high revs. swept volume unspecified.
These people seem to be now out of business. - Maybe they didn't get the Darpa Funding?


I have a two-stroke here which is 61 cc, basic 1350 g, 5 hp rated, usable range 6000-7500 rpm
if you add the muffler and ignition, it is still around 1500g total.

The Saito 84 cc 4-stroke seems to have similar or better performance, and easily spins a 24x10 prop.
its basic weight is about 3 kgs, but it looks fairly heavy construction.

The 74cc Saito 5 cylinder is only 2.5kg basic weight, and is probably close to the 61cc two-stoke in output at the same RPM.

from the APC tables, a wooden prop this size should absorb around 5 hp, and make 10 kg+ thrust at 6000 rpm.

----------------------------------------
Is this engine a worthwhile replacement for a two-stroke engine, or should I concentrate on two-stroke development?
There may be more opportunity to get good power/weight/smoothness stats at the chosen rpm???
Extreme efficiency and clean exhaust are not on my list of desirable attributes.
----------------------------------------

The seal system seems more complex than I first imagined.
The apex seals need to be carbon -ceramic, but relatively low friction, working on a steel rotor.
The rotor could easily be eutectic aluminium-silicon, with Nikasil coating, and reground.

The side-seals in this paper seems to be rigid.

Maybe the polymer side seals are a later development?

I think you could manage the chamber walls and seal enclosure so that a PTFE seal would be durable.
It can be in a flame quench area.
With frypan linings, you have to get them hellishly hot to cause deterioration.
- up around 400 degrees C plus.

They say that the steel rotor saves weight, but it would need to be really thin- like 1.5mm or less everywhere.
This could be difficult to machine at home like this.

Mild steel could be too soft and weak to handle.

It would need to be a semi-hard grade. Machined with full carbide tools.

The current rotor seems to float on its centre boss, and be side-located by the chamber side seals.

There are some good shaft assembly cross-section diagrams included here.


List of images.

They are a little out of order:

-analysis of chamber performance 1
- comparison of low and high speed chambers-analysis,
-cad images of the two chambers,
-Comparison of chamber performance,2,
-V3 on the dyno, showing cooling provisions.
-Compression-ignition version-diagrams,
-cross-section v3,
-cross-section v2, and v3 rotor,
-end housing photo with inlet and exhaust connections,
-view of housing inner coating.

 

[Owen_N]

More useful images:

Rotor, - this looks to be a floating drive,
end covers removed,
rotor assembly exposed,
polymer seal in place,
seals1 from paper,
seals2 from paper,
seals3-showing some dynamic modes of the apex seal-diagrams.
Version 3 improvements,
Version 3 performance,
xmd mini diesel, and some stats.

 

[Owen_N]

Remaining good images:
This shows some dyno setups.
It looks like the engines are surrounded in separate water-cooled casings.
I wouldn't think they would blow air through such small pipes, even though the engines are supposed to be air-cooled.
The casings look to be 3D-printed.

 

[Owen_N]

Continuation of LP engine monolog.

1) areas ripe for weight saving are the main crank shaft and eccentric.
This is a solid steel item.

The eccentric loading is quite high with a fairly short throw- under 1 cm?
It looks much heavier built than my comparable model RC two-stroke of similar swept volume.
It has a throw of 18.25mm.

3) the main eccentric only needs to be 1/2 an inch in width.- maybe up to 15mm (0.6")
The eccentric could be press-fitted onto it,
and could be a lighter material rather than steel - or it could be drilled and pinned through the shaft.
A 4mm tool steel pin should be adequate, and OK for a 12mm shaft.
What do you think?
if HP = 5 and rpm = 6000, then average torque =
3.8 kw, 628 rad/s = T = P/w = 6 N-m
peak torque is over about 1/10 of the complete cycle, so expected max = 60 N-m
force at 6mm radius? F- t/r = 60/ 6 x 10^-3 = 10Kn
Stress in single shear = f/a, a = 12.5 x 10^ -6 s = 10 x 10^3 / 12.5 x 10^-6 = 80 MPa
This is quite low, so a pin seems adequate.
If you are using the same crank for a diesel, or turning extremely high rpms, then a more solid
crankshaft is justified.
the two-stroke one should be OK up to 10:1 CR.
I will have a close look at the cross-sections, and scale them to the chamber volume, to check relative loading.

mass check for a disc of steel 25 mm in diameter and 12 mm wide?
with a 12mm shaft, full offset, this only gives 6.5mm full offset.
It probably needs a little more than this.
Den = 8000 roughly.
vol = 5.89 x 10^-6 cu mt
mass = 8 x 10^3 x 5.89 x 10^-6 = 47g, so this should not be a big contributor to mass.
scratch that idea. - stick with a 12mm shaft. and integrated eccentric.

a 12mm shaft 150mm long weighs: 16.9 x 10^-6 x 8 x 10^3 = 135g.
Another possible contribution comes from 1) the rotor counterweight, and 2) the flywheel.

An rc aero engine doesn't count the flywheel , and the prop acts as a flywheel, and is only 50g or so.

A proper flywheel may weigh 0.8 kg, if it is say 80mm diameter.

I still don't see why at 1.8 kg, it can't perform as well or close to a 2.5 kg multi-cylinder four-stroke piston engine in the 6000-8000 rpm range.
69 cc vs 80 -odd cc- about 15 % more swept volume.

Larger versions seem to hit the 5 hp per kg target, around the 35 hp area, so maybe it is a volume vs mass thing?

Tuning has an effect, too. If you are running Atkinson cycle, that will drop power off and increase efficiency.

Later versions they say "about 4 hp" and "about 2 kg", but that is probably invoking the excessive rpm thing, too.
A motor that turns 8,000 rpm or more is not suitable for "Lawn Equipment" ???

 

[Owen_N]

more monolog:

Analysing the design and the options is very long-winded and extensive.
This engine is not as simple as a basic two-stroke.

Here are the v2 and v3 crankshaft layouts, in a larger format.

The inlet area doesn't look too good in the V3 diagram- possibly a reason for lower output.

I am surprised that they managed to get it to work at 10,000 rpm with this inlet system.
I could revert to an inlet plate sliding on the rotor.

Edge-sealing the inlet plate could be tricky- use of flexible trapped gasket, or shaped o-rings in a smooth bore?
The trapped flexible gasket- sort of like a concertina, but it only has to flex about 1 mm, though.
This sounds easier to make.

Another issue is end location of the inner rotor assembly.
If the rotor has been made floating, the "big end" needs to restrained in both directions.
one reason to make the big end eccentric very long is to properly support the synchronising gear.
This does not seem to be highly loaded, but it should still mot rock around and vary in clearance.

They are probably well past "V3" by now, if the rotor inner layout is any indication.

That image of a rotor looked very much a floating design, with a gear tooth spline in the centre.

Since the rotor torque loading at that point are quite minimal- mainly from accelerating the rotor when increasing rpms-
it doesn't need such substantial engagement. A sliding key could be enough.
Also easier to make with home equipment.

again, any motion sideways is minimal.
There could be a very fast rocking motion, or constant angular misalignment.
I would have to think about that. This would make a symmetrical arrangement better

The V3 shaft is a vast improvement over V2. That looks rather unstable.

I see they do have two sets of needle rollers supporting the rotor inner assembly.
<edit>
Using my design criteria for a model aero engine, I could make substantial changes to the rotor internals, provisions for inlet and exhaust transfer sealing, and rotor interior cooling.
Having the hot exhaust port in the rotor is a problem for heat transfer.
Could mixture cooling alone deal with heat buildup in the rotor?
Is there enough room for a coolant flow system into the rotor, as well as inlet and exhaust?
Inlet and exhaust slide plates can be on opposite sides of the rotor.
Is the slide-plate alone sufficient for sealing?
Will the exhaust slide-plate get too hot?

The existing design mixed cooling air with the exhaust, then routes the lot through the end housing, with no need for an additional slide plate.
Mixing cooling air into the exhaust also helps prevent overheating of the end housing internals and main the crank ballrace on that side.

Having inlet and exhaust on the same side makes room for a flywheel-generator on the other side.
Routing the inlet air through the crank shaft and the offset eccentric makes room for exit ports on the same side of the engine.
This is done on very small two-stroke aero engines, with the rotary drum valve as part of the engine shaft.
However, there is a lot of obstruction between the inside of the crank and the inlet passages in the rotor.
This is right through the splineway for the rotor and leads to lots of small round air holes everywhere.
This is not as efficient as a single large port as in a rotary drum valve.

I would suppose roughly rectangular ports, covering about 2/3 of the crankpin surface, would be adequate.
Maybe 6 ports with webs in between?

for a 90cc engine, the individual ports could be about 8mm square.

this leads to a crankpin perimeter of 6 x 8 + 6x8x2/3 = 48 + 32mm, or 80, and diameter of 25 mm.

this sounds doable.- the port webs can taper inwards, with a wall thickness of 5mm, gives a bore of 15mm

Carburettors of these small engines are around the 15 to 18mm mark.

Area of the bore = 177 sq mm.
area of the ports = 64 x 6 = 384 mm.
if the wall thickness can be reduced to 3mm, bore = 19mm, and area = 283mm.
Thus the ports can be made smaller and the webs bigger to match, to line up with the bore flow capacity.

What are the losses likely to be, to get the mixture round the right-angle bends into the shaft ports?
maybe the shaft needs to be larger internally to limit losses.
However, mixture dropout in a large bore passage needs to be considered.
Venting outwards plus the spinning motion probably deals with any mixture dropout.


Maybe the ports can be distributed each side of the splineway, at say 6mm wide by 5 mm long, 2 rows.
The strength of the crank is also a consideration. It must be much larger in diameter than id it were solid, or fully tubular.
The rotor needs a collector ring and there needs to be a seal to the rest of the rotor interior.
This possibly clashes with the idea of making the rotor float on its elastomer seals.
Possible elastomer compressive seals each side of the spline to take up any rotor wobble?

- Maybe a composite main sealing system with springs and rear gas loading would allow the rotor to be firmly fixed to the rotor carrier.

Crankpin needle rollers also need an entry and exit for fuel-oil mixture to be lubricated.
The flow path needs to be divided, with a little flow through the needle rollers.
Normal two-strokes pick up oil from swinging the bearings through the mixture-filled crankcase.
They have notches and slots to let the mixture in. We need an equivalent in the LP design.

 

[Owen_N]

Update 4-02-22 (feb):
We have the generating procedure for the rotor drawing, and the one for the housing drawing, which seem to be simple epitrochoids with the marker half-way up the second wheel axis,
and the rotor spins at half the speed of the crank. - 1:2 ratio-

There is no way a simple pinion and ring gear will fit.

the radius of the smaller wheel would have to equal the offset of the crank, and from the drawing V3
its teeth would clash with the crank.

The offset shown is very small-the smaller wheel would have a pcd about 14mm, the same size as the output shaft stub.
If these sparkplugs are M10 thread, this looks like a full-scale drawing, but not necessarily in real proportions.

Either the offset is much greater than shown in the diagram, or they are doing something tricky with the rotor synchronising gearing.

There may be an intermediate gear shaft with two different sized gears on it.
Maybe 2-stage epicyclic stepdown?

For a Wankel engine, the 2:3 ratio seems to fit simply with the rotor and the centre shaft.

I will try some combinations and see what I come up with.
***********************************
Any ideas out there?
***********************************
Also, V3 as shown is quite different to the actual parts later displayed, for example in the "visible firing engine",
and "How it works".

I would presume that the left-hand crank web is pressed on, and the right hand is one piece with the output shaft.
There are some dyno versions which show output shafts at both ends, but the finished end housing with the 4 pipes sticking out
is not an output end.

I would suppose the external drive out is set behind the magneto/balance weight, or in some cases, it can drive straight out the centre of the magneto. This means some form of dirt exclusion is needed.

Also, I haven't seen where they put the cooling fan, or what fan shrouding they propose to use.


If you look at a chainsaw engine, they generally have a shaft coming out both sides.

Some Briggs and Stratton-type engines have a vertical split, so crankshaft bearings and seals can be
inserter from the sides easily.
The crankshaft can be one piece, and the big-end is split.

One pressed-on crank web makes more sense in this case (the LP X-engine).

Two-stroke chainsaw engines are probably fully pressed up.
I haven't seen the insides of one.

 

[Poppy Ott]

Update 4-02-22 (feb):
We have the generating procedure for the rotor drawing, and the one for the housing drawing, which seem to be simple epitrochoids with the marker half-way up the second wheel axis,
and the rotor spins at half the speed of the crank. - 1:2 ratio-

There is no way a simple pinion and ring gear will fit.

the radius of the smaller wheel would have to equal the offset of the crank, and from the drawing V3
its teeth would clash with the crank.

The offset shown is very small-the smaller wheel would have a pcd about 14mm, the same size as the output shaft stub.
If these sparkplugs are M10 thread, this looks like a full-scale drawing, but not necessarily in real proportions.

Either the offset is much greater than shown in the diagram, or they are doing something tricky with the rotor synchronising gearing.

There may be an intermediate gear shaft with two different sized gears on it.
Maybe 2-stage epicyclic stepdown?

For a Wankel engine, the 2:3 ratio seems to fit simply with the rotor and the centre shaft.

I will try some combinations and see what I come up with.
***********************************
Any ideas out there?
***********************************
Also, V3 as shown is quite different to the actual parts later displayed, for example in the "visible firing engine",
and "How it works".

I would presume that the left-hand crank web is pressed on, and the right hand is one piece with the output shaft.
There are some dyno versions which show output shafts at both ends, but the finished end housing with the 4 pipes sticking out
is not an output end.

I would suppose the external drive out is set behind the magneto/balance weight, or in some cases, it can drive straight out the centre of the magneto. This means some form of dirt exclusion is needed.

Also, I haven't seen where they put the cooling fan, or what fan shrouding they propose to use.


If you look at a chainsaw engine, they generally have a shaft coming out both sides.

Some Briggs and Stratton-type engines have a vertical split, so crankshaft bearings and seals can be
inserter from the sides easily.
The crankshaft can be one piece, and the big-end is split.

One pressed-on crank web makes more sense in this case (the LP X-engine).

Two-stroke chainsaw engines are probably fully pressed up.
I haven't seen the insides of one.

 

[Poppy Ott]

This is an interesting project, and I’ll be following. I have a question that is not really relevant but I have not been able to track down an answer; what does the ‘liquid’ part of liquid-piston refer to?

 

[stevehuckss396]

This is an interesting build. I'll be following along to see how this thing works.

 

[Owen_N]

This is an interesting build. I'll be following along to see how this thing works.

The actual build may be a while in the future.

Just buying and collecting machining gear will take me some time.
I have to figure out whether I can come up with a buildable version yet.

There is also some big gaps in the publicly available information about this engine, unlike the Wankel, which is pretty well documented.
All the versions shown from the company seem to be slightly different.
The most complete one seems to be the Micro, and it looks specifically designed to go in a drone.

A more practical lawn equipment engine would have shafts coming out both sides.

I haven't seen a complete one with the magneto, fan, fan shroud etc installed.

I have a rough idea how the crank mechanism works, but no idea on the layout of the synchroniser gears.
I have a theory that there is a crank- centric transfer gear- similar to an idler gear, about 2/3 the size of the rotor gear, with half the transfer gear driving the ring
rotor gear, and a gear train from the crankshaft driving the transfer gear forward 1/3 faster than crank speed, probably using the same teeth on the idler. The idler doesn't really need another set of teeth.
The main crank gear will be bigger than the idler, driving the smaller gear on a layshaft, and the larger layshaft gear
drives the transfer gear.

I will draw this up and see what teeth numbers, pitch diameters, and tooth module gives me a whole number of teeth.

There is probably a metric module that fits. the metric modules directly give the number of teeth as a

whole number related to the pitch circle diameter. for instance , a module 1, 36mm pcd has 36 teeth.

That is for gears between 15mm and 70mm in diameter.
The engine would probably use a module 1.5 or module 2 for the small version shown in the V3 diagram.
you don't really want to go below about 12 teeth on a gear, I think.

There a lots of other details that need to be reverse-engineered, adapted or designed from scratch as well.

It has taken them many iterations and thousands of hours of dyno work to get to this stage, so I
wouldn't expect anything I can make to be very durable or work well.

Another instance is their 3-piece "tip" seal. The only reason I can see for this is that the dragging action of the side seals
interferes with the sealing of the main seal, which has to switch sides on compression.
The main tip seal also seems to stick out more than the side tip seal, with a step in the rotor.
This may have to do with improving the switching action.

The side seals also have to be formed into shape, as they need a wrap-around form to contact and wipe against the outer
parts of the tip seals.

Making tooling to press-form the seals could be tricky, and would probably need a lot of trial and error.
The seal material is likely to be specially compounded, too, and not pure PTFE, for example.
Is PTFE a thermoplastic? does it need to be formed in an inert atmosphere? At what temperature?
Is the material ductile, and will it take a form if you stamp it?

Could this area be redesigned to make the home manufacture easier?

How good are fluoropolymers in this kind of situation? I have never heard of them being used as engine main combustion seals before.
They do get used in engine valve "air springs" for racing engines, but that is a cleaner environment, and cooler, too.

The shape of the rotor makes rigid seals a bit of a problem. They may need to be in several segments.
You still have the issue of sealing against the "tip" seals to consider as well.
The Wankel is easier, as the seals are fairly stationary relative to each other. the tip seals may cycle in and out a little.

 

[Owen_N]

Liquid? probably some arcane rotor sealing system that they later abandoned.
however, they had registered their company by then.

You could probably find out if you had a good google search.

 

[Owen_N]

Here is an image showing all the Parts in an x-mini engine.

I still can't figure out how the rotor synchronising gears work, though.
<edit>
It looks like the rotor centre has teeth all round, and there is a fixed ring gear bolted to one of the side cases.

However, if you roll the smaller wheel one circuit of the larger ring, it will turn more than one rotation.
according to me, the rotor should rotate in reverse to the crank, but only needs a half turn to visit all 3 outer chamber pockets.
The crank, in the meantime, has done one full turn. I don't see how a 2:3 ratio can do this.
the crank travels 1/3 of a turn for tdc in each chamber.

from the point of view of the rotor when the top of the rotor is centred in the chamber, the bottom, or far end of the rotor,
is centred on the inter-lobe seal. It looks like it needs 1/3 of a rotation to get to tdc in the next chamber along.
however, this rotation is in reverse.

How, if it moves 1/3 of a turn each chamber, it ends up only fully turning a half of a turn when it revisits the chamber it started at?
On the face of it, this is a bit of a mystery.

I could plot the engine in 1/6 of crank turns, and also plot the rotor gear motion,
and see if the 2:3 ratio as it looks, is correct.

You would think the rotor has in fact turned 1 and a half times in reverse, to one forward rotation of the crank.
This is a 3:2 ratio, which corresponds to the size of the gears.

Now this has to be subtracted from the forward one turn of the crank in order to really only turn half a turn in reverse.
It sounds logical, but is a bit of a brain-twister.

Another problem I have is with the inlet seal from the hub to the rotor.
The inlet passageway passes through a long gear, but there doesn't seem to be any side seals.
You can see why the rotor has lots of teeth in the centre- the gear port is quite small, and the rotor can bridge over that gap.
At some stage there has to be a divider between inlet , exhaust, and cooling air.
Where is it?
<edit>
A-ha!
I have figured it out. There are stacks of snap-rings each side of the gear port for the inlet.
the gaps can be staggered.
The rotor boss has a centre gear spline, but ring housings each side, which are a push fit over the snap rings.
The ring housings are not obvious on the dark rotor.

the snap rings can move sideways, and seal in the slots in the gear.

Lubrication of the internal needle bearing is from oil being spun along the inside of the gear from the fuel-oil mix.
the far ends of the needle bearing are also seals.

Now I just have the rotor sealing system to sort out. and see if there is a solution I can do without
a fluoroplastics company to make stuff.
I bet a fair chunk of their development budget went to sorting out those seals.

Also, is there a simple workable alternative to their interchamber "tip" seal with the the three pieces and the step each side?
having the main seal fairly loose means combustion gas can get under it , and press the seal against the rotor.

However, having the side seal drag going one way, often counteracts the sealing effect of the combustion pressure, leading to leaks or poor access for gas to the back of the seal.

The one-way drag of the seal against the rotor also has this effect, but it is probably not so bad.

You would think that as the seal tilts in its slot, this also affects gas access to the bottom of the seal.
They have done lots of simulations to get these interchamber "tip" seals balanced correctly.

Just shoving Wankel bits in there will not work as well. The vertical proportions , spacing, and depth of groove are probably quite different.

 

[Owen_N]

Thoughts on alternatives:
------------------------------
How about a radial engine 4-stroke, 90cc, 5 cylinders, with a disc-drum valve right round the outside?
I had a look, and the diameter would come in around 8 to 10 inches.
If each cylinder is 20cc, the diameter to the head split could be as low as 5 inches.
say stroke is 20mm, rod length = 36mm , piston is 15mm from the pin center to the crown, then this diameter becomes
15 + 36 + 10 to the center = 61mm, total diameter = 122mm.
if bore stroke is 20mm, bore diameter =?
bore = 20, area = 314 sq mm, x 20 = 6280cu mm. iml = 1000 cu mm, so capacity = 6.28 cc.
to get 20cdc, we want x 3.19, so D has to increase by 1.78, or 35.7mm bore.
bore to stroke ratio = 1.78.
compare with 46/36.5 = 1.26. - typical two-stroke.
maybe 1.78 could be a little much??
if we increase stroke to 25mm, x 1.25,
The bore can be reduced by the inverse of the sqrt of 1.25, 1.12, 0.89, x 35.7 = 32mm
ratio is now 32/25 = 1.28- better.
Cross check my calcs: area = 804, vol = 20,106 cu mm or 20cc - sounds good!
the head height can be say 30mm, not including cooling fins.
so diameter = 15 + 45 + 12.5 to the middle = 72.5, x 2 = 145, + 60 = 205mm
add another 30mm for fins, = 235mm , or 9.25 inches.
-------------------------------------------------------
how would this compare with a 90cc LP x-engine?

The basic 69cc version is about 6 inches in diameter, so if we increase evenly in all directions to the cube root,
90/69 = 1.30, cu.rt = 1.09, so diameter become 6.6 inches- a bit smaller than a 5 cylinder radial.

However, the 5 cylinder radial would be much easier to make, and could weigh about 2.5 kg if
Nikasil cylinders are used.
the weight of the LP engine will be not quite cubed, may be to the 2.5 power: of 1.09 =1.24 ,1.8 x 1.24 = 2.2 kg, I think.
Not a lot of difference in weight?
---------------------------------------------------------------
Should I continue to design the half engine speed disc-drum valve that surrounds the entire radial?
This could be supported by a bearing around the crankshaft, if it is supported by an offset cage.
What does this do to air cooling. can it pass through the support cage?
Part of the head structure has to "bend" around the support cage.
Maybe short head studs that side?
also, exhaust would have to come out the rear, inlet enters radially inwards, so the cage needs to be offset down a bit to clear the exhaust.
Having a rear exhaust is potentially a problem for air cooling, which is partly obstructed by the ring disc.

I think a radial with a big sliding disc valve is easier to make.
----------------------------------------------------------------

The diameter could be traded for length as in a barrel engine, but that is worse for air cooling.
One side of the cylinder has much better cooling than the other.
Have there been any good air-cooled barrel engines?
--------------------------------
Is a 5 cylinder radial better balanced than a 4-cylinder inline engine?
-------------------------------
These have been done as air cooled, but you need an extra fan, even in an airplane engine.
-------------------------------
Flat-four airplane engines are good without a fan. As a twin-boxer, they balance well.

 

[Owen_N]

Alternative engine number 2:
air cooled 3 cylinder barrel engine with a rotary plate valve at the end.
This would be quite compact at 90cc capacity.

This needs fairly substantial thrust bearings at each end, as well as a hefty double ballrace centre for the swash plate.
Would angular contact be appropriate?
It also needs a swashplate gimbal, to stop the rods winding up. The gimbal takes side-thrust from the piston rods
the piston-rod ends need to be ball-type. Both ends is easier.
this means the ball-ends should screw into the piston-rod, so the ball caps can be installed.
There may be existing ball-ends which will do the job- the ball part, anyway.
Would a though-bolted ball off a Heim joint be adequate?- you want about 30 degrees deflection.
half-speed gearing for the valve plate could be a crownwheel and pinion, straight reared.
If the valve plate is solid, all the combustion/compression thrust can be taken by one bearing, which is more efficient.
Should the valve plate have diverting passageways? This would be a bit like overhead drum valves , but laid flat.
the valve channels could exit in a conical interface, to clear the end housing.
If the entire barrel rotated, single inlet and exhaust ports could be used, and the spark plugs could be in the valve plate.

I would like to pinch an idea off the disk-drum valve, as it allows fairly unconstrained combustion chamber and spark plug location.
This means separate ports for each cylinder, and a fixed "Barrel" part.

Having the valve axis at right angles to the cylinder constrains sparkplug location, if I want flat seals at the top of the cylinder.
You end having 2 sparkplugs as they are offset in the cylinder "head".
The sparkplugs also need to come in horizontally, unless the valve plate is not extended right out to the side of the cylinder.
more convoluted seals have to be used elsewhere, unless the exit plane is also flat, and the inlet and exhaust ports on the outside,
run at separate radii,
There also needs to be provision for an end box housing for bearings and gears..
An alternative is to have a hollow main shaft, and put the gearbox and thrust bearing at the other end of the engine.
That could interfere with power takeoff.

If they are jammed between the piston rods, the engine will need to be longer in this area.
Flat seals in the cylinder are simple, as they can be formed like side-on piston rings.
on air intake/exhaust outlet outer side, the exhaust and inlet can be continuous manifolds, as there are no seals needed
between cylinders.
This could cause some problems with the exhaust, as one exhaust opens while another is partway through blowdown.
maybe individual exit seals on the exhaust are desirable..

I need to look at a plate valve design that allows 3 lots of 4 stroke cylinder cycles to go on at once.
I am sure something similar has been done. The trick is finding drawings.
Maybe the plate has to go slower than half speed- maybe 1/3 speed?

Just valve-timing one cylinder takes quite a few degrees on a single half-speed rotary valve.

 

[Owen_N]

I am terminating my updates on this thread, as it looks that the sealing system is a bit too high tech for home reproduction.
I will start a new thread considering a 4-stroke build with 3 round cylinders 90 to 100 cc,
and with some kind of rotary valving.

It is not exactly novel, though.

I like the air-cooled barrel engine, as the rotary plate valve can be quite small in diameter.

There are the complications of needing a gimbal-like linkage to stop the swash "spider" from turning.

Also, bottom end lubrication could be tricky, if a full set of rings is not used.

It would be nice to draw premix through the swash case.

 

[Owen_N]

Add-on:
An idea for the tip-sealing problem:
1) The side seal can be rectangular- spring loaded.
2)The extreme side parts of the between-chamber seal hang over the sides of the chamber, and extend down to the main side seal.
this needs a fairly precise gap, say 1mm.
3) the hanging -over section takes the form of a small disc, so that the tip seal can rock.
This end piece could be keyed into the main tip seal.
4) The critical problem in this case is: how good a tolerance can we maintain with an assembly like this?
What is the relative expansion between the rotor and housing as it heats up?
If a ring 45mm in diameter needs a 0.2mm ring gap. and the perimeter is 141 mm, then relative expansion is
0.1 percent. This is for cast iron.
A eutectic aluminium-silicon rotor will probably expand more than this. more like 0.1mm in 45, or 0.2 %
the actual rotor is likely to be about this width.
effectively, this makes the leakage gap about twice what you would get with a piston engine- this may be acceptable,
considering the gap is likely to be about 6mm x 0.1mm
The engine wold have to be fitted together very carefully, with interchangeable side discs, and a very good finish and sizing of the rotor edge.
A one-piece side seal will need some restraint to stop it creeping around the rotor.
Maybe interlocking step pieces in the groove, and small slots in the back of the seal.
Could the main seal be machined from a thin slab of commercial PTFE?
It would be hard to hold as it was machined.
A stock 2mm PTFE rectangular section, bent to shape and inserted, would be better.
I don't think anyone actually makes these. I can check whether cut-to-length sealing material is available.
An alternative is to have several curved segments of stainless steel, bent to shape from square rod then surface ground, with
ptfe joining pieces under slight compression.
The joining pieces could be sunk into round recesses in the rotor.
The cross-section of the seal should be only 2mm square.
Does the fit to the rotor need to be a ground quality, too?
The side channel for the seal could have drillings into the chamber area, as combustion pressure needs to be compensated.
This sounds doable.

Does anyone else have any other ideas on dealing with these small gaps down to the seal groove?

Also, can we evaluate in some way whether PTFE or a compounded fluoropolymer would work?
The chamber side needs to have a bonded PTFE coating as well, to stop sooty particles sticking and being drawn into the seal, causing rapid wear.
I will look into locally available ptfe coating services.

2.5mm Square FKM O-Ring Cord 75D
https://www.theoringstore.com/store/index.php?main_page=index&cPath=117_2409
This is a fluoro rubber. Would it have the correct properties, high temperature resistance?
would making a deliberate gap above it for the seal connecting piece cause direct flame contact a degradation?
<edit>
Better idea. "V" the edge down to the seal.
This gives the side piece more support, and the actual gap can be really small.

 

[Mat H. Physics]

Give me a hot minute to review the entire post, but I think things are about to accelerate.
Hope you've seen, and played with, this mathematics engine.
Would love to follow this build to learn how to determine shapes by function equations.

 

[peterl95124]

Give me a hot minute to review the entire post, but I think things are about to accelerate.
Hope you've seen, and played with, this mathematics engine.
Would love to follow this build to learn how to determine shapes by function equations.

that's really neat !!!, or at least potentially neat as it seems the author of the app used a set of parameters that IIRC are different from the ones used in Ansdale's book written for engineers rather than mathematicians, but who can afford a Mathmatica license, or can put up with its absurdly terrible user interface.

 

[Corsair_Pilor]

I know this was a year ago, but curious if you got anywhere with it?
I am tempted to have a go at making one, I want to use it in a rc car application.
I wanted to do the standard rotary, but this one seems kinda cooler, and possibly have better performance.