The Bourke Engine 4

[SBBlue]

Okay, kids, here's your assigment: To learn all about the Bourke Engine.

Go visit this website and watch the animation;

http://bourke-engine.com/ani.htm

Supposedly this engine has superior fuel economy and emissions characteristics. As far as I can tell, it is a two stroke engine. Instead of using crankcase scavenging, it uses the lower portion of the cylinder to pump the air necessary for charging the combustion chamber.

It is also claimed that the "Balanced Precision Reverse Cam Effect Roller Crankshaft" did good things too, although from the animation it looks to me that it just results in the piston spending more time at TDC and BDC, and I'm not sure the advantage of that. It apparently is also known as a "Scotch Yoke." The engine is also supposedly "self supercharging."

It is also claimed that the exhaust gas temperature is 190 to 240 deg F, the compression ratio varies from 8:1 to 20:1, and that the air/fuel ratio is 30:1 to 50:1.

I have spent some time looking at the animation, and while I would believe the engine would run, I just don't see how it is "self supercharging", I don't see the advantage of the "Scotch Yoke", and I don't quite understand how it would operate with a 30:1 to 50:1 fuel/air ratio.

If you could get ignition at an air/fuel ratio of 50:1 I could understand why the exhaust gas temperature would be quite low. In addition, the combustion temperature at an A/F ratio of 50:1 would be quite low, which would explain low nitric oxide production, and the PM's might be low by virtue of the excess oxygen present in such a lean condition. But it looks to me that the power density of such an engine would be absymal.

More information on the engine can be found at:

http://www.constant-pressure.com/Home.htm

I must also confess that I don't quite see how this would be a "constant pressure" engine, since it looks to me closer to a "constant volume" engine.

An interesting and somewhat different view of the history of the Bourke Engine can be found at:

http://www.niquette.com/books/sophmag/bourke.htm

Does anybody have any insight about this that I am missing?

 

[notnats]

Pat,

The only thing I can see that gives it an advantage as an HCCI engine over a conventional crank and conrod is that the rigid coupling between the opposed pistons might mean that the compression forces are transmitted directly from the firing cylinder to the compressing cylinder without loading the scotch yoke mechanism. This might be how such a flimsy looking mechanism does not wear out quickly.

Still doesn't make it balanced.

Still haven't seen any power or fuel consumption figures.

Jeff

 

[ivymike]

...except that the combustion force of one won't match the compression force of the other, since they're both higher near TDC, and TDC happens for one cylinder 180deg before the other. Energy from one cylinder will still be transfered to the crank/flywheel/vehicle for storage, then transferred back to the rod assy to finish compression at the other cylinder.

Somewhere along the way I must have missed the discussion about why the scotch yoke helps this thing. It doesn't seem to me that trading a couple of piston skirts for a scotch yoke helps to reduce friction. The piston motion isn't going to be tremendously different either (sinusoidal vs sinusoidal+higher orders)...

 

[schwee]

Thanks for the explanation, Pat -- that makes a lot of sense to me.

I can't see the advantage of a scotch yoke arrangement, either, unless it does in fact let the piston spend a longer time at TDC. And if it does, what advantage would this provide -- better approximation of constant volume combustion?

Is there any reason not to consider the action of this thing simply that of a standard non-turbocharged, ported 2-stroke, short of the (possibly) HCCI combustion?

And -- what makes this different than a two-cylinder 2-stroke or 4-stroke opposed with a conventional crank setup? Or are conventional opposed engines balanced (firing from both sides simultaneously)?

 

[ivymike]

there was an old engine used in fighter/trainer planes in WWI that used opposed cylinders with a single crankpin - it would shake planes apart fairly rapidly. There were some conversion kits that came out later, changing it to a two-pin crankshaft. I don't think the cylinders fired simultaneously, but they reached TDC at the same time.

I believe that the planes that used these old 2-cyl shakers were called Petenpols or something similar. There was also the "penguin?" trainer aircraft.

 

[jmw]

The link to Constant Pressure Engines in the opening post provides a rational and clear description of how it works and why the Scotch yoke is considered valuable, though it doesn't answer all the questions raised here.

The data on this site shows exhaust gas temperatures at 700[°]C while the another site indicates much lower temperatures.

But note: two of the sites apparently reate to the original Bourke engine or to modern manufactured copies while the Constant Pressure Engine site relates to the later version and to R&D on this later version with ongoing development.

(One of the sites is pretty miffed at the third site...)

I am intrigued to see that the engine has sufficient going for it, apparently, that at least two companies are embarked on R&D with plans to manufacture.

It is interesting to note how novel engine ideas are received and there are plenty of examples in these threads. I wonder how well the Wankel Engine compares for novelty to benefit ratio? and how well it would have been received here at a similar level of development.




JMW

http://www.viscoanalyser.com

Eng-Tips: Pro bono publico, by engineers, for engineers.

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[j2bprometheus]

Hi-
The Bourke engine website mentions that the engine could "...run in super lean detonation mode...". This sounds like it operates like Honda's "active radicals" engine ( the charge autoignites or self-ignites without aid from the spark plug). Honda's engine only operates in "active radicals" mode at relatively light load. Honda's engine is a two-stroke, and it is available to the public in Japan. I don't think that Honda's engine is sold in the US.
The Bourke engine could indeed operate as lean as claimed if it is operating in the same combustion mode as the Honda engine. The Honda engine does have some emissions but the emissions are relatively low. Hence, the claim that the Bouke engine has no emissions is not reasonable.
Power density would be low for any engine operating that lean. So a very large engine would be required to produce a large amount of power.
Their claim of "Fuel Consumption: 1/4 pound or less of fuel per horse power hour " is vague (they don't say what kind of fuel). Assuming that the fuel is gasoline, that would mean that the engine achieves 53 % brake thermal efficiency --- very high (large diesel engines can reach this level of efficiency). This is much higher than the efficiency achieved by the Honda engine, which operates in a similar mode. This makes the claim of "1/4 pound or less of fuel per horse power hour" a bit suspect.
All in all, the website does not seem to me to be very reliable.

 

[jmw]

Interestingly the Constant Pressure Engine development is based on testing as spark ignition gas engines but they claim they think it would make a good diesel.

JMW

http://www.viscoanalyser.com

Eng-Tips: Pro bono publico, by engineers, for engineers.

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[schwee]

From the original link:

http://www.constant-pressure.com/RD.htm

"As the piston reaches the bottom of it's [sic] stroke, the exhaust port is uncovered and the remaining gases are exhausted into the atmosphere. The exhaust gas temperature is much lower than normal engines, due to the fact that as the piston is traveling down on the power stroke, the volume in the cylinder in expanding . . ."

Uh??? I think this means that exhaust gas temp is low because the charge has already completed its [there we go] burning while there was a pause at TDC. So no burning as cylinder volume expands, so lower exhaust temps and better emissions (because closer to constant volume combustion) (this might also have something to do with a faster burn rate?).

The above makes sense for the "Vaux" engine, which appears to be identical to the Bourke except perhaps that it is spark instead of HCCI. But by no stretch of the imagination do I understand the free piston claim. It's not a free piston engine as I conceive of it. Maybe only free in the sense that the scotch yoke allows a dwell at TDC instead of being tied to a standard articulated crank setup, with no dwell. I'd love to see the reciprocal motion of the scotch yoke plotted -- it could be that it reduces reciprocal loads by accelerating and decelerating more gradually than a traditional setup.

 

[Bourke30]

If anyone believes this hobbiest is fraudulent, you are mistaken. I have tried to unravel the Bourke engine since 1979. Bourke said we would have to "shake off our preconceived notions" while working with them. Bourke is an unusual animal in that it can produce a smooth flow of power from seemingly unusable detonation. I would agree with the term "knock engine", except for the fact that that lack of piston side loading allows the Bourke to detonate without crushing the skirt oil film and slamming the skirt into the cylinder wall. This is what causes conventional articulated pistons to melt, even for a short period of detonation. If conventional articulated engines are optimized for F/A ratio, then where does the extra heat come from to melt the top out of the piston? For that matter, where is the extra oxygen in the charge to form oxides of nitrogen? Leaner burn requires either more compression, or a longer duration of compression. The Scotch yoke gets the piston to TDC faster, allowing more time at fixed volume. Try putting a rod off vertical in an arbor press and it will kick sideways. Simple vector analysis says that the Bourke rod can withstand higher BMEP's. Detonations are inaudible in a sealed vessel. Detonations are inaudible in the Bourke, also. It can tolerate "knock" pressures easily, without any noise. The rigid rod assembly transfers charging and compression forces directly in a straight line. Thinking about how this is done in the conventional recip, and the term "Rube Goldberg" seems appropriate. Lots of going around corners, with rods being accelerated parallel and laterally to the centerline. The yoke assembly is simply being thrown between two violent, alternating pulses.It represents a single mass. Variation between piston weight is not critical. Because the flame speed in the Bourke is supersonic (all molecules being inflamed almost instantaneously) energy is released much faster than with a conventional burn rate if 150' sec., in which the inflamation passes relatively slowly from molecule to molecule. This has been thoroughally researched using Pratt and Whitney engine manuals. Engineers and pilots have been taught that detonation has no place in conventional engines. The faster release of energy allows more TIME for thermodynamics to come into play after the charge is spent.Expansion causes cooling, and this is the reason for the low EGT. More cooling is done in the cylinder, and less in the pipe or atmosphere. You will never see flaming gasses in a properly tuned Bourke exhaust. My lowest temperature using Fluke digital equiptment with the correct gas probes was 196 deg F. Sure, you can argue about probe location, errors, etc., but even with this in mind, the exhaust temp is still much lower than conventional. I have been using Coleman lantern fuel, with varying dilutions of diesel fuel or furnace oil. Bourke calls for "white gas", but this is unobtainable. As far as vibration, if everything is spot-on, it will run virtually motionless, and will not vary as RPM changes. Other researchers claimed they shook violently, as it did for me in my early runs. This is because the "null" was not achieved, one cylinder not producing the same BMEP as the other. The Bourke is a delicate balancing act. Cylinder pressures and phasing are critical for smooth power delivery. A paper cup put on the camcase during operation will not fly off, even at idle (3750 RPM). There is no "bum-ba-bum-bum-bum", as with a regular two stroke. Charges cannot escape unburned, as in four-stroking. Our test equiptment specs are posted on the
<

http://bourkeengine.net>

site.
This forum was listed after searching "Bourke engine", and I thought that adding my experiences with real running models would be helpful. I think we have gone past saying that it "might" run. I have shown that it will run smoothly and produce cooler exhaust temps. Ten units are being prepared for testing by outside labs. I'm afraid that even with certified results, there will still be those who say it cannot run at all. Such is the plight of those of us who see that Russ Bourke was "on" to something about internal combustion that has been avoided since the beginning. If this were not the case, we would not have to "doctor" fuels to keep the burn rate within limits.
Internet postings are a little long, but clarity and quality were more important than download speed. Careful study of the Solid Works animation will reveal a reason for the smallest detail.
If a hobbiest with more successful running Bourke Engine hours than all the others combined is of no importance to this discussion, I will retire to my garage and put some more hours on the 15 Kw generator test bed.

 

[SBBlue]

Reading your description of the engine makes me think that it is just a slightly different flavor of an HCCI (homogenous charge compression ignition) engine. They have a very lean air/fuel ratio, they have a power density so low that they are virtually unusuable, and they make quite a racket. They don't really seem to be all that efficient.

But the combustion and exhaust gas temperature are quite low, and they have quite low particulate matter and nitrous oxides level.

Another characteristic of HCCI engines is ignition control can be quite a challenge.

 

[patprimmer]

At last an attempted explanation instead of just claims.

That is all that was ever asked for.

I did not say your claims were fraudulent, but either misconceived or fraudulent. I will always say that of claims that are not supported by data or evidence. I now withdraw that statement in the light of some evidence.

I agree, there are several advantages to keeping the forces in line, not the least being virtually eliminating side forces on the piston and bore, and also greatly reduced bending forces on the rods, and removal of lateral acceleration and deceleration.

I agree that most of what we hear as knock is piston slap which will be greatly reduced with a Scotch Yolk, but the instantaneous peak loads on the piston crowns will still be very high, unless the power density is very low.

If the piston loads are very high, diesel engine type pistons will be required.

Both Scotch Yolk and conventional con rod engines instantaneously stop the piston at TDC and BDC, but in both cases this is an infinitely short time. Whether one infinitely short time is shorter than another infinitely short time. The rates of acceleration before and after that instant are lowest for the Scotch Yolk, and increase in conventional con rod engines with decrease in rod to stroke ratio, but at reasonable rod ratios, the difference between the piston acceleration rates is of little practical importance.

I still don't see a supercharging effect unless there are valves that don't show in the drawing.

I guess I will once again try to download the Solid Works animation, that is when I am about to leave the computer unattended for an hour or so. Maybe it will become clearer.

I certainly do not want to stifle nor criticise invention, but I certainly like to see claims supported with credible data, logic and evidence.

Regards
pat pprimmer@acay.com.au
eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[Bourke30]

Were it possible that Russ Bourke was unknowingly doing HCCI in 1936? Homogenous charge compression ignition, about which I learned from this forum, fits his descriptions perfectly. Honda's "active radical" engine can maintain this condition for 15% of the powerband. Once "in cycle", Bourke's HCCI is continuous throughout the entire powerband.
The transition to "cycle" is clearly audible in the "proto" video posted on
<

http://bourkeengine.net>

Again, please forgive the long download time. I wanted to preserve as much detail as possible.
During start-up, the familiar two-srtoke "bum-ba-bum-bum" is followed by the distinct "jump" to detonation mode, where the engine smoothes-out and the exhaust pulses are uniform. All exhaust smoke stops. I actually have to lean the fuel screw to get this to happen. I use a Matrix air clutch and variable regulator to engage the motor-generator A.C.and D.C armatures. The engine responds better to fuel flow adjustment than to throttle adjustment. As far as dynamic balance, slight oscillations decrease as RPM increases. This is also clearly visible on the video.

patprimmer

"Both Scotch Yolk and conventional con rod engines instantaneously stop the piston at TDC and BDC, but in both cases this is an infinitely short time. Whether one infinitely short time is shorter than another infinitely short time. The rates of acceleration before and after that instant are lowest for the Scotch Yolk, and increase in conventional con rod engines with decrease in rod to stroke ratio, but at reasonable rod ratios, the difference between the piston acceleration rates is of little practical importance,"

Bourke 30

These "infinately short" times are extremely important. I would use milliseconds per reversal, Since the I/C formula calls for "fixed volume", and since Bourke flame rates exceed 5000'/sec., the additional time at confinement can make a huge difference. Bourke chose the Scotch yoke mechanism for this measurable diffrence, to maintain volume, avoid pressure drop, and provide mechanical advantage during pressure maximum. Mother nature dictates that in order for a reaction to go to completion, it must be violent. Bourke is better positioned to absorb these detonations by virtue of the straight line solid connection between the pistons, provided by the yoke.

For Proto video and Solid Works animation;

<

http://bourkeengine.net>

For historical photos and Proto model:

<

http://bourke-engine-project.com>

Technical information contained in the "Bourke Engine Documentary", by Lois Bourke and Elwin Coutant, is available elsewhere on the web. Graphs and plots along with many magazine articles and references give good reason to listen to what Bourke preached.

I'd like to thank all of you for the fair hearing.

<

http://roger@rogerrichard.com

 

[patprimmer]

Bourke 30

Can you provide the bore, stroke, and volume over the piston at TDC, then one of the younger whizz kids who still remembers the math or who has the appropriate software, can run the numbers re comparative changes in volume over the piston during the burn time, for a convectional engine of same dimensions with say a 1.5:1 and a 2:1 rod to stroke ratio, as this range should cover most production motors.

A second note. If that is you in the video, who is feeling the exhaust, adjusting the carby and putting your head down near the rotating parts, I would caution you as to the possibility of serious injury if anything goes seriously wrong with this experimental engine while you are so close to the moving and stressed parts.

Regards
pat pprimmer@acay.com.au
eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[patprimmer]

Bourke 30

Unless I lost a decimal place or 2, a quick calculation says that an engine rotates through 36 deg in a millisecond if the engine is doing 6000 rpm. A millisecond is a long time compared to infinitely short time or instantaneous.

36 deg does involve a significant change in the volume over the piston, and the percentage increase is higher for a high compression engine as the original volume is smaller.

As I say, with actual dimensions this can all be calculated accurately

Regards
pat pprimmer@acay.com.au
eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[Bourke30]

patprimmer

I have what you request on floppy disc- Microsoft spreadsheet. It will be posted as soon as we can get it to our site manager.
Thanks for the heads-up on safety. Early-on it was a little scary, but as I gained confidence that it wasn't going to explode, I got a little closer to it.

 

[GraviMan]

Now that we are all friends again (aren't we? )...

Here is my take on the Bourke Engine concept, as shown in the linked website:

1. The high efficiency claim may result from the fact that as a detonation engine, the P-V cycle is the theoretical otto cycle. Since the compression ration is 18:1 and the heat injection is near instantaneous, the peak pressure will far exceed anything seen in a diesel.

2. The total balance results from the crank having a balancer equal to the total piston/con-rod mass. This leaves a vertical component to balance. In the case of the engine shown, the main "scotch yoke" bearing has an offset mass, also equal to the total piston/con-rod mass. Due to contact driven rotation the vertical balance mass effectively moves in a vertical line. An epicycloidal mechanism is suggested as a better solution, however.

3. In order for the balance mechanism above to work, yoke bearing contact against the piston nearer TDC must be maintained. Since the engine runs at high compression ratios, and has 2-stroke detonation combustion, this can be maintained up to higher than expected RPM.

4. The engine is apparently 2-stroke in layout, with yoke mechanism providing a sealed lower chamber. This allows the engine to scavenge effectively, and is sometimes misquoted as supercharging.

5. The ideal fuel for the engine is the subject of debate. The ideal fuel allows the detonation point to be maintained at TDC, regardless of load and rpm.

6. By running very lean, this engine has much of the cooling in the active gas stream. In this way cylinder walls may be kept cool in the same way that a gas turbine combustion chamber temperature is controlled.

7. Since the detonation occurs throughout the mixture near instantaneously, no NOx may form. Since the working fluid is far below stoichiometric throughout the cylinder, there are no regions of high temperature. This gives a very favourable combustion environment for emissions.


Hopefully this (brief) analysis of the evidence presented in this thread, and on the site will rekindle this discussion. I promise not to bit anyone's head off if they don't quite understand any of my points above...

Sincerely,

Mart

 

[Bourke30]

GraviMan

Bourke is Otto cycle taken to its' extreme.

The roller cam assembly spins as a flywheel, independent of the rod-yoke assembly. Counterweights correct only for the mass of the journal bearing. It is strobe balanced to 10k by removing small amounts of material from the counterweights.

For every action, there is an equal and opposite reaction. In the case of the Bourke, the action of the rod-yoke assembly (which includes pistons, rings, pins, rods, and yoke plates being one mass), is thrown down the bore, causing the engine casing to want to go in the opposite direction. Within milliseconds, depending on RPM, the opposing cylinder fires, and the rod-yoke assembly is thrown in the oposite direction, causing the casing to try to fly outward. This is why I call it exchange of momentum. The engine casing transfers momentum to the opposite cylinder where it almost instantly reversed, again. The time between reversals is the frequency of vibration. This can only be accomplished if the cylinders are in a straight line, horizontally opposed, common pin. This is why the Bourke gets smoother with RPM increase. Less time between reversals. This dynamic balancing scheme is absolutely elegant, as no percent balance factor is necessary. With a little mental gymnastics, this should become clearer. I got a much better handle on it after I got the proto to fire-up and observed the oscillations. Any misfire translates into shaking and vibration. This pits the Bourke against established practice. It was stated further up in the replies that cylinder pressure has little effect on dynamic balance. Bourke changed the rules. What I have described is visible in the proto video engine start-up.

Bourke is definitely a two-stroke. The scavebge chambers below the pistons induct the charge, which does not have to carry oil for bottom end lubrication, since the bottom end case is sealed. An added benefit is that oil never sees the carbon from combustion and does not get contaminated.

Although the rod assembly is not fixed to the reverse roller cam, I still don't think it can be called a free piston design. The reverse roller cam serves as a PTO, converting the oscillating motion of the rod assembly into rotational motion, and limiting rod travel in case of misfire. The gas dynamic is a fundamental part of the workings, as, apparently, is not the case in conventional articulateds.

There is only so much energy in the inducted charge and Bourke is not reinventing thermodymics. His mechanism converts a larger percentage of this available heat into useful work by eliminating dead strokes and other "power thieves", as he describes it. He is not squeezing more energy from the charges, just more useful work.

Bourke's fuel curve is linear. Fuel demand is determined by load and not RPM, being similar to current demand of an electric motor.

Although I have not tried for the claimed 20k RPM, it appears possible due to the lack of tensile forces. At this level, induction stall would seem a real possibility. My model was intended to do 4K, and induction was sized accordingly, limiting RPM to 8k tops. Larger intake tracts can be installed for more breath. Bourke parts are hard to come by, so the reason for being conservative.

The large cam roller rotates unidirectionally, a round bearing rolling in a circle in a square box. Why the Collins design opted for a square rubbing block is beyond me. Why slide when you can roll? They have also defeated many of the Bourke benefits by going with a conventional four-stroke cycle top end, with its' aditional complexity and expense. One comment called the rod-yoke assembly flimsy. It is adequately sized, constructed, and constrained. Rods are hollow for side wall strength and minimum heat transfer. All torsional moments are borne by the guide bushings. Obviously, minimal mass is beneficial. Less power required to operate the mechanism means more power out.

Even if Bourke were on par with conventional outputs, its' simplicity and relative ease of manufacture makes it an ideal alternative engine candidate. Hopefully, upcoming water brake testing will put hard figures on its' true character. Let the chips fall where they will.

Again, thanks to the forum for allowing me to present the case for Bourke. My goal was to reproduce Bourke's work and get to the bottom of it- one way or the other.

 

[GraviMan]

"Counterweights (on the roller cam assembly ) correct only for the mass of the journal bearing"

I take it you mean the crank balance by this?

"...the action of the rod-yoke assembly ... is thrown down the bore, causing the engine casing to want to go in the opposite direction."

Hmmm, technically this is actually using the engine as an inertial absorber. I understand your point, but the ideal is that the block remains absolutely static. Put your hand on a Jaguar V12 to see what I mean.

"Any misfire translates into shaking and vibration. This pits the Bourke against established practice."

I read this as the roller cam assy goes out of synch, since it momentarily loses contact. This means it cannot produce the required vertical balance component.

"The scavebge chambers below the pistons induct the charge, which does not have to carry oil for bottom end lubrication, since the bottom end case is sealed"

Yes, I like this feature. It is another reason I personally favour epicycloidal big end 2-strokes...

"The gas dynamic is a fundamental part of the workings, as, apparently, is not the case in conventional articulateds."

Again, I think this is due to the need to keep roller cam in contact with the yoke.

"His mechanism converts a larger percentage of this available heat into useful work by eliminating dead strokes and other "power thieves", as he describes it."

Well, lets just say that it is a high compression ratio otto. Most here will accept that. 4-strokes do suffer some pumping losses, but 2-strokes also suffer scavenging losses.

"Fuel demand is determined by load and not RPM, being similar to current demand of an electric motor."

Basically this means the thing has very good volumetric efficiency across the rpm band. It does not "come on cam" so to speak.

"Although I have not tried for the claimed 20k RPM, it appears possible due to the lack of tensile forces."

Well balanced motorbike racing 2-strokes regularly get 20k...

"The large cam roller rotates unidirectionally, a round bearing rolling in a circle in a square box."

I don't see how it could balance any other way.

"Again, thanks to the forum for allowing me to present the case for Bourke..."

Anyone that has put this much effort into a project, hobbyist or not, will have a great deal of practical experience to offer...

Mart

 

[ivymike]

I read this as the roller cam assy goes out of synch, since it momentarily loses contact. This means it cannot produce the required vertical balance component...
I don't see how it could balance any other way.

I still don't see how the engine shown could balance this way either... wouldn't the "roller" have to complete one revolution for each engine revolution? Wouldn't this require that the roller diameter be approximately (2/pi) times the stroke of the engine (about 2/3)? In other words, wouldn't the roller radius have to be precisely 2/3.14 the crank radius (instead of ~1/1 as shown)? If the roller diameter is even the tiniest bit too big or too small, then the weights will go "out of sync, right?

Also, for this to work the way you describe, wouldn't the firing pressure halfway through the firing stroke have to be precisely equal to the compression pressure halfway through the compression stroke (so that the roller switches from one side to the other at exactly 1/2 stroke, every single time)? (no inertia load at 1/2 stroke, right?) If it switched a little late or a little early, the roller would skid and the engine would go "out of sync," right?

Don't yokes like these work much better with very little clearance? If you used teeth to "sync" the roller, you could get around roller size problems, but wouldn't that require excessive clearance?

 

[ivymike]

"precisely 2/3.14" -> meant precisely 2/pi