Atkinson Differential Engine - Making it work?

[dsage]

I've been busy trying a few things. All of them failures. I tried a slightly modified version of the book carburetor. Conclusion - complete junk. I should have used the time to build the one Gordon presented a few posts back. The book carb doesn't have a venturi so the moving air does not create a vacuum with which to suck in fuel through the orifice. The air just knocks fuel off the needle and orifice in hopes it will be vapourized into the air stream. I blew some air through the finished carb and it actually pushed fuel back out the fuel delivery hose.
SO I made an adapter for a spare carb I have for my Howell V4 / V8 engine. That carb makes a spray of fuel as nice as a spray gun when you blow air through it and it is small and has a throttle.
Still no luck getting the engine to run.
Then after observing the motion of things again in my CAD program I had a revelation.
If the engine does not fire, then when the right piston goes through it's power stroke where the gasses are supposed to be expanding, instead (because it didn't fire) there is a huge vacuum created in the cylinder. More than the amount of volume compressed because the right piston moves very quickly to the right.
I always wondered why there seemed to be a pretty strong vacuum on the carburetor - much more than could be explained by the fairly short and relatively slow intake stroke.
What I think is happening is that the strong vacuum created in the cylinder by not firing is pulling fuel/air in through the carb when the intake port is uncovered. The vacuum is relieved but the fake charge is quickly pushed out the exhaust when the left piston follows along in the exhaust stroke.
I think this was pretty obvious when I looked at the intake valve popping open. It had a huge excursion - almost the full 0.1" limit of the spring. I would have expected it to barely open being self regulating.
So I think there is a fundamental problem with having two atmospheric valves especially on this engine.

So you might ask why doesn't a normal (four stroke) engine have a problem with vacuum in the cylinder when it does not fire. On a four stroke engine with cam actuated valves, the vacuum in the cylinder cannot open the intake valve so the vacuum persists until the the exhaust valve opens.
On a two stroke engine if it does not fire then the vacuum might help bring in the new charge or at least the fuel is coming is anyway by it's normal action of pressurized crankcase. And it's the compression stroke that follows. With this engine the intake stroke follows and it is being messed up by the missed power / exhaust stroke.

Gordon I believe you observed fuel dripping out your exhaust port. This may explain that. i.e the fuel is being sucked in and immediately blown out the exhaust port. I have a similar problem.
I think you've been lucky that the intake stroke actually pulled in some new charge or there was enough left around after the failed exhaust stroke to get it to run.

I'll admit that if it does fire i.e. you somehow stumble on the exact mixture then in may be self sustaining. But I think the first time it misfires it's doomed to failure unless conditions are just right for it to survive the screwed up mixture.

Let me know if you think I'm missing something. My brain hurts.

You might need to have a look at the video I posted a few posts back that explains the motion of the pistons to see what I'm guessing is the problem. Keeping in mind the action of a misfire condition.

I'm going to have to do some thinking about this before I go any further. I'm not going to do any more work on it without a clear direction. Maybe some sort of mechanically actuated intake valve??

Comments welcome.

 

[teeleevs]

I've been busy trying a few things. All of them failures. I tried a slightly modified version of the book carburetor. Conclusion - complete junk. I should have used the time to build the one Gordon presented a few posts back. The book carb doesn't have a venturi so the moving air does not create a vacuum with which to suck in fuel through the orifice. The air just knocks fuel off the needle and orifice in hopes it will be vapourized into the air stream. I blew some air through the finished carb and it actually pushed fuel back out the fuel delivery hose.
SO I made an adapter for a spare carb I have for my Howell V4 / V8 engine. That carb makes a spray of fuel as nice as a spray gun when you blow air through it and it is small and has a throttle.
Still no luck getting the engine to run.
Then after observing the motion of things again in my CAD program I had a revelation.
If the engine does not fire, then when the right piston goes through it's power stroke where the gasses are supposed to be expanding, instead (because it didn't fire) there is a huge vacuum created in the cylinder. More than the amount of volume compressed because the right piston moves very quickly to the right.
I always wondered why there seemed to be a pretty strong vacuum on the carburetor - much more than could be explained by the fairly short and relatively slow intake stroke.
What I think is happening is that the strong vacuum created in the cylinder by not firing is pulling fuel/air in through the carb when the intake port is uncovered. The vacuum is relieved but the fake charge is quickly pushed out the exhaust when the left piston follows along in the exhaust stroke.
I think this was pretty obvious when I looked at the intake valve popping open. It had a huge excursion - almost the full 0.1" limit of the spring. I would have expected it to barely open being self regulating.
So I think there is a fundamental problem with having two atmospheric valves especially on this engine.

So you might ask why doesn't a normal (four stroke) engine have a problem with vacuum in the cylinder when it does not fire. On a four stroke engine with cam actuated valves, the vacuum in the cylinder cannot open the intake valve so the vacuum persists until the the exhaust valve opens.
On a two stroke engine if it does not fire then the vacuum might help bring in the new charge or at least the fuel is coming is anyway by it's normal action of pressurized crankcase. And it's the compression stroke that follows. With this engine the intake stroke follows and it is being messed up by the missed power / exhaust stroke.

Gordon I believe you observed fuel dripping out your exhaust port. This may explain that. i.e the fuel is being sucked in and immediately blown out the exhaust port. I have a similar problem.
I think you've been lucky that the intake stroke actually pulled in some new charge or there was enough left around after the failed exhaust stroke to get it to run.

I'll admit that if it does fire i.e. you somehow stumble on the exact mixture then in may be self sustaining. But I think the first time it misfires it's doomed to failure unless conditions are just right for it to survive the screwed up mixture.

Let me know if you think I'm missing something. My brain hurts.

You might need to have a look at the video I posted a few posts back that explains the motion of the pistons to see what I'm guessing is the problem. Keeping in mind the action of a misfire condition.

I'm going to have to do some thinking about this before I go any further. I'm not going to so any more work on it without a clear direction. Maybe some sort of mechanically actuated intake valve??

Comments welcome.

Silly question but are you trying to run it the wrong way?

 

[dsage]

Good question. And you had me worried for a minute.
But I went back to the book and I am turning it the way it says to. (flywheel clockwise looking at the motion works).
That said, some two strokes - namely my old full sized marine engine -will run in the reverse direction just fine. (And this engine is as much a two stroke as a four stroke).
SO I went back and analyzed the motion works when turned backward and it seems it goes through reasonable motions but on the power stroke when turned backwards the pistons are moving more together down the cylinder rather than one moving fast to absorb the rapid expansion of gasses.
You have tweaked my interest in what might happen if I turn it "backward" though. I'll try flipping it by hand tomorrow but to give it a good try I'll have to make a new one-way bearing starter for my drill. I have a spare bearing. I'll see what happens.
Thanks for asking.

PS> I'll have to change the spark timing as well. Luckily I made that easy enough to do.

 

[dsage]

Silly question but are you trying to run it the wrong way?

teeleevs:

I will get to trying turning it backwards.
BUT

Back in post #98 Awake posted:

Dave, I found myself thinking about the video you posted above, early this morning when I woke up for my nightly OMB* call. Not sure why that was what popped into my head, but as I pondered it, I got to wondering about the way the spark plug is exposed just before it fires, and the fact that it is a very narrow space between the pistons that is acting as the combustion chamber. Is it possible that the mixture does not have a chance to come into adequate contact with the spark? This may be a complete red herring, but just in case ...

*OMB = old man's bladder.

SO I woke up this morning thinking about this and I think there might be something to it. Especially since I Have no firing at all.
At least it's something I have to consider. I'm not sure how to address it though.
To explain what Awake was trying to point out is that In order to get the compression up I ended up with only 1/16" between the pistons at TDC. Compounding the issue is that the spark plug is off to the side of the cylinder and not even in the combustion "chamber" (there really isn't one).
ALSO I had to reduce the spark plug size to 1/4-20 from the 10mm original plug because of it's new location and it's sitting on a boss not much bigger than the plug. So there is even less volume of charge being collected at the tip of the plug.
This is compounded again by the fact that when the left piston goes to the right the spark plug is literally "hanging out in free air" behind the piston. So my thinking is that if anything, there is fresh air trapped around the spark plug not combustible charge when the pistons come together briefly in front of the plug again. At a minimum it will not be what's between the pistons. Then there is the issue of if the combustion will propagate between the pistons.
The latter issue is somewhat taken care of by the multi-spark ignition (another contributor suggestion) because the spark is now still firing when the piston gap increases as they move apart and are well down the cylinder.
SO I don't really know if this is the real issue but it's something I need to look at. I don't really know how to solve it considering the current configuration of the moving parts.
On the original design there was a huge space between the pistons at TDC becasue there was almost no compression. This might be why the original works to some degree.

I'll have to think on it some more.
Some more suggestions are welcome.
Thanks

The spark plug is off to the side of the cylinder. The gap between the pistons is only about 1/16"

The spark plug is in free air at the end of the power / exhaust stroke. Both pistons are at the right.

 

[stragenmitsuko]

Yes the position of the spark plug is at a far from ideal location .
Just thinking out loud , what if something like a couple of tig needles would be inserted from opposite sides of the cilinder meeting each other in the middle .
That would put the spark where it should be .

Hope you nail it , would be a shame to make such a beautifull design go to waste .

 

[Gordon]

I think that you are correct on your vacuum theory if there is no firing. Unburned fuel keeps accumulating until it is super flooded.

I tried running the engine in reverse and that did not work. It had better compression in reverse rotation but it did not fire.

I did find that the engine seemed to fire better when the timing was very retarded. Perhaps 15° ATDC. It would not run but it fired more consistently. That may be because of your theory of vacuum on the return stroke which dilutes the mix from the unburned gasses as the chamber becomes larger the mix air/fuel gets closer to a combustible mix. I also wondered about the "fresh" air trapped in the spark plug boss. I shortened up the boss to minimize the dead air space. My spark plug boss is threaded into the main cylinder so I could play round with that.

Like everything else I tried nothing seems to make any appreciable difference. Mostly the engine looks cute as a shelf elf.

Gordon

 

[dsage]

stragenmitsuko"
The rods in the cylinder is a thought except that (unlike a normal engine) both pistons move down the cylinder so the left piston would clip the rod(s) off. Thanks for thinking about the problem though.

I had thought for about a nano second that the spark plug could go in the bottom of the left piston with a hole in the piston top for the electrode to peak out. But then the spark plug would have to ride along up and down the cylinder with the piston. (Plus there is no room inside the piston because of the con rod.). Lots of silly thoughts come to mind when you're grasping at straws.

Gordon:
Thanks for giving me more insight into your engine and what you've tried. I did notice in my animation that it might have more of an intake stroke running backwards (i.e. more compression). Interesting that you noted that. Did you also change the timing? to go along with that. Because it would then be on the wrong side of TDC.
But I think the biggest killer of the backward running idea is that the pistons are more or less moving together down the cylinder and if it does fire I think the pressure and the mechanism might actually stop the engine.

It's also good to know that (if you are running the normal compression) that yours does not fire reliably with more gap between the pistons at TDC (where mine has almost none). This tells me that it could still be a mixture thing because when yours it at TDC with the large(r) gap the spark should be able to get to the mixture. I have somewhat accomplished a work around for this with the multi-spark. But that just leads to ignition when there is less compression so it's counter productive.
The mis-fire leading to vacuum is certainly going to mess things up.
I had a thought of an electrically controlled intake valve that is locked out until the real intake stroke starts (BDC). Then the vacuum can be released out the exhaust valve when it opens. Then unlock the intake and let it work normally.

I had thought about the spark plug and it's distance from the pistons when I made the boss. The tip of the plug is just flush inside the cylinder before I added the copper gasket. So it's still about as close as I dare make it. (My cylinder walls are very thin so I couldn't screw it into the wall itself?

All food for thought.
Thanks for participating.

 

[awake]

So my thinking is that if anything, there is fresh air trapped around the spark plug not combustible charge when the pistons come together briefly in front of the plug again. At a minimum it will not be what's between the pistons. Then there is the issue of if the combustion will propagate between the pistons.

Yes, this is what I was thinking, but you've said it much better / clearer than I did.

Here's a thought, FWIW - what if you clipped the corner of one or both pistons, so that there is a bit more of a chamber at TDC? Particularly if you cut off a corner from the left-hand piston, it would expose the plug sooner, perhaps allowing some swirl to occur at the plug - ?

Not that I actually know what I'm talking about ... !

 

[dsage]

Not that I actually know what I'm talking about ... !

No, actually that's a good suggestion. I just wish I cold get it to at least respond a bit first. I'm surprised that it does not fire at all especially with the multi-spark is firing from just before TDC to well beyond when there should be mixture moving around. Even though the compression is going down. I think there's a nasty combination of things going on.

And for the doubters that think I don't have a spark. (a logical thought).
I put my spark tester (gap) in the circuit and there is lots of spark. i.e it's jumping a gap in series outside the plug much larger than the plug gap on it's way to the plug.
And no it's not sparking to ground outside the plug. I looked at it in the dark.

I'm going to ponder a bit on simple things to try before I get carried away modifying any parts.

Thanks for the suggestions.

 

[Cogsy]

I forget what engine it was, but I once built an engine that for some reason had the end of the plug down a fairly deep 'hole' in the combustion chamber. I was concerned that even if the plug got enough of a charge to fire once then it would be surrounded by exhaust gases and it wouldn't fire again. It turned out to run just fine though, which suggests to me that your plug probably isn't trapping fresh air around it's electrode (I think that's what you're thinking). It looks like an interesting engine and I hope you get it sorted.

 

[teeleevs]

stragenmitsuko"
The rods in the cylinder is a thought except that (unlike a normal engine) both pistons move down the cylinder so the left piston would clip the rod(s) off. Thanks for thinking about the problem though.

I had thought for about a nano second that the spark plug could go in the bottom of the left piston with a hole in the piston top for the electrode to peak out. But then the spark plug would have to ride along up and down the cylinder with the piston. (Plus there is no room inside the piston because of the con rod.). Lots of silly thoughts come to mind when you're grasping at straws.

Gordon:
Thanks for giving me more insight into your engine and what you've tried. I did notice in my animation that it might have more of an intake stroke running backwards (i.e. more compression). Interesting that you noted that. Did you also change the timing? to go along with that. Because it would then be on the wrong side of TDC.
But I think the biggest killer of the backward running idea is that the pistons are more or less moving together down the cylinder and if it does fire I think the pressure and the mechanism might actually stop the engine.

It's also good to know that (if you are running the normal compression) that yours does not fire reliably with more gap between the pistons at TDC (where mine has almost none). This tells me that it could still be a mixture thing because when yours it at TDC with the large(r) gap the spark should be able to get to the mixture. I have somewhat accomplished a work around for this with the multi-spark. But that just leads to ignition when there is less compression so it's counter productive.
The mis-fire leading to vacuum is certainly going to mess things up.
I had a thought of an electrically controlled intake valve that is locked out until the real intake stroke starts (BDC). Then the vacuum can be released out the exhaust valve when it opens. Then unlock the intake and let it work normally.

I had thought about the spark plug and it's distance from the pistons when I made the boss. The tip of the plug is just flush inside the cylinder before I added the copper gasket. So it's still about as close as I dare make it. (My cylinder walls are very thin so I couldn't screw it into the wall itself?

All food for thought.
Thanks for participating.

I'm thinking that the spark plug should be in its own little combustion chamber so there is a rush of mixture into the chamber as the piston uncovers the hole. Just a thought from my Diesel experience.

 

[dsage]

I'm thinking that the spark plug should be in its own little combustion chamber so there is a rush of mixture into the chamber as the piston uncovers the hole. Just a thought from my Diesel experience.

Yes, that's a thought. Compression is a real problem on this engine and anything that increases the combustion chamber volume would lower the compression. Also I'm trying to solve the problem without doing any major work on the engine. An added combustion chamber would mean work on the cylinder etc. etc. (it never ends.).
Also I think my multi-spark ignition being fired for along period when the pistons open up is sort of like a huge combustion chamber. Somewhere in there I figure it should fire.
It's all guess work at this point. But I'll keep your suggestion in mind. Thanks

I have something I'm going to try soon to eliminate the vacuum issue on a mis-fire (I hope). I'm pretty sure it is one of the issues I need to eliminate if I'm going to move forward.
I did a lot of flywheel twisting today starting from various positions and have proven that there is a pretty serious vacuum on the carb being created on what is supposed to be the power stroke during a mis-fire condition. IT's many time greater than the proper intake stroke vacuum. And all that fuel / air is pushed straight out the exhaust a few degrees of flywheel rotation later.
I need to do some work around the house so I won't be working on it so often.
Later.

 

[teeleevs]

Yes, that's a thought. Compression is a real problem on this engine and anything that increases the combustion chamber volume would lower the compression. Also I'm trying to solve the problem without doing any major work on the engine. An added combustion chamber would mean work on the cylinder etc. etc. (it never ends.).
Also I think my multi-spark ignition being fired for along period when the pistons open up is sort of like a huge combustion chamber. Somewhere in there I figure it should fire.
It's all guess work at this point. But I'll keep your suggestion in mind. Thanks

I have something I'm going to try soon to eliminate the vacuum issue on a mis-fire (I hope). I'm pretty sure it is one of the issues I need to eliminate if I'm going to move forward.
I did a lot of flywheel twisting today starting from various positions and have proven that there is a pretty serious vacuum on the carb being created on what is supposed to be the power stroke during a mis-fire condition. IT's many time greater than the proper intake stroke vacuum. And all that fuel / air is pushed straight out the exhaust a few degrees of flywheel rotation later.
I need to do some work around the house so I won't be working on it so often.
Later.

I fear that the combustion chamber will remain filled with spent gasses with no way of it scavenging, so the spark plug would have to be near the cylinder so the mixture would rush past it to repressure the chamber.

 

[dsage]

I understand your concern and it could be an issue at some point. I can cross that bridge when I get there.
But you gotta burn something first to get trapped spent gasses. And we're not there yet. No pop yet.
On the exhaust stroke there is even more tight clearance between the pistons - also round about 1/16" so I think all the gasses will be getting out ok when there are any.
The vacuum in the cylinder on what's supposed to be the exhaust stroke (when it doesn't fire) is real. I hooked up an open ended hose with fuel in it to the carb and opened the needle valve and the throttle. When I spin it over by hand just in the area of the exhaust stroke it makes a huge suction and the fluid comes right up the tube to the carb. All that charge would be going right out the exhaust an instant later when the exhaust port is uncovered.
The real intake stroke an instant later seems normal and barely moves the fluid.
I suspect the real intake charge has to be getting messed up.

I'm working on something pretty complicated to try to get rid of the problem but it's going to take some time.
I'll try this first. Hopefully it results in a normal charge that fires. I'll write down your concern.
Thanks for the feedback.

 

[Ken Brunskill]

Dave,

I am building Dave Perallt's design of an Atkinson Differential, at this time have machined all the parts except the pistons and rings, did a trial fit (see photo), during this trial fit discovered I was short some bearings, which arrived yesterday. Our mutual friend Paul D. suggested you contact me, and got me reacquainted with the HMEM posts, and I thank you for reaching out to me.
Having read all the posts, and your progress, the one thing that I see as a potential and worth some thought is, on the Perrault design the POWER piston (3/4" Dia.) has three (3) 1/8" x 45 deg angles at the top of the piston, one centered in line with the Con-Rod, the other two at 20 deg from the one on center, the spark plug is at 20 deg., which would put the upper most 20 deg. chamfer of the piston in line with the Spark Plug, with the other two below it to align with the exhaust port. This appears to square with the comments & observations in earlier threads. I'd like to share the details with you but do not want to run afoul of copyright issues.

 

[dsage]

Nice looking work. Seems to be a fair bit different.
I see no reason you can't take a picture of the finished piston you made (if possible) and post it here.
A picture is worth a thousand words. But no drawings please.

Gordon:
Is this the engine design that you have combined with the Gingery to arrive at what you have?
Comments?

 

[Gordon]

Nice looking work. Seems to be a fair bit different.
I see no reason you can't take a picture of the finished piston you made (if possible) and post it here.
A picture is worth a thousand words. But no drawings please.

Gordon:
Is this the engine design that you have combined with the Gingery to arrive at what you have?
Comments?

That is different than the design I partially used. The design i partially followed was by Brooks Pendergast. I also would like to see the piston design. It sounds like it is completely different than either engine we have. See post #21 for a picture of mine.

 

[dsage]

Thanks Gordon:
So yet another one lurks out there
Ken contacted me by email with some details.
I think it would be safe to post a piece of one of his CAD drawings of the piston.
He posted above a description of the angles involved. I'm hesitant to implement that change until I really have no other choice. Compression is a precious commodity on this engine and it looks like it would steal quite a bit.
His engine must be considerably better in that respect.
Perhaps I should have started with those plans.
Looks like what Teeleevs is suggesting.

 

[Ken Brunskill]

Gave this design some more thought, and thought that a bit of a description and/or explanation might help one to understand what my opinion is of this design.
1. the upper angle coincides or is perpendicular to the Spark Plug hole & Intake Port.
2. The small center angle is essentially a path for the exhaust to reach the Exhaust Port.
3. The lower angle coincides or is perpendicular to the Exhaust Port.

Dave rightfully is concerned that such a large 'relief' might negatively affect to compression, so on my build I'll start small and increase in 25% increments (looks as if I'll need to fabricate a holding fixture to accomplish that.) Dave suggested I start a thread on my build, will follow up on that, not nearly as well as he has done, so be patient with this old man.

 

[bluejets]

Compression is a precious commodity on this engine and it looks like it would steal quite a bit.

There is a way to overcome this by using methanol fuel with castor (castrol M) as a lubricant mix and varying amounts of nitromethane.
There is also available, commercial mix at any hobby shop which uses synthetic oils.
The nitro is substituted for the equal amount of methanol in the mix.
On ignition, the nitro releases oxygen a bit like supercharging only with fuel.
With the rc aircraft engines, a 10% nitro mix usually cleans up idle and transition etc. and desensitizes the needle setting, all without any compression setting alterations.
By some quirk of nature, increases of 10% seem to have the most noticeable effect however, once at the 20% mark and above, normally compression lowering is required.
I have used it in volumes of up to 60% on the small cox engines but for others, 20% was the limit.
Thought it might help to mention.
Just one drawback is , burning causes acid release so necessary to use after run oil or simply flush out the internals after using it.
Can't imagine you'll be using bucketfuls so cost would be irrelevant.