Another Atkinson Differential build

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Ken:

I've had good luck stopping ring rotation by controlling the amount of clearance between piston groove and ring.
I've done my set of rings using trimble method, then machining them in a very accurate and well centered spigot with the right dimensions based on the ring and bore and ran my engine for several hours to the point of having some pops but no sustained run, so I opened the pistons to replace rings and they were in the same location where they were installed..both pistons had decent be in of rings but compression was low (1round 10-12 PSI) I think we need to achieve over 20 PSI's to get a decent pop..
Funny thing is most times I start with 20 plus PSI and then trying to bed in the rings, compression goes down to almost 5 PSI and then slowly start to come back up when the bed in the cylinder.
At this point in time, I think the need for pinned rings is not necessary. Good rings and controlling the ring groov gap to ring to be under 0.005 in would be sufficient to stop the rotation, my rings are 0.062 thick once polished on the sides and the groove in the piston is 0.063 in.

Kind regards
Well ,I have the blanks ready to make new rings using the 'No Heat Treat method'. Where the rings are over size to begin with, compressed to cylinder Dia. for machining both ID and OD at a preset end gap. Will go to this step if don't get a POP, and the vacuum and compression indicate the need.
 
Ken, no matter what method you use to make your rings you can always fix them the way I did, put some extra fine lapping compound in the cylinder and slide the pistons with rings up and down in it, disassemble and use an ultra-sonic cleaner to be sure to get all the grit out of everything, reassemble, first pop. (works better with pinned rings so you can rotate a bit while lapping up and down, but I didn't and it still worked for me) Pete.
 
Ken, no matter what method you use to make your rings you can always fix them the way I did, put some extra fine lapping compound in the cylinder and slide the pistons with rings up and down in it, disassemble and use an ultra-sonic cleaner to be sure to get all the grit out of everything, reassemble, first pop. (works better with pinned rings so you can rotate a bit while lapping up and down, but I didn't and it still worked for me) Pete.

Thanks for that suggestion Pete, this is getting as interesting as the build itself 😏 . I hope that others are enjoying these exchanges.
 
That lapping is what they did in the Hepworth and Grandage factory - last operation before washing, packing and selling their rings.... They called it "barrelling": perhaps as the cylinders they made for the purpose were the "barrels", or because the rings had a microscopic "barrel" profile afterwards? I can't remember...
K
 
Try toothpaste for the finest lapping compound you probably have ready access to. (It is used for polishing diamonds as well). Toothpaste also makes good tapping lubricant, especially for Aluminium.
Or "Brasso".... or T-cut car body-paint polish.
K
 
There's nothing like a bit of hand-cranking for satisfaction. Having always been an Engineer - designing and preparing stuff for others to make - I seriously enjoy doing the calcs, then making the stuff myself that proves my calcs are good. So personally I avoid CAD, CNC, digital readouts, digital micrometers and verniers, etc. and much prefer the slide-rule, paper and pencil, drawing board, etc. and the micrometer with real marks, or vernier caliper with a vernier scale. AND a good file and hacksaw! I just love the tools and skills, both mental and physical. It is a hobby, not a business, isn't it? Something so we can play and make things as we did as kids?
But for the "technophobes" - enjoy the CAD and CNC digital world....
Just remember the artisan skills that have a natural beauty in "hand-made", rather than the modern "Machine made". Sometimes the imperfections give the object "personality", or "character"?
Enjoy life! - It's the only one we know we experience.
K
 
Thanks for that suggestion Pete, this is getting as interesting as the build itself 😏 . I hope that others are enjoying these exchanges.
The discussion is indeed interesting but unfortunately much of the discussion is relevant to this engine but a lot of the suggestions apply to a standard two stroke or four stroke design and the Atkinson Differential is in a completely category. Talk of compression, vacuum and ring seating etc. are almost impossible to test with this design. After spending far too much time trying to get one of these to run it is obvious to me that the design in intriguing but is not really a practical design. A very minor change makes the difference between running and not running. I seriously doubt that the design could actually be used to power anything.
 
Here's a tip i use with Piston Rings. I use the Trimble method, but what I do is half the width of the ring and place two rings in the groove with the gap set at 180 degrees apart a couple of strokes in the bore and they seal very well. Give it a try as if you follow the Trimble method they work and seal very well. Cheers.

John
 
Has anyone plotted the "pressure graph" and tried to determine theoretical efficiency of this cycle? - From what you are saying, there is some intriguing geometry and motion, but the "engineering" of the gas cycle is in question?
Before I start from scratch - never having done this before on an infernal Otto or Carnot cycle.... - or steam for that matter - I am sure some of the Engineers out there have done so and some may have the software to make it easy?
For a High Voltage circuit breaker (back in the 1980s) I designed a new configuration of linkage to non-linearise the mechanical advantage (inverse of velocity ratio) to re-engineer the motor-to-contacts ratios so that the motor did about 3/4 of it's work while the contacts did 40% of their motion, then in the last 1/4 of the motor's stroke the contacts were decelerated to rest over the last 60% of their motion. I worked with a Doctor of Mathematics who evolved a computer programme so I could "tweak - iteratively - the parameters between various chambers of the motor so achieve a 2-stage deceleration that enabled the circuit breaker to become an off-load interrupter, by increasing the number of cycles from 500 to 5000, before mechanical servicing was required. The computer programme was eventually turned "inside-out" so instead of my numerous iterations to optimise loads from the gas flow and pressures inside the motor, I input the required motion result and the programme did the iterations - working backwards - to a solution for the various orifi that gave the motor its required force variation through the stroke. - But I take no credit for the maths and computing! - I think this can now be done on a spreadsheet.... but Desktop computers were still at the Amstrad 64kB capacity.... or HP 500kB capacity... ! - An "Apple" was a piece of fruit... Some younger people cannot imagine how we put a dozen men on the moon 10 years before we had computers!
Is there someone out there who can do the Maths for this Atkinson differential engine? - I have no plans, but will "have a go" if no-one else is bothered, and wants to send me the plans?
It sounds like a problem yet to be solved? (Non-linear Geometry and varying volume gas chambers). I.E. following Ken's suggestion of 16th Feb? - (Maybe I have missed a pressure/volume analysis somewhere in the thread?)
Ken Br. - maybe you can send a geometric plan of the linkage with dimensions of arms and pivot locations, - a pdf will do - and I'll get on my drawing board. (No CAD here! - I was born when we used pencils and slide rules (Mechanical logarithm calculators). "Computers" were the girls NASA used to do the calculations and "Compute" the trajectory of rockets! - Strangely, I enjoy the mental challenge of using my brain to work things out....
Ken
 
STUPID me... Some of it has been done - Atkinson cycle - Wikipedia. Just a matter of using my computer as a reference book!
https://upload.wikimedia.org/wikipedia/commons/c/cb/Atkinson_Opposed_Piston_Engine.gifBy MichaelFrey - Own work, CC BY-SA 3.0, File:Atkinson Opposed Piston Engine.gif - Wikimedia Commons
But it doesn't show valve timing?
Also:
1601459442178.png
By Richard Rusnák (Rios) - Own work, Public Domain, File:T cycle AtkinsonMiller.png - Wikimedia Commons
The ideal Atkinson cycle consists of:



A few curious comments in WIKIPEDIA... suggest that the apparent efficiency increase of Atkinson cycles, have been achieved now due to modern design and manufacturing of "conventional" Carnot cycle engines... Is this an example of Darwin's laws of natural selection? - We make the cheapest things that sell best? - And scrap the rest?
I wonder if the limitations of this engine were simply the Capital cost? - or perhaps the huge energy required to accelerate and decelerate the large levers of the main linkage? - This all requires large bearings (for durability) and substantial beams for fatigue resistance. The substantial beams in themselves are major inertias driven by the pistons and crankshaft so the bearing losses will be larger than in a "simple" crank-to-con-rod-to piston configuration. The extra cost of these, plus the structure, possibly were impractical compared to the fuel savings, for a buyer's budget?
But all credit to Atkinson as the British Gas Engine company did make a lot of the Cycle and Utilite engines...
Reading Wiki, I wonder if much of the running problem enjoyed by Ken B is due to valve losses, rather than piston losses? - Something I have experienced on both model Steam engines and full sized motorcycle and car engines. Perhaps the piston seal can be proven as the culprit by a single wrap of PTFE tape around a piston ring (Top one?), then check for a significantly improved compression? That trick worked for me on a worn single acting oscillating steam engine. It confirmed the oscillating face-joint spring force was too low, then when I fixed that it proved the pistons were leaking too badly and needed new rings. But as a trick it only works for a few strokes.
K
 
Steamchick:
All interesting Ponderings.
But were you researching the Atkinson "Cycle" engine or the one we are discussing here which is the Atkinson "DIFFERENTIAL" engine? Two very different beasts.
The differential engine we are talking about has two pistons in the same cylinder which for the most part work against each other.
I built the Atkinson cycle engine and it works fine. Not exactly efficient IMHO but it at least runs.
As for some Engineer that has studied it. I would assume Atkinson did. Unless it has been a scam since the beginning. If it weren't for the fact that there are a couple I've seen running I wouldn't believe it could run based on the workings I've observed in the one I'm building.
BTW the version Ken is building is a bit of a cheat from the original design. It has some extra valve gear that may make it better.
 
Cool. I didn't read past the animation mentioned above. Looks like in the third attempt he abandoned the wacky mechanisms in favor of what looks like a two stroke.
So if you have what you need to analyze the differential version I'd be interested to know if it is theoretically able to run. :)
Practically it doesn't seem so.
 
Last edited:
The discussion is indeed interesting but unfortunately much of the discussion is relevant to this engine but a lot of the suggestions apply to a standard two stroke or four stroke design and the Atkinson Differential is in a completely category. Talk of compression, vacuum and ring seating etc. are almost impossible to test with this design. After spending far too much time trying to get one of these to run it is obvious to me that the design in intriguing but is not really a practical design. A very minor change makes the difference between running and not running. I seriously doubt that the design could actually be used to power anything.
Gordon,
True enough about this thread 'Another Atkinson Differential Build' being relevant to the Atkinson Differential, curious of why one would consider it "unfortunate", after all it is specifically about this design.

No offence intended, just a thought. For me, and I hope others, the post have provided some delightful exchanges, all friendly and very informative.

I'd exchanged a few email's with dsage, and started this thread at his suggestion. Based on his contributions, I'd venture to say he is enjoying the exchanges as well.
 
STUPID me... Some of it has been done - Atkinson cycle - Wikipedia. Just a matter of using my computer as a reference book!
https://upload.wikimedia.org/wikipedia/commons/c/cb/Atkinson_Opposed_Piston_Engine.gifBy MichaelFrey - Own work, CC BY-SA 3.0, File:Atkinson Opposed Piston Engine.gif - Wikimedia Commons
But it doesn't show valve timing?
Also:
View attachment 119727 By Richard Rusnák (Rios) - Own work, Public Domain, File:T cycle AtkinsonMiller.png - Wikimedia Commons
The ideal Atkinson cycle consists of:



A few curious comments in WIKIPEDIA... suggest that the apparent efficiency increase of Atkinson cycles, have been achieved now due to modern design and manufacturing of "conventional" Carnot cycle engines... Is this an example of Darwin's laws of natural selection? - We make the cheapest things that sell best? - And scrap the rest?
I wonder if the limitations of this engine were simply the Capital cost? - or perhaps the huge energy required to accelerate and decelerate the large levers of the main linkage? - This all requires large bearings (for durability) and substantial beams for fatigue resistance. The substantial beams in themselves are major inertias driven by the pistons and crankshaft so the bearing losses will be larger than in a "simple" crank-to-con-rod-to piston configuration. The extra cost of these, plus the structure, possibly were impractical compared to the fuel savings, for a buyer's budget?
But all credit to Atkinson as the British Gas Engine company did make a lot of the Cycle and Utilite engines...
Reading Wiki, I wonder if much of the running problem enjoyed by Ken B is due to valve losses, rather than piston losses? - Something I have experienced on both model Steam engines and full sized motorcycle and car engines. Perhaps the piston seal can be proven as the culprit by a single wrap of PTFE tape around a piston ring (Top one?), then check for a significantly improved compression? That trick worked for me on a worn single acting oscillating steam engine. It confirmed the oscillating face-joint spring force was too low, then when I fixed that it proved the pistons were leaking too badly and needed new rings. But as a trick it only works for a few strokes.
K
K, I will contact the originator of the plans I have for the 'Mini A' Atkinson Differential to obtain permission to share those specifications for this purpose, I respect his proprietary rights and would not want to offend him, his plans are well done and reasonably priced.
 
Gordon,
True enough about this thread 'Another Atkinson Differential Build' being relevant to the Atkinson Differential, curious of why one would consider it "unfortunate", after all it is specifically about this design.

No offence intended, just a thought. For me, and I hope others, the post have provided some delightful exchanges, all friendly and very informative.

I'd exchanged a few email's with dsage, and started this thread at his suggestion. Based on his contributions, I'd venture to say he is enjoying the exchanges as well.
Yes. Unfortunate was not really the proper choice of words. My point was the some of the discussion was relevant to a standard engine design but not the differential design. Some of the points relevant to measuring compression, vacuum, stroke etc. on a standard engine are difficult or almost impossible to obtain in an engine where the intake stroke is just a fraction of an inch and the compression stroke is dependent on one piston moving faster than the other. Some of the suggestions in this and other discussions show that it was not obvious that this design is unique.

I also got into an offline discussion in an earlier thread with Ray and we certainly did end up batting a lot of ideas around but never came to any firm conclusions.

There. I think I just dug a deeper hole for obscurification. Sorry if I offended anyone. That was certainly not my intention. You are certainly correct that many of these discussion can be very interesting and useful.
 
Ken:
I enjoy the general discussion. It's all good. But sometimes I wonder if it's getting off topic so I try to be sure we are all talking about the same thing. I get confused.
I am waiting for what to try on my engine next. You are my guinea pig :cool:
The pressure is on you Ken :)
All in good fun.

In the end it is your thread so I should just step aside and let it take it's course as you see fit. ;)
Thanks all.
 
Last edited:
Although I worked as an Engineer for 45 years, it wasn't studying combustion cycles. Does anyone on the thread understand the "Atkinson cycle" - as shown in the "Pressure/Stroke" diagram? As I do not have a drawing of valve gear - just Wiki sketches of crank to piston linkages - and I am learning from your best info in this thread. - Dsage: Yes, you are correct - the diagram shows the Atkinson Cycle, but as it shows an engine "cycle" it could be related to any configuration of engine that creates the particular "cycle". In Innocence, I read it that the "Atkinson differential engine" is one configuration (probably an early configuration?), followed by the "Atkinson Cycle engine", and latterly the "Atkinson Utilite engine"...? - All of which emulate the "Atkinson cycle" for combustion.
To be able to determine the thermodynamic cycle for an engine needs some knowledge of the valve and ignition timing, as well as the configuration of piston motion - versus shaft rotation, or other datum. As I wasn't planning to "buy the book", I wonder if any of you kind folk have that information in a simple diagram or table? - My interest is purely intellectual, as I am not planning to make a version of any of these engines. I only have experience of "mid-to-late 20th century conventional" engines, and their components, so am curious about some earlier designs.
Ta,
K
 
No diagram, but I can verbally describe it:

The Atkinson cycle is designed around the notion that what makes an engine efficient is the expansion ratio, not the compression ratio per se. As it was conceived at a time when compression ratios were, necessarily, low to avoid knocking, it solves the problem of low compression ratio & high expansion ratio by having a differential action -- the intake & compression strokes are short, followed by long combustion & exhaust strokes, then repeat.

It's basically the Otto cycle, but with different strokes (for different folks?).

My thermodynamics isn't strong enough, but because the expansion stroke is larger than the intake stroke, it ought to be releasing exhaust that's cooler than could be achieved in a fixed-stroke Otto cycle engine. I suppose that in theory, with a whole lot of expansion, the exhaust valve could open at zero pressure differential, or you could even have the thing pull a vacuum at the end of the expansion stroke -- presumably you could have a vacuum and a cylinder filled with rarified air that's cooler than atmospheric pressure, although then you'd be losing efficiency to pumping losses, not gaining it.

I think, though, that the real reason for the Atkinson Differential Engine was to dodge Otto's patents.
 
Back
Top