Another Atkinson Differential build

[Gordon]

I never got the engine to run for more than about two minutes. At this point it is on the shelf and I have "borrowed" some parts like carburetor and ignition module. The engine seems to run best if timing is ATDC. The carburetor adjustment is super sensitive. 1/64 turn on the needle makes the difference between running and not running. As the engine heats up the carburetor has to be readjusted. I have tried several different throttle body sizes for the carburetor but nothing works well. As many of you know once you get most engines to run a minor adjustment in fuel mix or ignition timing makes the engine run better but does not necessarily make the engine not run.

As I said. Phase of the moon and plane alignment.

 

[peterl95124]

observations regarding Dave's (dsage) recent video post

1) I'm assuming both intake and exhaust have at least one-way-valves if not cam operated valves,
otherwise both intake and exhaust are open during the intake stroke

2) the ports on the cylinder look minuscule compared to for example a typical "model RC airplane engine" whose ports form a ring around the cylinder that is perhaps 50% open for an old engine and probably 75% open for a Schnuerle ported engine

so my scientific-wild-assed-guess is that these engines need massively widened cylinder cut outs for the ports, although I acknowledge that the wide ports on RC engines are for high RPM and power which isn't the goal with an Atkinson, but still I think the carburetor should be the smallest cross section to the flow and everything else should be much larger

also, since this is a low RPM and small displacement engine getting the carburetor to work properly is most likely problematic, switching to propane to eliminate carburetor "draw" problems has been done by many modelers and seems to work well.

just my 1.414-cents,
Peter

 

[Ken Brunskill]

A first attempt... just a picture of my representation of Chamber Volume versus Pressure.
The peak pressure is just after ignition.... I have assumed "No losses", but "guestimated" a chamber volume at max compression based on 4 bar attainable pressure, and to simplify things uses a 2cm diameter bore. The peak pressure after firing is just a number to make a picture.
2 things: As Atkinson intended, the Pressure-combustion cycle (Top loop) is very large compared to the intake cycle, (bottom loop). Thus the limited charge will do a lot of work efficiently in expanding to ~25% of it's original "fired" pressure before the exhaust is released. Of course, as Otto and Carnot both knew, the compression ratio is "almost everything" and modern petrol engines running at 10:1 compression ratio are twice as efficient as the numerical model I have used.
I haven't plotted curved lines - just for speed to get something of interest here.
View attachment 119913
Hope this is interesting to some?
More when I have done some tarting-up.
Ken

Steamchick, et al;

Being that I am an novice at Solidworks, I need to get advice from their forum on how I can get the 3 deg. step increments you ask for. So silence from me on this topic for a bit, is expected. Then there is actually doing it , as I am still trying to get my engine to provide signs of life.

I think there are a few members of our club following this thread, hopefully it will spur some more interest and construction of a model of this engine, knowing machinists as I do, we are prone to think outside the box, which in this case may well bring some fresh ideas to light, ha, some might even work.

Ken B

 

[Ken Brunskill]

I never got the engine to run for more than about two minutes. At this point it is on the shelf and I have "borrowed" some parts like carburetor and ignition module. The engine seems to run best if timing is ATDC. The carburetor adjustment is super sensitive. 1/64 turn on the needle makes the difference between running and not running. As the engine heats up the carburetor has to be readjusted. I have tried several different throttle body sizes for the carburetor but nothing works well. As many of you know once you get most engines to run a minor adjustment in fuel mix or ignition timing makes the engine run better but does not necessarily make the engine not run.

As I said. Phase of the moon and plane alignment.

Gordon,
I know that you have a different variation (Gingery I think) from the Perreault design that jquevedo & I are building, but the 1/64 turn on the needle indicate to me that you have a different taper on your needle than I have, could you elaborate on that? On this aspect I doubt that model design makes much of a difference (piston diameter might be a factor so please. provide that as well.)

Ken B

 

[Gordon]

My engine is roughly following a design by Brooks Pendergrast. Not an exact copy but I used most of his locations for the pivot points and length of lever ratio on the pivot arms. The piston is 1 1/8 dia. The needle on the carburetor is a #18 darner needle.

Attached is a layout of the action. That will probably just make everything more confusing.

 

[Ken Brunskill]

observations regarding Dave's (dsage) recent video post

1) I'm assuming both intake and exhaust have at least one-way-valves if not cam operated valves,
otherwise both intake and exhaust are open during the intake stroke

2) the ports on the cylinder look minuscule compared to for example a typical "model RC airplane engine" whose ports form a ring around the cylinder that is perhaps 50% open for an old engine and probably 75% open for a Schnuerle ported engine

so my scientific-wild-assed-guess is that these engines need massively widened cylinder cut outs for the ports, although I acknowledge that the wide ports on RC engines are for high RPM and power which isn't the goal with an Atkinson, but still I think the carburetor should be the smallest cross section to the flow and everything else should be much larger

also, since this is a low RPM and small displacement engine getting the carburetor to work properly is most likely problematic, switching to propane to eliminate carburetor "draw" problems has been done by many modelers and seems to work well.

just my 1.414-cents,
Peter

To say, I am a step ahead on that would be an exaggeration, maybe 1 #9 foots worth, I am constructing fittings to try both propane and mapp gas. This inspired by dsage's comment on having "fuel dripping out of the exhaust port" (last Tuesday). Hope to have that done by next weekend.

 

[Ken Brunskill]

steamchick et al;

Met today with a mechanic friend we discussed the issues regarding the Atkinson Differential, he is going to allow me to use some of his sophisticated analysis tools in his shop, and intriguing tool is a 'Smoke Tester' where smoke gets blown into the Spark Plug hole. Will post pictures of that test, in theory this may be a simple enough of a device we 'handy men' can make a home made unit for ourselves.

Steamchick: I sent you an message separate from this thread, we need to work out the details of the data in the Excel spread sheet. I have learned what to do to get the model to stop at 3 deg. increments, & have started collecting the data, but don't want to go any further until we are in sync on the spreadsheet content.

 

[jquevedo]

Hi Guys!
I hope my musings - as attached - will give you something to ponder? (Hope you can open the file!).
I may easily be wrong with my observations and ideas, but it may give you some new insight to get your engines running?
I realise that changing valve timing and/or porting is no small task.... but it is just my humble idea.
I'm not offended it it is all trash.
Enjoy?
K

Ken:

Excellent study, diagrams and insight on what may be the key to get the engine to run..

I have put my best effort in cylinder construction and finish, the same goes for the rings, using the best rings produced from the batch for the the engine, after many hours of trying to get the engine to break in rings and cylinders and then reduce the drag from pistons and rings I have made some progress but remain having trouble getting the engine to run, some pops but not enough to make the engine run..

I share your thoughts regarding timing been one of the key issues, specially in low compression engines, when you have more compression, it usually hides some of the imperfections on timing and allows you to run an engine even when lean/slightly flooded.
In this case, the timing of the engine, should be as close to specifications as possible.

Will use the power piston as the one to determine TDC..

I have checked some of the basic specifications for Atkinson engines . This numbers are for not the atmospheric model but the model with mechanical valves, this engine with a sliding valve should belong to the same group.

Intake valve opening at 0 Deg Closing at 70 Deg, compression cycle 70 deg - 220 deg Spark at 140 Deg. Exhaust valve opening 220 - 335 Deg.
I will check the timing of the engine against the numbers you came with in your Diagram and see how they look..

I'm traveling so may take some time to get back to my shop but thanks so much for taking the time to create all these diagrams and calculations, I think we need to address this engine a bit more scientifically to see how can we improve the engine running..

Kind regards

Jaime

 

[Steamchick]

Hi Ken,
I have sent an email this afternoon (Been away from the desk for a while).
Thanks gents for the encouragement. - I hope with Ken's new data I can get a more accurate picture of what is happening. But I realise that sometimes it is labour intensive and time consuming getting a lot of data out of computers, for "numb-nuts" like me to play with, unless you are an IT geek.
Jaime, I think your engine may be a bit different from Ken's? - I also think you started the crank zero at 180 degrees from where Ken's video started. - But good stuff! - I'll see if I can make a picture using your data?
Just give me a little time to get the grey stuff working (A lot of booze last night has left me "cloudy" today).
Thanks guys!
K

 

[jquevedo]

K..

Building same engine Perrault mini Atkinson with slding valve, just realized I was taking the TDC from the wrong piston yesterday based on your input. Need to consider the pistons on the RIGHT ( pumping Piston ?) instead of power piston to set the events.

Ken:
Some of the blow by- dripping fuel in the power piston is due to the geometry of the engine and the fact that the intake port gets uncovered when the power piston travels all the way to the RIGHT right before the Spark happens, this exposes the port to the piston skirt right below the rings, so gas drips from the intake port, I have not found any vaccum leaks, but funny thing is why does the sliding valve has to open the intake again at this point, well to move upward to get in position when the exhaust port will have to open.
BTW this fuel drip tends to get better after the engine has started bedding rings in cylinder. Not sure why.

I think an easier to control valve mechanism would be to separate intake and exhaust valves.
I have been toying with some ideas, need to make some designs ad dimensions based on engine capacity and then figure out if this is practical.

Cheers.

Jaime

 

[Steamchick]

Just going through some of the latest discussions....
There is a comment about port sizes, versus carb choke sizes. My understanding: (any engine tuners with more precise theory or experience please teach us all something better.) - The carb choke is typically 80% - ish of the intake tract before - AND AFTER - the the carb, in order to get the pressure drop in the carb by increasing the air velocity - and thus draw-in the fuel charge (being petrol, etc.). Unless the choke of the tract increases back to original size after the carb it won't work as well. However, to increase the VELOCITY of the charge and get it past the valve (in the case of poppet valve ICEs), it is normal practice for the intake tract to smoothly and uniformly reduce in cross-sectional area to that of the valve-port CSA. This encourages turbulent mixing of the charge in the cylinder, because there will be some residual exhaust gas, and from the carburettor there will be inconsistency of charge (air and fuel) mixing.
HOWEVER - the porting of "conventional" 2-strokes is designed for the intake tract to feed the crankcase first - where mixing is good - then transfer to the cylinder (which is near full volume, albeit with exhaust port open creating some vacuum from the pressure wave in the exhaust) in a very short time (compared to 4-stroke engines). Thus the transfer port area from crankcase to cylinder is not so representative of the Atkinson, that takes the intake through the carb, past the intake valve directly into an expanding small volume cylinder.

So my "theoretical" intake for the Atkinson would be something like Intake area "A", through carb of choke 0.8 x A, into the intake tract CSA = A, followed by intake tract reduction to a CSA of "valve/port" OPEN CSA... which may be of a similar magnitude to the Carb. choke CSA of 80% intake tract.
Does this make sense with the various Designers' drawings you have? The other problem - comparing current conventional engines with the Atkinson differential - is that the volume between the cylinder (port), when the intake valve is open, to the carburettor "jet" position, should be a small fraction of the volume of the cylinder when the valve opens. As the cylinder volume at valve opening is so small, it means the carburettor will need to be very close to the valve, which in turn will need to be very close to the cylinder port. This makes matching the physical shape of the intake to the theoretical shape very difficult. - So I recommend you follow the designer's idea very closely - assuming he made good working designs.
Further, if you are trying to make "suitable" carbs without a design, then you need to copy something like a small carb for an engine of only (say) 3cc engine, even through a 3/4" bore engine like Ken's has a larger displacement (>5cc compression).

I haven't made any "Gas" carburettors, only "wet fuel". So I don't know if any of this is relevant to Propane or MAPP gas. But I can throw in a "2-pennorth" for gas ignition. Butane and Propane are small "stable" molecules that need a lot of spark energy - and correct mixture - to make them fire. "Petrol", Acetylene, and coal-gas ignite very easily, as Hydrogen is free or loosely bound to the fuel molecule, so much less spark energy will work with these. - I think MAPP gas is also a mixture that will ignite using a lower energy ignition? So my comment is to make sure the ignition gives a really good spark - at the pressure within the combustion chamber. If you can make a "clear" spark plug (I have one for 14mm thread from an old "Colour-tune" kit) then you can see what the spark is like when you spin the engine. There are 2 aspects to a good spark: The voltage - you need 4 times the volts at 4:1 compression than in air to break down the molecules and make a spark, or 10 times the volts at 10:1... And the current in the spark. - conventionally, we talk about a "fat" spark - lots of current = lots of ignition energy, or a "weak" spark, where the spark just looks "thin", like a fine hair of spark. Weak sparks don't ignite fuel. No sparks means you need more volts, or a smaller spark gap. I'm sure most know this, but there may be something here for the innocent?

I hope some of this makes sense?
K

 

[Steamchick]

Jaime, I think that for the valve operation, if you make independent levers from independent crank-pins you can drive the valves independently. ( Effectively the same logic for "twin-camshaft" engines as opposed to "single camshaft" engines.). It's just a matter of how authentic you want to be, or how much complexity you like to make. I like to make steam engines with a single eccentric to run one way only, rather than a pair of eccentrics to allow a reversing gear to reverse the engines. But I like to make engines to power generators, which tend to run just one way. (Making lots of levers doesn't excite me so much).
Enjoy!
K

 

[Ken Brunskill]

My engine is roughly following a design by Brooks Pendergrast. Not an exact copy but I used most of his locations for the pivot points and length of lever ratio on the pivot arms. The piston is 1 1/8 dia. The needle on the carburetor is a #18 darner needle.

Attached is a layout of the action. That will probably just make everything more confusing.

Gordon,

A couple of questions:
Observing your comment ". . . roughly following a design by Brooks Pendergrast. Not an exact copy . . . "
And your pistons being 1.125" Dia. - .1372" apart at ignition, presume that to be TDC (Closest point).
D. Perreault pistons are .750 Dia. - .1077" apart at ignition, presume that to be TDC (Closest point).

The DP design is 66.75% smaller diameter vs. the your version of the BP design.
The TDC are 78.49% smaller for the DP design vs. your version of the BP design.

So, the question is:
How much of a difference does this make.
Are both too large a combustion area, or too small?
From your experience of getting yours to run at all is a sign that that ratio is on the right path.

Maybe Steamchick's analysis will shed some light on this, lets hope so!

 

[Steamchick]

Hope my light is a bright as you wish... but usually I am dimmer and duller than expectations!
Thanks for the confidence!
K

 

[Gordon]

Gordon,

A couple of questions:
Observing your comment ". . . roughly following a design by Brooks Pendergrast. Not an exact copy . . . "
And your pistons being 1.125" Dia. - .1372" apart at ignition, presume that to be TDC (Closest point).
D. Perreault pistons are .750 Dia. - .1077" apart at ignition, presume that to be TDC (Closest point).

The DP design is 66.75% smaller diameter vs. the your version of the BP design.
The TDC are 78.49% smaller for the DP design vs. your version of the BP design.

So, the question is:
How much of a difference does this make.
Are both too large a combustion area, or too small?
From your experience of getting yours to run at all is a sign that that ratio is on the right path.

Maybe Steamchick's analysis will shed some light on this, lets hope so!

I am not sure of anything in regard to making the engine run. The drawings/layouts which I attached were one of several I tried to determine what changing the pivot link (the link between the crank and the oscillating arm) would do to operation. I tried making the RH link shorter and the LH longer and vise versa. In some cases I actually made new links but nothing I did made any real appreciable difference. It has been at least a year since I have tried running the engine so I am a little older and a little fuzzy on the actual results once I made minor changes. The thing which seemed to make the biggest difference was making changes in the carburetor throat dia. I think that the carburetor changes mostly reflected the fact that fuel mix is critical and like I said in a previous post once you do not fire the mix keeps getting worse until it no longer will fire. With the low compression and the limited intake stroke everything becomes critical.

I keep watching this thread in the hope that someone comes up with a magic bullet and determines something which makes this engine run. I am leaving for a few days so I am counting on the folks here to have an answer by the time I get back.

Gordon

 

[Steamchick]

Enjoy the break Gordon,
We are restricted to "local and essential travel only" - so no holiday breaks for us in the UK. So MAYBE we'll have an answer... but I doubt it very much. On paper one cannot absolutely decide things like ignition timing, carburation, etc. as those traditionally have always been fine-tuned on the test-bench/dynamometer. The study is called Engine Mapping, and takes in various loads, speeds, atmospheric pressures, fuel qualities, etc. - but the sum of knowledge here could do the same, if we had running engines, which we do not...
All I am striving to achieve with my study is a simple explanation of the combustion cycle of the engine, so the various modellers can appreciate better how the engine runs. (Or not). The usual attributes - like minimising friction and leaks, will of course always apply and are often critical in modelling. (The main reason I have some non-runners!). The subtleties of carburation, both for starting and running (seldom the same) are really a study of a "supposedly running" engine, with a lot of chance thrown-in.
An aside: I once spent 4 hours working on a friend's CZ motorcycle that he had acquired as a non-runner and had re-built the engine. Ignition timing and carburation were set "as the book", but the engine would not run despite all checks - including compression - seeming to be OK. We even pre-heated spark plugs, tried a few different grades of plug and changed the HT connections, - all showed good sparks but no ignition of the fuel. Finally we got it to start with a completely wrong spark plug set at a small gap, but it would not run for more than about a minute. He bought a new spark plug, and it started straight away, and ran for a couple of years before he sold it. It seems that it didn't like "Wrong grade" spark plugs as they had the wrong position of the spark gap in the cylinder, and the original spark plug - although it worked in my motorcycle (we even tried that!) that he bought new when he did the re-build would not work in his engine.... and we didn't know why.
So getting something like an Atkinson Differential to run is a REAL hard-task.
Actually, I suspect it will run with some compression, as long as the fuel air mixture is correct and ignition timing close to "right". But the fuel air mixture is probably the hardest to achieve...
An engine that pops or runs for more than a few sputters is very nearly there... so WELL DONE. so far...
K

 

[Dave Perreault]

Hi, this is Dave Perreault..... Thank you for all the interesting conversation about my Mini A Atkinson Differential engine.... I thought I would put a like to the engine running video.
Thanks Dave

 

[Ken Brunskill]

Hi, this is Dave Perreault..... Thank you for all the interesting conversation about my Mini A Atkinson Differential engine.... I thought I would put a like to the engine running video.
Thanks Dave

Dave P.,
Thanks for posting the video, shows that when everything is right, your design does produce a runner. I am finally at the stage, with all the ignition components in place, tomorrow I'll commence starting attempts, at getting mine to run.

I have a spreadsheet where I have used Solidworks to take Piston-Piston & Pump Piston to Port measurments for every 3 Deg. of Flywheel rotation, unsure if every one here would be interested, will post if if there is enough interest. Alternatively, I'll post just the valve timing.

 

[Ken Brunskill]

OK Folks, as mentioned earlier today, this is the valve event timing for Dave P.'s Mini A. Keep in mind these are nominal (perfect) dimensions and fabrication tolerances must be considered. Leaving this, strictly a start point.

 

[Gordon]

Just to add another one of my wild theories which I came up with when I was working on the engine a year or so ago. Since ring sealing is dependent on gases getting behind the ring to force it against the cylinder wall the rings are not being forced against the cylinder wall when the LH piston is moving in the compression stroke. This theory is probably as valid as my theory about phase of the moon.

Gordon