Atkinson Differential Engine - Making it work?

[propclock]

Thanks Dave , very informative.

 

[dsage]

Making of the clear cylinder turned out to be a good idea as it demonstrated visually that the drawings were accurate and also gave me some more thoughts on changes that might be beneficial. But it was not perfectly accurate for making measurements.
One of the stumbling blocks was, other than trusting the drawings, I couldn't figure out how to measure where to put the ports in the cylinder. The light went on when I noticed that everything could be measured from the tail end of the pistons and knowing the dimensions of the pistons I could measure exactly where the front of the piston was going to be (which is where the ports are located). The only thing that could throw that off is the height of the cylinder above the base plate because altering that dimension pushes the pistons in or out of the cylinder. I trusted the drawings for that measurement.
In any case it took quite a bit of checking to be sure I had everything setup and measured before possibly trashing a good cylinder.
The first picture shows milling the hole in the water jacket that the cylinder slips through. My machine is very sloppy especially when milling steel. I also had to use a very long 1/4" end mill in order to reach through the water jacket to mill the other side to ensure they were concentric. In the end even though it worked the hole turned out to be too small and out of round. Not a disastrous outcome because I then used a boring head to bring it perfectly to size (and round). (Time for a new machine).
After marking the cylinder with the port locations I used a 3/16" end mill to make the ports. You'll notice I've decided to make the exhaust port a slot as well and that decision came from observing the motions which indicated that (if the combustion was poor) there might be a point where the exhaust is being compressed again at the right side of the cylinder. So I thought maybe I'd "open the exhaust port a bit early. This whole port placement is sort of like a two stroke engine which I have no experience with. It's all a guess.
In any case the next series of pictures shows that my measurements for the placement of the spark plug hole landed it right between the pistons at "TDC" where I had hoped. The next picture is of the exhaust port slot placement where the left piston and right piston come very close together (about 1/16 apart). The right piston is almost gone to the right of the slot while the left pistons rings are visible. There will be no problem getting all the exhaust out.
The next picture shows the start of the intake motion. There is a bit of motion where there will probably be a bit of vacuum build up because the exhaust valve will be closed and the intake port is not uncovered yet. But I think a bit of vacuum build up will be ok when the intake port is finally uncovered.
The next pictures shows how (I hope) the slotted intake port will allow for a longer intake time and it's length is designed to take advantage of all of the motion that "appears" to be when the pistons are still creating vacuum before the port is covered up.
My next question is going to be what to use for valves. There is no need for conventional poppet valves because in this engine neither of the valves has to seal against compression and only residual combustion pressure. Those pressures all takes place between the pistons when they are both at the left end of the cylinder far away from the two ports.
I could use poppet valves but the springs would likely need to be very very light.
So I started thinking maybe a free floating ball and seat might be easier than messing about with spring pressures as the ball valves would be self governing.
Then I thought of reed valves. They might be most effective and are certainly used in two strokes but I have no experience with them. The good part is that the valves will be enclosed in pieces that are fastened to the cylinder over the slots so they can be experimented with. Also if I find the slots need to be modified I can just cover them up to a certain extent by the pieces that hold the valves.
Ideas are welcome.
PS> You might notice that the cylinder is at this point held in place by JBWeld spots. Since the cylinder is so thin I didn't want to risk silver soldering it in place and distorting it. I'm going to completely JBweld it in place eventually and form a nice fillet at the joint and eventually paint the water jacket (and maybe the cylinder assembly).

 

[dsage]

I guess my thoughts on using reed valves is going to have to be implemented. I hadn't noticed that raising the cylinder to increase compression ended up with the intake valve making contact with the oscillating arm. Also sliding the exhaust manifold over to line up with the new slotted port put it in contact with the intake manifold.
Below are a couple of pictures of the collisions and then a few pictures of new manifolds using reed valves. The reed valves are in orange inside the manifolds.
There is still a bit of interference with the carburetor requiring it to be cocked on an angle to get the fuel hose on it but that might be solved with a 90deg hose pipe.
I pondered over what to use for the actual reed material. It needs to be very flat and springy so it will cover the port, seal properly and spring out of the way when pressure acts on it and reseal reliably. Then I thought of a small piece of feeler gauge. I made the dimensions of the manifold interior such that I could just cut the end off of a standard 1/2" wide feeler gauge and clamp it in place. This will also give me the ability to choose a thickness from a set that suits the application.
The problem right now is that I don't want to cut up my current set and considering that we are supposed to be isolated at home it's probably a bit irresponsible to go out to a store to buy a new set. If only they stocked feeler gauges at the grocery store
The idea might have to remain as drawings for a while. Sometimes not a bad idea because I might think of a better idea. I could probably make one of the manifolds (they are fairly simple) and substitute the feeler with piece of mylar sheet to test the idea.

COLLISION OF THE INTAKE VALVE

INTERFERENCE OF THE INTAKE AND EXHAUSTS MANIFOLDS

CLOSE UP OF INTAKE MANIFOLD WITH REED VALVE (flat orange)

BOTH MANIFOLDS WITH REED VALVES (carb on an angle to accommodate the fuel hose connection)

 

[awake]

Here you go - delivery to your front door:

https://www.amazon.com/OEMTOOLS-250...id=1585250822&sprefix=feeler+g,aps,198&sr=8-8
Of course, there are plenty more options available from Amazon if you want something different, but the one shown is cheap (you are planning to cut these up, right?) and better yet is an Amazon Prime item. Of course, even with that, how quickly it will arrive in these perilous times is a question.

And of course, plenty of other on-line stores you could order from, starting with HF - but then you have to pay the shipping.

https://www.harborfreight.com/feeler-gauge-32-pc-63665.html

 

[propclock]

Following . Enjoying your process. And progress.

 

[dsage]

Awake:
Yes I guess that's an option. I never thought of it.
Thanks

 

[awake]

A totally selfish gesture on my part - I want to see where you take this design, and how it turns out!

 

[Gordon]

I guess my thoughts on using reed valves is going to have to be implemented. I hadn't noticed that raising the cylinder to increase compression ended up with the intake valve making contact with the oscillating arm. Also sliding the exhaust manifold over to line up with the new slotted port put it in contact with the intake manifold.
Below are a couple of pictures of the collisions and then a few pictures of new manifolds using reed valves. The reed valves are in orange inside the manifolds.
There is still a bit of interference with the carburetor requiring it to be cocked on an angle to get the fuel hose on it but that might be solved with a 90deg hose pipe.
I pondered over what to use for the actual reed material. It needs to be very flat and springy so it will cover the port, seal properly and spring out of the way when pressure acts on it and reseal reliably. Then I thought of a small piece of feeler gauge. I made the dimensions of the manifold interior such that I could just cut the end off of a standard 1/2" wide feeler gauge and clamp it in place. This will also give me the ability to choose a thickness from a set that suits the application.
The problem right now is that I don't want to cut up my current set and considering that we are supposed to be isolated at home it's probably a bit irresponsible to go out to a store to buy a new set. If only they stocked feeler gauges at the grocery store
The idea might have to remain as drawings for a while. Sometimes not a bad idea because I might think of a better idea. I could probably make one of the manifolds (they are fairly simple) and substitute the feeler with piece of mylar sheet to test the idea.

COLLISION OF THE INTAKE VALVE
View attachment 114725


INTERFERENCE OF THE INTAKE AND EXHAUSTS MANIFOLDS
View attachment 114726


CLOSE UP OF INTAKE MANIFOLD WITH REED VALVE (flat orange)
View attachment 114727


BOTH MANIFOLDS WITH REED VALVES (carb on an angle to accommodate the fuel hose connection)
View attachment 114728

I am amazed at how many errors there are in a design which has been around for such a long time. I encountered the same problems. I took some off from the oscillating arm and shortened the valve a little. I mounted the exhaust valve horizontally.

 

[dsage]

Gordon:

Concerning your comment that you noticed the same issues, I went back and checked my backup drawings. When I drew the engine as per the book drawings there was no issue with the the intake valve colliding with the oscillating arm. It was close but not a problem.
It was only after making changes such as moving the cylinder up that it showed up.
So I'm at a loss why you saw that problem.
I haven't built the reed valve manifolds yet. I'm still deciding if I'm going to implement them.
I should have noted before that I don't have any piston rings installed yet and I do notice some compression on the compression stroke.
I haven't made the rings yet and will probably leave them to almost the end when I'm done playing around with the cylinder and I'm sure I won't be needing to remove the pistons.

 

[rsholl]

Hi Dave Sage,

I've been trying to follow your build, definitely very interesting stuff but I'd expect no less from you.
As for the reed valves, my first thought was to the small reed valve used in the Cox model airplane engine. I see they have change some but about mid way down on the link is the new reed valves. On the old engine, they just used the small cir-clip to hold them in, they actually floated a small amount. Anyway, I figured if they worked in the little Cox .049 engine at 21,000 rpm they might work for your application. BTW, the Cox dealer is in Canada.

https://coxengines.ca/back-end/

 

[dsage]

Thanks Roy. I'll check them out.
I should probably post a cut-away (or something) of what I came up with so far. Maybe I'll get some constructive criticism on what I have.
Maybe tomorrow.
Thanks

 

[dsage]

Below is what I have designed for the intake manifold with a reed valve.
The charge from the carburetor comes in through the hole in the back but it can't proceed until the vacuum pulls the orange reed valve (down). The reed is the cutoff end of a feeler gauge.
There is quite a bit of space around the sides of the reed so the charge can travel around it. There is also a bit of space at the curved end (not really visible).
There is a bit of a cavity below the reed to accumulate the charge. But the cavity is mostly there to make space for the clamp that holds down the one end of the reed.
The trick will be getting the thickness of the reed chosen properly. But there will be a range of thickness available in a full set of gauges.
The reed could be any material instead of the metal feeler gauge. Maybe a thin piece of mylar would be better?
Looking at rsholl's suggested page from the Cox website. Plastic seems to be the material of choice.
I was trying to avoid burrs on the edge of the reed caused by cutting it that would keep it from sealing. (Just being lazy).
With the feeler gauge there would only be one cut and it would be under the clamp. A typical feeler gauge is 1/2" wide with a full radius on the end.
Also not really visible is that the clamp has been narrowed to about 1/16" width between the hold-down screws which are #0-80. Hopefully so leave the most flex possible.
The whole thing could be scaled down to reduce the mass of the reed. I just started with the original manifold size.
Comments are welcome.

 

[propclock]

I forgot about the circular reed valves"with ears" on the old cox engines. They worked well.
Just for what it is worth. As a kid I gathered old .049 engines that had poor compression
due to ingesting dirt etc. They would not run on the original "cox" fuel. Fairly high nitro?
But being a " Mad Scientist" I added some of my fathers Ronson Lighter fuel to the mixture. Amazing !! even really really worn out engines would run?? They would
run backwards most of the time so I guess they were " Dieseling" ? I really don't know
but I thought I should pass this along. I got a lot of fun out of "dead" engines.
What is in "cox" fuel or Ronson lighter fuel circa 1966? I have no idea. But it works!
Sorry Dave the reed brought up old memories that I thought might be useful to others.

 

[dsage]

Sorry. It's been a while since the last post.
Spare time is high but motivation has reduced.
I used the slower (and stronger) JB weld to affix the cylinder into the water jacket. So it was time to make the bosses that go over the port holes in the cylinder for exhaust and intake.
One thing that is nice about a CNC machine is that you can produce a large radius on the face of a piece of material without having to set up a rotary table or by some other cumbersome method. Of course you have to make a drawing of the exact conditions for the tool path.
The bosses need a radius on them so they sit nicely on the side of the cylinder.
In the first picture after the machine is set up it makes multiple passes in an arc to the full depth of the boss. The second picture shows the result. The third picture is the finished boss with blind #2-56 tapped holes for the carburetor body to screw to. I used JB weld again to glue the bosses to the cylinder. As you will remember, the hole in the boss is round but the hole in the cylinder is a slot. But because of the thickness of the cylinder wall the air will follow the slot to the hole in the boss.

I also started work on the reed valves since it looked like they might work.
The next picture is of the finished exhaust valve body with the recess where the reed will sit. You can see a couple of #0-80 tapped holes that will hold the clamp that holds the reed along one edge.
The reed covers a hole in the bottom of the recess. I purposely offset the hole thinking that air pressure would have more leverage to push the reed open.
In the second picture you see yet another method of holding material while it's machined. In this example the aluminum plate is contact cemented to the piece of MDF and the finished parts - once cut free of the stock - will be held by tiny finishing nails the heads of which turned out to be just the right size for the 1/16 cutter to cut around without hitting them. These parts were about the limit for my worn out machine. They are about 0.6" long, 1/16" wide and 1/16 high. It's always good to make more than you need in case one gets messed up.
The last picture shows the finished exhaust reed valve. I used a piece of clear mylar for the reed and marked an X on it in case it was not visible in the picture. You can see the tiny clamp on the left used to hold it down.
I hold no hope that this reed will survive hot exhaust. It's more of a test to be able to manually crank the engine over to see if the two reed valves operate. They do seem to work quite well when you try to suck air through them.
There is no need for them to withstand high pressures since they never see compression. Their only purpose is to route outgoing exhaust and incoming air.

 

[propclock]

Yea ! Dave.

 

[awake]

Now that's some tiny, fiddly work. If it were me, I'd say you need to make extras because, inevitably, one of them is going to wind up missing ...

 

[dsage]

Now that's some tiny, fiddly work. If it were me, I'd say you need to make extras because, inevitably, one of them is going to wind up missing ...

As far as the tiny clamps are concerned I made one extra. I should have made more. But I have the CAD file so it's just a matter of glueing and nailing down another piece of aluminum, setting the reference point and pressing go.
(At least that's what everyone says is the case for CNC - isn't it ?)
Luckily I can just loosen the clamp and slip in another reed so hopefully the clamp(s) won't go missing.
Thanks for tuning in.

 

[dsage]

All of the moving pieces of the engine are finished so I decided to assemble it - hopefully for the final time. I plan to use a one-way clutch bearing on the end of the crank shaft for starting the engine. They work best if the shaft is hardened. The shaft is made from 1/2" O-1 oil hardening drill rod. I've had a Banggood induction heater around for quite some time so I thought it might be a good time to give it a try on the end of the shaft.
The fist picture is all stuff required to completed the job.
The induction heater that I having running on four car batteries batteries (48 volts)
A magnet held in the vise grips to tell when the drill rod becomes non-magnetic (the metal has reached the desired temperature).
A tin can full of cutting oil to quench with (it smokes but won't catch fire - and I have it around the shop)
The shaft with a wet rag around it to limit the heat to the shaft end.

The end of the shaft is stuck inside the coil. It heats up pretty quickly. It never really got as hot as I would have heated it with a torch but it did become non-magnetic and I let it stay hot for a couple of minutes.

A quick dunk and swish in the oil to cool it off. There was minimal scale on the surface easily removed with steel wool. I tested the shaft with a file and it was very hard. I tried tempering it but it was very hard to see the colour of the shaft being that it's inside the coil. I think it also heats up much quicker with the induction heater since there is no way to reduce the heat. So after annealing it ended up being a bit soft. So I went through the hardening process again and just left it hard. I suppose it might only be surface hardened because I normally heat the metal much hotter and for a longer period of time. But in this application surface hard is good enough.

 

[propclock]

Way cool again. I bought one of those induction heaters and a big switching power supply.
So far it has just been a dust collector. I believe you have motivated me to finish an enclosure for it , Thanks for the motivation/ training.

 

[awake]

Interesting induction heater - didn't realize they were available on sites like Banggood.

For tempering, I generally use the kitchen oven - just make sure the part is very clean and won't emit any fumes.