Rotary Valve Engine

[MJM460]

Hi Paul, another interesting concept on which to develop your skills.  I don’t comment much but do read every post, however I do have some comments on the first post in this thread.

First, if I understand correctly, the valve is intended to turn through 360 deg, driven by the coupling rod.  You are quite right to look closely at how the valve will be driven past the 3 and 9  o’clock positions where the driving torque on the valve is zero.  The flywheel stores the energy required to turn the main crank past these positions, but the flywheel is on the main shaft, not the valve shaft, and will not apply any torque to the valve shaft at those points.

I suspect the result will be a lock up, or perhaps oscillating motion, depending on the tolerances achieved on all the components.

Remember that locomotives generally have at least two double acting cylinders with the cranks at 90 deg, so there is always significant torque.  Secondly, the coupled wheels are connected not only by the rods, but by the rails, so the linear motion of the locomotive along the rails ensures that the coupled axles are forced to turn even when there is no torque from the coupling rod. 

One solution to the problem is the well known rocking valve, where the valve has a longer crank arm than the main crank so it turns less than 180 degrees, so oscillates.

Another possibility is a belt or gear driven rotary valve.  And an idea out of far left field, a second flywheel on the valve axis might do the job.  Other forum members may have other ideas for you to consider.  We all want to see your engine successful.

MJM460

[Jasonb]

I was under the impression it is a rocking valve engine, see the link in the first post.

[PaulR]

MJM460: Thanks for the comments. I had thought about a secondary flywheel to avoid it stalling; I was going on a gut feeling but you've explained the reasoning clearly 

Jason: Yes the original was a rocking valve but I thought it might be possible to use a crank or disc to do a complete 360 in unison with the big end to achieve the same effect (with flats/holes on the valve set accordingly).

[Jasonb]

In that case then you are likely to get unpredictable valve movement either 180 or 360deg

[PaulR]

In that case then you are likely to get unpredictable valve movement either 180 or 360deg

OK, I'll stick with the lever valve then

[Charles Lamont]

If you want a fully rotary valve (and why not, as Barry Norman would say) then you could use a timing belt or chain drive.

[PaulR]

If you want a fully rotary valve (and why not, as Barry Norman would say) then you could use a timing belt or chain drive.

Sounds more difficult, I definitely need simple at the minute! 

[PaulR]

Right, after much frustration with the few online linkage simulators I figured out how to do it in Solvespace. I've settled on these essential dimensions which give around 90 degrees of valve rotation. I think I'll have to spot the position of the valve flats through the cylinder once all the parts are made and the valve pin set in the right position; even though I've got the co-ords of the traced arc from Solvespace and can work out the exact angle, I've got no way to mark that out on the valve. These dimensions give bags of room for a simple piston/wrist pin connection, and a short cylinder making manufacture easier (I might make it longer so the clevis doesn't pop out the end which gives a slightly vulgar appearance even to my blocky constructions!  ).

[Jasonb]

Looks like it is back to the original design with the eccentric "pin" approx 90deg from the crank pin which I have always said was needed to get the valve lever at it's extremes of movement when the crank is in the middle of it's movement.

Visually it will also look more interesting than what you posted yesterday as the conrod and valve rod will be moving out of phase with each other.

I would keep the wrist pin part of the piston full diameter, if you do want to reduce weight/friction take some out the middle as I showed. But I would make it as a simple cylinder first, the waist can always be added later. Keeping it short you don't get the extra guiding effect and the side loads are likely to increase friction!

[PaulR]

Thanks again Jason 

[Jasonb]

If you did want to keep all the piston inside then you could use longer material for the cylinder and turn it down so it looks a bit like a stubby trunk guide.

[PaulR]

Good idea...that might also be a good place to make an oil hole.

[MJM460]

Hi Paul, that’s much better.  I would suggest you now play with the valve rod length and the angle of the valve lever to get the lever at about 90 deg to the engine centreline at mid travel, and travel about equal distance each side of this line.   To get your valve timing right, you also need the crank driving the valve rod to be about 90 degrees to the crank connected to the piston.  I think this is the same as what Jason is saying.
This will give more symmetrical valve events, and avoid the extreme angle at one end of the stroke you show in the new drawing where the valve con rod is getting close to a straight line with the valve rocker arm.  This will reduce the maximum force in the rod, and so reduce the friction losses in the valve mechanism.

Looking forward to seeing firsts chips now you have refined your design.

MJM460

[PaulR]

Hi Paul, that’s much better.  I would suggest you now play with the valve rod length and the angle of the valve lever to get the lever at about 90 deg to the engine centreline at mid travel, and travel about equal distance each side of this line.   To get your valve timing right, you also need the crank driving the valve rod to be about 90 degrees to the crank connected to the piston.  I think this is the same as what Jason is saying.
This will give more symmetrical valve events, and avoid the extreme angle at one end of the stroke you show in the new drawing where the valve con rod is getting close to a straight line with the valve rocker arm.  This will reduce the maximum force in the rod, and so reduce the friction losses in the valve mechanism.

Looking forward to seeing firsts chips now you have refined your design.

MJM460

Thank you, this looks like it fits the bill:

[Jasonb]

No I was meaning that yesterday's sketches had both the crank pin and valve rod pin in the same place. Todays has the crank leading the valve by 90deg.

It is a lot better, I have altered my parts to your latest dimensions and this is how it moves.

Myself I would take the crank throw back up to the 8mm and also possibly reduce the valve from 9mm down to 8mm as it is possibly swinging a bit far but that will depend on the valve and ports. I think having the inlet dead centre of the head as shown in your sketch may give problems.

<a href="https://www.youtube.com/watch?v=YxwwqmX192g" target="_blank">http://www.youtube.com/watch?v=YxwwqmX192g</a>