Now, let me preface this by saying that I have about 4 beers and shot of tequila in me. Pretty average friday actually, but all my freinds are out of town so here I am

But seriously. I was talking about how the number of valves affects curtain area which is a huge part of why most makers have goine to DOHC for hi-po engines. I suddenly realized how the assumption that all valves (regardless of layout) are circular is a huge assumption and possibly a dumb one. Seriously, the geometry is critacle in the amount of curtain area you get. If you can use shapes other than circles in a given bore, you can get quite a lot more flow given enough RPM.

Think about it, and look at the pics of a mean ported head. You can't tell me that the airflow is evenly distributed around each bit of the circumfrence of the circular valve as it opens and particularly at its highest lift point. Just look at a pic of an intake port where you can see the valve. The air is obviously going to rush in in one particular reigon of the valve edge instead of another.

So here is my question: Why not make a wedge or fan shaped valve that maximizes the circumfrence in the area where the airflow will naturaly want to go?

I can think of two reasons off the top of my head. One, by maximizing the flow in that one bit of curtain area you would get really uneven cylinder filling compared to a circular valve. Two, the uneven weight and inertia would eat guides and seals very quickly.

Ideas? Ridicule? Am I on to something or has it already been thought of, researched, discarded, and laughed away?

 

Also, once the shape of the valve becomes anything but circular, you then have to design a valvetrain that does not allow the valves to rotate in the guides. This would be costly and unreliable as it adds more moving parts to the system. Circular valves are free to rotate any way they want and they will still seat. Also, circular valves and seats are simple and cheap to manufacture.

 

Kind of like manhole covers in the street. Round is the only shape that will never fit through the corresponding hole it is supposed to seal.

 

Round valves offer the most curtain area (circumference) and most flow area for a given circumference. That means the most area for a given weight. That's a biggie! Any non-round poppet valve makes for a very difficult sealing job, as well as big problems manufacturing the seat.

As far as the airflow being biased toward one side or the other of the valve due to chamber shape and placement in the bore, the airflow may vary over the engine operating range, being the fickle stuff that it is.

 

I think it has more to do with the stress and heat build up in the valves than manufacturing. **edit** Manufacturing meaning actually making the seat- design wise its going to be a PITA to get it to seal. **When you have corners, things heat up, break down, etc quicker. You would be adding considerable stress to that valve by restricting its motion in 5 degrees of motion instead of 4.

Modeling flow around any shape valve would not be an issue with CFD programs of today; no one does this stuff by hand anymore. Hand calcs would be wildly inaccurate relative to CFD and time prohibitive. Turbulentflow is not that big of a deal to simulate either. Considering the lengths OEMs go to for higher output and efficiency, i doubt manufacuturing the valve seat would be a problem.

 

That's where you hit the nail on the head. Pretty easy to machine(manufacture) a perfect circle, and even more, two perfect circles to line up and seal perfectly. Just a rotating cutting tool, and you get a perfect circle. If you tried to make an odd shape, you would have to have the valve, and the seat the same exact odd shape which would be basically impossible from a manufacturing perspective.

 

Something to consider here...engineering analysis software is only as good as the software design teams. Among other programs we use at my job, we use autocad mechanical, which includes some nice 2d fea tools that we use for analysis. Funny thing I noticed one day, that when I entered the material thickness, I typed in a fraction. Didn't get any error. But it turns out that the software isn't made to use fractional input, so it actually just disregarded the denominator, and used the numerator as the part thickness!

Moral of the story: analysis software isn't magic, you still need to know how to do the analysis manually to check quality control on the output from your software.

 

It allows the valves to rotate somewhat and still seal. Any other shape would have to return to "exactly" the same place on the seat to seal every time. With out some sort of indexing, it wouldn't happen. Years ago, as I remember, Honda tried oval valves on a small multi-cylinder GP bike. It was mostly so they could fit a 6 valves in a very small combustion area. These valves were tiny, without a lot of mass and they spun the engine to 18K RPM. It didn't work very well and they abandoned it.

 

I remember also honda years ago experimenting with oval cylinders and pistons which they also abandoned. Probably close to 20 years ago.

 

They used like 20 piece rings on those things.

 

there are a couple reasons why they are round
1: cheap to machine
2.provide the best heat transefer(if you have a corner it has to be thicker otherwise it will get hotter. it will be like having to little of a margin, causing the valve to melt and warp)
3. you want your valves to be able to rotate to remove carbon deposits.
4. to the best of my knowledge valves have always been round and manufactures dont like changing thier process.

 

instead of changing the shape of the valve why not knife edge the bottom of the stem? we all know hoow undercutting helps but what u if you knife edged it like an airplane wing?

 

With the great deal of spring pressure causing friction on the valve stem from the retainers how much do you guys figure the valves rotate?

Has anyone ever marked it and checked it a couple 1000 miles later?

Would be interesting to see the results.

 

Be very very careful with the word perfect when it comes to machining anything. There's lots of deflection in cutting tools (even the most rigid machine tools with the most rigid cutters) as well as in the parts while they're being cut. How accurate do you want the measuring tools to be?...the more accurate they get, the more you'll find it's almost impossible to cut a perfect circle. You can get very close with most conventional means though (and MUCH closer with grinding)...for sealing valves it's good enough, but I wouldn't want to even think about trying to seal non-circular poppet valves...especially when you take into consideration the tolerance on degrees between whatever on the valve stem is allow you to keep it from rotating and the face of the valve, how well the face of the valve is cut, how well the valve seat is cut, how well the seat pocket in the head is cut, how accurately the pocket is degreed relative to the valve guide (to keep the valve degreed properly to the seat)...WAY too many variables.

 

oh, I know. Believe me, I know. I use the stuff regularly

your problem is autocad though; might as well circle it in blue

 

Even better question- Why are sewer cover round?

There are actually 2 right answers.

 

1. Most of the people who climb down into the sewer (or understreet areas) are basically round, so they fit better thru a round hole.

2. Round sewer covers are easier to move; you can roll them down the street. Polygon covers thump a lot when they are rolled.

3. Like with valves, you get the most area for the least amount of materal used.

4. You can't drop the round cover thru the uncovered hole, so the folks working below can wear regular hard hats, not the 50 pounders it would take to deflect a dropped cover. Fewer workers on disability that way.

 

I've heard some people complain about autocad, seems to me similar to how mac users complain about windows.

I have been using many forms of autocad for years and years, and swear by it for the most part. I also am good with solidworks and inventor too, so it's not the only software I know. By the way, I also prefer inventor over solidworks too. Even for 3d work, which I've done a ton of in autocad, I actually like it better than solidworks in many ways, mostly because in solidworks it's so easy to paint yourself into a corner. Autocad gives you freedom.

 

Because due to the rotation of the valve, it would cause the knife edging to become a hinderance to flow rather than a help. Not to mention stress risers and the like.

 

Actual Situation: In 1979, the Honda NR500 had an 8-valve per cylinder 2-piston design that used conventional valves. Only the piston was oval, not the valves. It actually worked very well but was too heavy and the honing process (patented) was extremely expensive with the technology available at the time.

Simple fact is, more conventional designs caught up and surpassed it in power-to-weight ratio, so they abandoned it.