Edwards Radial 5 build thread --- PHOTOS!

[JRaut]

Hey guys, I've been on quite the hiatus on building my radial. It's currently sitting over in the corner of my seldom-used shop.

I'm working on finishing my basement (and rearing a 6- and 3-year-old son and daughter), so shop time is hard to come by in this busy season of life.

I'll get back to the engine one of these days and intend to continue to provide updates when they eventually happen.

 

[woodchucker]

yep, more important you spend time with the kids, you can't get back time.. but you can always pick up the project .. Kid time will pay you back over and over again.

 

[Dhal22]

My last kid (sophomore) decided not to tryout for high school basketball this fall, even though he plays year round. He just needs a break I guess. With my older son now a junior in college I for the first time in years and years have no kid sports. Sad but definitely knew this was coming. Hobby time about to go way up.

 

[starr256]

Is it possible to do all of the work on just the lathe?

With a milling attachment and some creative piece holding. maybe. More effort than I would be up to. That said, it has taken me over 2 years to get maybe 40% done with both a mill and a lathe.

 

[griggsfamily5]

With a milling attachment and some creative piece holding. maybe. More effort than I would be up to. That said, it has taken me over 2 years to get maybe 40% done with both a mill and a lathe.

Good to know. My biggest holdup is finding time to work in the shop at all. Space is also at a premium (same as a lot of hobby shops, I guess). Currently I am ether touching up edges with hand stones or using an angle grinder to shape them. I don't really have space for a mill that would be big enough.

 

[rgaal]

Fantastic job, gives a lot of ideas how to approach this engine build.
I have thought about the machining of the cams and I think I have found the solution. The idea is not to machine at the height prescribed by the tables but at the level of the tangent plane at a given position. This plane, except at the circular part of the cams and top dead center of the lobes is not above the center, so one has to sweep the cam under the mill. In the annexed file I give some more detailed explanation and the new coordinates I have calculated.

 

[Dhal22]

Fantastic job, gives a lot of ideas how to approach this engine build.
I have thought about the machining of the cams and I think I have found the solution. The idea is not to machine at the height prescribed by the tables but at the level of the tangent plane at a given position. This plane, except at the circular part of the cams and top dead center of the lobes is not above the center, so one has to sweep the cam under the mill. In the annexed file I give some more detailed explanation and the new coordinates I have calculated.

So far over my head I don’t even know how far….

 

[rgaal]

The basic idea is that one should machine tangent to the cam profile. In a standard setup with a rotary table by setting the cutter above the center of rotation this is only the case at the circular parts of the cams and dead center. At the raising and falling parts of the cam this is not the case, which means you invariably cut into the cam on one side, and the flat you machine will be at an angle to the true profile you want to obtain. The tangent point in those regions is somewhat higher and out to the side.
Just imagine a bump on a circle, touching a horizontal plane and look at it from the side: when you rotate it and you are in the circular part, the bump far below, the plane will touch right above the center of the circle. As the bump moves upward, a moment will come when it will hit the plane somewhat sideways. At this moment the bump will start to lift the plane, and the contact point will remain to that side until it reaches top dead center, then said contact point will move to the other side and linger there until the circular portion will come to bear again.
For calculating the distance of these tangent lines is a bit involved but I am trained for that one.
If you use the elevations given in my table it is important to make a horizontal sweep so that you mill in the region of the tangent, as suggested by the very simple drawing.
One remark, the units on the graphs are mm, as I am in a metric country, but the elevations in the tables are in thousandths of an inch.

 

[rwm]

I did not understand your explanation at first. After reviewing your drawings, I think you have a good idea here. That is a complex part.

 

[petertha]

@rgaal there probably is some merit to your method, but its probably also worth pointing out some of the some other issues in model engine radial cams.

- the cam plates are actually quite simple in principle: the inner diameter with no action (valve closed), the outer diameter 'bump' dictating (valve open) throw through the rocker assembly geometry. And the ramp curve connecting these 2 features which ideally has some smooth transition curve. The transition profile is more critical on engines with smaller cylindrical cam followers that see the cam as a point tangent. The Edwards is somewhat unique in that it has a profiled follower 'shoe'. But you will notice they are in line vertically coincident to the cylinder center line vs radially outward which is maybe a bit more common on other engines.

- I used to think cam timing was super critical on these radials. But I'm now of the opinion there is actually pretty wide margin where they run well for the task s long as the timing is reasonable & proven. Part of this stems from these radials are generally lower RPM, broader torque curve outputs, not high RPM 'peaky' horsepower type engines where cam tuning can moreso influence breathing. Back in post #140 I showed a comparison of OS (methanol glow) 4-stroke timing. You can see there is actually quite a wide deviation even normalized to the same manufacturer. (I have since found data for Saito so will add that in at some point). The Edwards has a slightly different exhaust vs intake profile but many model engines are the exact same cam plate for simplicity, just positioned different relative to TDC.

- valve lash (gap between valve stem & rocker) has quite a significant effect on actual valve open/close timing at model scale. Anywhere from 0.003 - 0.010" cold gap depending on the engine. Lash reduces open duration timing by a surprising amount on model engines, so we can't really look at the theoretical cam open/close duration in isolation. This is magnified on models because because typical lash is a higher percentage of cam throw vs full size engine.

- even with identical cam timing, different methanol (glow plug) engines will fire 'whenever they want to' within limits close to TDC. Its not an ignition driven event like a spark ignition, its a function of compression ratio, fuel etc. So an 8:1 will fire slightly different than a 10:1, which is why there is a requirement to equalize CR among the cylinders heads because the Edwards has an uncompensated master rod which would result in significantly different CR's that basis alone.

So not taking anything way from your observations or cam making proposal. I'm just saying (in my opinion) fortunately its probably not as critical as one might think. Now of course you cant have a jerky motion or a radical cam without consequences, but it seems like there is quite a generous margin within reason among similar model radial engines.