Valve Cage Configurations

[petertha]

I've been trying different valve cage configurations on my radial & going to have to make a decision & live with it soon. The head has a hemispherical shaped combustion chamber. The valves enter at a specified angle, so that's all locked down in the design.

The assembly method calls for machining the cage with a perpendicular edge, so its basically like a squared off sharp edged tube. The cage is permanently inserted into the head, then the valve seat cut with a 45-deg tool, then lap as necessary for final seal. This allows for any slight distortion of the cage to occur before cutting & dressing the seat. The cage is now permanent so if seat is buggered up or seal can't be achieved , then frowny face. This method has been documented by better builders than me in other posts, but my question is more about the assembly final layout.

With the hemi head & valve angle / placement geometry its hard to get the valve totally flush. I've tried to strike the happy medium so one valve edge extends slightly into the CC dome. The opposite side is inside a bit. What I'm concerned about is the resultant open space nooks & cranny's, specifically where combustion debris can potentially build up. Hopefully the sketches illustrate this.

I could make the cage+valve stick entirely into the chamber. I've seen that done on another engine. That would keep crud buildup to the outer area of the cage, ie harmless. It would increase the CR a bit but mostly I'd have to check for piston interference. Another radial that used that method have relief arcs cut into the top of the piston.

Looking at real heads, they come in all flavors but generally they have a smooth profile kind of transitioning to the seat area. I could Dremel some of this profile, but not near the outer edge where there is a lip in the head.
Any comments or opinions?

 

[petertha]

Extended valve cage length (exaggerated)

 

[petertha]

Somewhat similar 7-cyl radial with valve cages close to flush with combustion chamber, therefore protruding valve faces into combustion chamber, therefore requisite valve clearance crescents in piston tops.
http://philsradial.blogspot.com/2012/01/cylinder-heads-5.html

 

[petertha]

Another option is to make the valve cages thinner wall thickness. This would reduce the trapped gap volume. But the wall thickness is only 1.0 mm right now (.039"). Comparing to other designs I'm thinking thinner would be more prone to distortion.

 

[Charles Lamont]

Pockets in pistons are quite common, specially with 4-valve heads. (Currently ready biography of Keith Duckworth).

 

[mayhugh1]

I've put valve cages in both conical and hemispherical combustion chambers. You can't get around the math of their intersecting surfaces. The results are never as esthetically pleasing as when they're mounted in a flat head especially when there isn't enough room to deeply sink the cages and clear the valves. I always account for the valve height protrusion when calculating compression ratio, and its effect can be dramatic - sometimes as much as one and one-half points. - Terry

 

[gbritnell]

I have never been a fan of valve cages for a few reasons. It seems like if you press them in you get distortion and I don't like using Loctite on everything. Generally when cutting a seat in a conical shape the area around the valve opening is spotfaced to provide a flat surface for the valve seat. My radial airplane engine has aluminum heads with hemispherical combustion chambers and pressed in seats (steel) I cut a spotface slightly larger than the seat diameter then machined for the seat. When pressed in it sits flat against the counterbore.
gbritnell

 

[WOB]

I don't get the "locked down" part. Why not change the valve angle a couple of degrees? Or if that's not possible, change the shape of the combustion chamber. Why does it have to be a hemisphere? Why not conical? Or a combination. Or throw out the original head design entirely and do your own from scratch. That's what I did on my Hodgson 9 engine. I thought Hodgson's head design was poor at best. It might have been closer to a real full size radial design-wise, but that was not what I was building. I went with conical combustion chambers, phosphor-bronze valve cages and centrally located plugs in easily removable heads ( 5 bolts). Runs great.

WOB

 

[petertha]

Hi WOB. I admire you modifying the Hodgson. A head re-design could be tackled I suppose. I've made some necessary minor tweaks here & there along the way, but this is my first engine build & I don't really want to stray too far off the path because it turns into a bit of cascade effect. Different valve angle / placement translates into different rocker perch facet, rocker arm mounting, push rod action, probably even valve timing if it moved significantly. The CR is kind of critical being a glow engine so that's another factor modifying the hemi shape to something else. I know the engine runs in its 5-cyl & 9-cyl form.

It's not the end of the world, more of a visual thing unless the gap presents running problems with any extensive carbon buildup. I've ran RC engines hard for a season & they just have a light, thin sooty coating if mixture if all is well. I've attached a view sliced along the valve axis which makes the conformance look better than along the head half where its the worst.

Also some pics where I alternately machined a 30 & 60 angle into the cage which then leaves 45-deg seat to occur on the crown. This layout effectively raises the valve into the cage a bit without thinning the cup too much & kind of a more gradual shape. But it doesn't noticeably improve the gap & is a lot of extra fussy steps to do this properly & consistently across 10 cages. I tried making some hardened D-bit like tools to cut these angles by advancing the tail stock. But thus far have had so-so results compared to single point cutting.

 

[petertha]

It occurred to me that maybe there is an optimal relationship between the valve diameter and cage ID with a defined maximum seat length. Here I reduced the valve OD slightly. It helps a bit I suppose. But now the contact point occurs out further outboard towards the valve lip where it is thinner vs. more in the middle of the 45-deg seat area as before. I have already thinned the vertical section a bit after seeing some comparable sized RC 4-S valves.

 

[petertha]

After some more fiddling around, I've come to the conclusion that reducing the cage ID slightly is the single most beneficial factor. Of course that comes at some expense - the cage has slightly less wall thickness so is more fragile in terms of potential distortion. So the hope is that permanent installation into the head's counter bore with say .001-.002" annular clearance will limit the egging & once filled with HT Loktite will give outside cage support as though it were a thicker solid. And then the 45-deg seat cut from there.

The stock design calls for 10mm OD (0.394") and 8mm ID (0.315") for a 1 mm wall thickness (0.039"). I'm not sure if 8mm was a nice round metric number corresponding to a reamer, but that's how I've interpreted it & been making testers thus far. But there is nothing stopping me from single point boring the ID or buying an IMP decimal reamer that come in very small progressive increments. After a roughing drill, single point boring might actually be a better way to go in the bronze anyways. So, enough sketching. I'll give this a shot in my test block.

 

[petertha]

Here are some machining pics corresponding to the drawings. I think I'm happy with the end result of how the valve sits in the head all things considered. The valve seat can be nicely controlled to a smallish band 0.010-0.015" across using the gun tool.

I found pre-drilling the bronze is best done with a dubbed tip to mitigate grabby-ness that bronze likes to do with regular sharp drills. Boring to final ID was easily done with a carbide micro bar. One thing I had not considered is that this head has a lip that extends into the liner so in order for me to chamfer the 45-deg seat with cage bonded in, it will also have to simultaneously relieve a bit of one lip side as well. Thus far I have been chamfering the cage when its been outside the head just for the evaluation. Next test is the whole procedure from beginning to end on my mockup head. I also realized I should be boring the cage ID and then without changing setup in lathe collet, center, drill & ream the valve stem hole so the seat & stem stand the best chance of being concentric.

 

[tornitore45]

[QUOTEI also realized I should be boring the cage ID and then without changing setup in lathe collet, center, drill & ream the valve stem hole so the seat & stem stand the best chance of being concentric.][/QUOTE]

That is the way I did and turned out well.
I machined the two OD first. That make sure they are concentric.
Then grab in a collet by the small end which is longer and presumably hold better on axis. Absolute centering is not necessary, all it does it misplace the the valve location a thousand or two.
Then do all the ID ream the stem hole and single point bore the cavity
I do not like to use chamfering cutters for the seat.
When I finish boring the inlet cavity, the boring bar is right on the radius.
I back off 0.010" that will stop the bar from entering the cavity.
At this point move the carriage to touch the cage front and lock the carriage
Go in with the compound previously set at 45*
Polish up with scotch pad before unchucking.