Valve Seats

[petertha]

My goal has been to make one complete cylinder tower ‘prototype’ to contain my learnings & blunders. If that works out, then carry on with repeated clones for 5-cyl radial. Right now I’m now at the stage of valve & valve cage making. I was particularly interested in Terry’s 18-cyl build because he demonstrated a method for testing valve/cage seal prior to fixing the cages into the head. His build post link here, post #143.
http://www.homemodelenginemachinist.com/showthread.php?t=21601&page=15

As it stands, I have my doubts that I’ve achieved proper seal. The results are kind of confusing, so I’d like to solicit experienced help & improve my procedure. Maybe some of you have walked this path. Terry, if you’ve tuned in, thanks for the PM reply. Hopefully the pics will help clarify.

This is the general assembly head blank with cage & valve. These tester cages are 12L14 steel for now, the real ones will be bronze. The valve is 416 stainless steel.

 

[petertha]

This shows the Brownell 45-deg cutting tool. I was able to make a 0.010 – 0.015” seat in the valve cage quite nicely 8 why I chose 12L14 just to get started. Actually I’m still experimenting here. The cutter is intended to be used with a center guide pin that would be a slide fit in the cages stem segment. I had a cage being made in the lathe & decided on-the-fly to cut a seat with cutter in tail stock under low speed. The recommended method seems to be do this carefully seat by hand.

Here is my attempt at lapping tool to condition the valve face which was 45-deg cut in the lathe. The plan was to use the same Brownell cutter to make the 45-deg lapping face so the geometry would match the seat using same tool. This plan didn’t work out. It developed a chatter & ugly surface. My hunch is maybe because it’s cutting a wider face ~ 0.100”. The plans call for a valve face of 0.067”, so with 0.018” on either side, that’s how I arrived at ~ 0.100”. The 0.010 – 0.015” chamfer on the cage lip which worked like a dream. So… just to keep going I finish cut over the chattered surface with my lathe compound just so happened to be sitting in the same 45-deg position as the valve.

 

[petertha]

Here are the cad pics of valve in cage & lapping tool

 

[petertha]

Here is the MityVac test setup. I have silicone fuel tubing over the stem + valve & short 3” hose to vac body. The idea is, with the valve seated you pull a vacuum from stem side to some reference level, say 25”HG, then observe the leak-off time to some terminal pressure. I won’t repeat Terry’s numbers but he quantified this. A sufficiently long duration means good to go. A rapid leak off means bad seal to fix. A fixed needle means you walk on water.

First I thought beginners luck, it took a nice long while for fall-off. Then I noticed something odd as I repeated the cycles. When it down to say 10”Hg, if I manually pulled the valve off its seat, it did not snap back to atmospheric immediately like I assumed it would. I suspect what’s happening is the longish valve OD stem ID annulus is acting like a choke, so it’s tricking the actual valve seat seal. (This is where Terry did something I’m still not clear on, I want to say he modified a sacrificial valve stem with a flat to give it an unencumbered flow path).

 

[petertha]

Forgot to mention I blanked off the valve altogether (green silicon tube). The needle stood rock steady at vacuum, no drawdown at all. Actually I should have done this at the onset just to make sure everything was working.

 

[petertha]

Then I got fiddling around, trying to determine valve contact visually & lapping compounds. I don’t think the Prussian blue stuff is really a very good enough indicator, at least with my eyes. Even without lapping & slight machining marks it shows a pretty consistent transfer blue ring. A black felt pin tended to highlight a bet better. But I noticed the lapping tool eventually showed a matt ring contrast to the adjacent virgin shiny cut surface & same finish along the valve face.

I figured maybe some lap grit was remaining on surface & allowing leak. After careful cleaning I got the idea to put a single drop of oil on the seat & lightly rotate it once so it was coated. Amazingly the vac needle stuck frozen like the perfect seal test. It took 1min-14 sec from 20-15”Hg, 2min-24 sec to 13” Hg. My valves will be at least this oily with pre-mix fuel. Have I been beating my head against the wall testing dry?

 

[petertha]

I had another idea to drill a separate vac test hole in the body of the cage, blank off the stem altogether. This would confine leak off test to the chamber above the valve seat (thus the seat) & not the stem segment. Hopefully sketch illustrates (picture the valve/gage outside the head like the regular vac test). I haven’t thought this through, but it would be nice if the little test hole would also line up with the eventual (larger) head port drill hole so it just gets bored out after serving its purpose.

 

[petertha]

I also fitted tubing over the valve cup & applied (mouth) pressure on the valve, somewhat mimicking positive combustion chamber pressure. I couldn’t detect any leak or hear any hiss, but this is completely unreliable. But I was wondering if maybe instead of vacuum, could one hook up a small pressured chamber like an airbrush compressor tank, set at say 50 psi & do the reverse test? That would still have the valve stem in the equation but facilitate higher pressure difference to see leak off?

 

[petertha]

Some random pics of my valve stem lapping trials & tribulations. I have yet to give the latest ‘nutcracker handle’ a try but crossing my fingers. The M3 screw at the end seems to give a nice, graduated clamping feel. The brass laps are easy to make & in progressive ID hole, so wearing them out isn’t an issue.

The hand held split lap idea was kind of a fail because the shaft size has to be very closely matched. Stainless isn't the nicest material to cut but I'm getting better with the tool selection.

 

[Cogsy]

After careful cleaning I got the idea to put a single drop of oil on the seat & lightly rotate it once so it was coated. Amazingly the vac needle stuck frozen like the perfect seal test. It took 1min-14 sec from 20-15”Hg, 2min-24 sec to 13” Hg. My valves will be at least this oily with pre-mix fuel. Have I been beating my head against the wall testing dry?

I'm not sure your valves will be that oily, at least not the exhaust valve as the combustion heat should burn the oil off. On full size car engines it is/was common practice to check piston ring wear by doing a compression test dry, then adding a few drops of oil to the cylinder and testing again. The increase in compression gave a fair indication whether compression was being lost via the rings or the valves.

I'm far from an expert but from what I've read, and my own meager experience, if the valves and seats are well made, the first few combustion pops will go a long way towards ensuring a good valve seal and after running a short while the seal should be perfect. Trying to get a reliable seal without the aid of combustion pressures can be very difficult.

 

[mayhugh1]

Peter,
I think you've lost all the advantage of the seat cutter if you don't use it with the pilot to manually cut the seat using very little pressure and a few turns of the cutter with your fingers. I would start with a .005" seat width.
When I tested the leak-down times of my cages before installation I used a test valve with a flat cut along its stem. The flat is required because if you have a good fit between the valve stem and valve cage it will dominate the seal you are trying to measure and since its leak is in series with the leak at the seat, you won't be able to see the one at the seat. The photo below is a copy of the one I used in my original post showing the flat along the valve stem. - Terry

 

[mayhugh1]

Peter,
I polished the back of the valve head (the seating surface on the valve) while the valve was still in the lathe. I was able to get a brilliant finish free of machining marks with a scotchbrite pad. At that point I considered the valves finished and did nothing more to them. The lapping I did was on the valve seat in the cage. The manual seat cutter leaves light scratches on the seat which will affect the leakdown numbers. These are the only imperfections that lapping will help with. You can't correct geometry problems with lapping. You will destroy your lap or your valve if that is what you are using for a lap long before you correct any geometry problems. I made a few extra valves but with spigots on them to use as laps so I could conveniently hold them while doing the lapping on the seats. By the way, my method for lapping is to twist the lap back and forth into the seat only +/-90 degrees several times, lift, turn the lap random fractional turn, back down on seat, twist +/- 90 degrees, lift, turn the lap, etc. I use only moderate pressure. When checking the contact patch with bluing I carefully place the new valve in the seat after cleaning the seat and lightly spin 20-30 degrees back and forth a few times with light pressure. If there is a high spot on the valve it can sometimes scrape the bluing away from the seat giving a false impression of a 360 degree contact patch if the valve is rotated completely 360 degrees around. Frankly, I'm not sure the bluing check is reliable. I had many cases where the contact patch was 360 degrees and the valve still leaked due to machinng marks left on the seat. That is why I only relied on the leakdown test.
I agree with Cogsy about invaludating your testing with oil. The oil is more viscous than air, and you can easily cover up a significant valve leak with oil. You likely won't have oil in the combustion chamber to help with valve sealing if the engine is working properly. Rings-yes, valves-no. - Terry

 

[Rustkolector]

Peter,
I'm not that experienced at making valve cages, but have had good luck so far as follows. Turn the OD to final dia. Drill the valve cage internal diameter, then drill the stem bore about .015" undersize. My stems are usually .125" so I then bore them a few thousandths more using a miniature carbide boring bar to assure stem bore concentricity before reaming the final stem diameter. Without disturbing the cage in the chuck, I use the compound to cut the seats .010" - .015" wide. Wider seats are problematic. Use as many passes as needed to get the seats smooth. I cut the valve seats using a very sharp miniature carbide boring bar with the compound set at 45 degrees. The seats and stem bore should now be concentric. I keep the compound at this setting until the valves are made, or vice versa. I lap my valves using a pin vise on the valve steam using the same method as Terry using a fine lapping compound. I do, however, heat shrink my cages in the heads before testing for leakage and draw the vacuum on the intake port using those little soft fittings MityVac provides. This allows stem leakage to influence the reading so I lube the stems with 90# oil, but the seats are kept dry. I look for a very slow leak down using valve spring pressure to hold the valves shut. Never measured the time, but most probably about 5-8 seconds from 25" to 20" Hg.

I use the MityVac also. I noticed early on that the clear soft flex tubing they provided on mine would often collapse under vacuum giving me odd readings like you experienced. If you experience this, switch to stiffer vinyl tubing, but don't kink it when storing it or you may have the same problem.
Jeff

 

[petertha]

..you've lost all the advantage of the seat cutter if you don't use it with the pilot

ok thanks. I will make an appropriate pin & do that. The tailstock method was an experiment.

- you mention a useful hand technique for lapping, is there a similar technique for seat cutting? ex - is the objective smooth rotation through 360 before lifting the cutter etc?
- anything special to know about doing this on bronze (cutting fluid vs dry?)
- to check seat width progress since you have no depth control, do you eyeball the shiny cut surface under magnification against a pre-set 0.010" vernier setting or matching wire?

Peter,
The flat is required because if you have a good fit between the valve stem and valve cage it will dominate the seal you are trying to measure

Aha! I understood the relief requirement to open up annulus, but it wasn't clear how this then applied to subsequent sets of stock round valve shafts thereafter. I thought you vac tested every valve + cage set insitu. But sounds like you use this dedicated reduced shank tester valve to validate cage seats only. As you say the valves are declared done on the lathe, only the cage seats are tweaked.

I polished the back of the valve head (the seating surface on the valve) while the valve was still in the lathe. I was able to get a brilliant finish free of machining marks with a scotchbrite pad. At that point I considered the valves finished and did nothing more to them. The lapping I did was on the valve seat in the cage.

I thought about pushing a wet/dry paper backed tool against the valve seat while at the compound setting to clean it up. But then I thought, what's the point. If my compound is actually +/- 0.5 deg off relative to the Brownells cutter angle, then its not going to mate properly anyway. This clarifies things, thanks.

- did you endeavor to somehow match the lathe compound to Brownell cutter angle when you made your valves & lapping spigots, or just strive for as close to 45-deg & assume that's what the cutter was?

 

[petertha]

Another question. If you look at the head section, last pic, post #3 - you will see the ceiling of the valve cage is a semi-circle dome. I didn't have the right size metric ball end mill at the time, but had a similar diameter imperial size 2-flute. It cut the shape no problem.

My metric one arrived, I ordered 4-flute thinking smoother finish. But it left me a chatter serrated finish & seemed to be grabby. I thought maybe because I didn't have the valve stem hole drilled the 2nd time & the center was dragging. So drilled the 3mm valve stem hole but still ugly. I moved to some scrap & tried a slightly undersized pilot hole first, then the ball endmill, same result - nice & clean hole segment but a spirally, starburst looking knurl on the dome surface & vibration on the tailstock.

What's going on here? Is 2 flutes a magic number when plunging or something I missed in the endmill specs? I tried traversing the ball endmill in the mill & it cut like a dream. Its the plunging/bottoming it doesn't seem to like.

 

[petertha]

I'm not sure your valves will be that oily, at least not the exhaust valve as the combustion heat should burn the oil off. On full size car engines it is/was common practice to check piston ring wear by doing a compression test dry, then adding a few drops of oil to the cylinder and testing again. .

That never even crossed my mind, thanks. I was thinking of the exhaust slime on the side of my RC fuselage after a run & premix oil on the intake side. But maybe there are other reasons to explain exhaust slime (induction flow crossover?). Anyway, dry it is. Back to work I go.

 

[mayhugh1]

Quote:

- you mention a useful hand technique for lapping, is there a similar technique for seat cutting? ex - is the objective smooth rotation through 360 before lifting the cutter etc?
- anything special to know about doing this on bronze (cutting fluid vs dry?)
- to check seat width progress since you have no depth control, do you eyeball the shiny cut surface under magnification against a pre-set 0.010" vernier setting or matching wire?


- did you endeavor to somehow match the lathe compound to Brownell cutter angle when you made your valves & lapping spigots, or just strive for as close to 45-deg & assume that's what the cutter was?

Peter,
When cutting the seat I hold the cage vertical in my hand and then let the weight of the seat cutter cut the seat for 3 or 4 revolutions. This will generally produce a seat width of .002" to .003" which, really, is good enough. Sometimes, depending upon how hard your seat material is you might have to apply an additional slight pressure. You want to do this in a practiced single motion with the rotation so you don't gouge the seat with too much pressure. I wet my phosphor bronze with WD-40 as a lubricant. I don't believe the seat width is critical. I'm still learning but I'm beginning to lean toward .005" as a max. I don't think it is super critical to match the angle of your valve to the angle of your seat cutter. As long as you're within a degree or so, it will just have the effect of narrowing your seat. After your engine starts running the seat will explosively form around your valve and all will be good.
It's just my opinion, but you can take two approaches to seating valves. The first, and the one I follow, is to over-do the sealing process and not depend heavily upon the combustion pressure doing the job I should have done myself. My personal goal is to get the 10-15 sec leakdown time that I discussed earlier. My experience is that this is an easy thing to do with a piloted seat cutter and a well polished valve on the lathe. This will be more than enough to seal the cylinder for any rpm range that you'll want to run the engine. Of course, combustion will improve it even further over time. The second method is to make the seal as best you can using whatever method you want and then hope for the best. But without some way of measuring your result you won't know for sure if you have a geometry problem. And you won't be able to fix a geometry problem with lapping. I don't think checking the contact patch with bluing is 100% dependable, and checking the seal using lung power means very little unless there is a really BIG leak. If you're building a single cylinder engine you might get away with number two. I wouldn't use number two for a multi-cylinder engine, though.
Remember also that I'm talking about testing the cages before they're installed. In my case, installing an untested cage in a head that had 10 hours of machining time on it wasn't acceptable. But this also meant that I had to install the tested cage in the head without distorting it. So far, high temp Loctite bearing retainer hasn't let me down. - Terry

 

[mayhugh1]

Quote:
My metric one arrived, I ordered 4-flute thinking smoother finish. But it left me a chatter serrated finish & seemed to be grabby. I thought maybe because I didn't have the valve stem hole drilled the 2nd time & the center was dragging. So drilled the 3mm valve stem hole but still ugly. I moved to some scrap & tried a slightly undersized pilot hole first, then the ball endmill, same result - nice & clean hole segment but a spirally, starburst looking knurl on the dome surface & vibration on the tailstock.

What's going on here? Is 2 flutes a magic number when plunging or something I missed in the endmill specs? I tried traversing the ball endmill in the mill & it cut like a dream. Its the plunging/bottoming it doesn't seem to like.

Peter,
The chatter in the 'dome' area probably won't hurt anything except the appearance. It's probably a sign that your rpm is too high for so much contact area in a set-up that isn't rigid enough and your feedrate was too slow. Cut the rpm down to as low as your machine will allow or even turn the spindle manually when you reach that area. Be careful to feed in only very slightly and only while you're manually turning the spindle. You can probably clean up your practice parts this way. - Terry

 

[petertha]

Awesome information. Internet sharing is a beautiful thing. Going back to work with a better plan.

 

[MachineTom]

A couple comments on your valve seating plan. You are building a model engine which in its life will run 10's of minutes per year. In full scale world with use at hundreds to thousands of hours per year things are different, and reliability and low downtime need to be built in each design.

It is normal practice to have the valve seat and valve have an interference angle of .5 to 1 degree, this provides only line contact at first start, but the valve seats in very quickly making a very good seal. When you start with matched valve/seat angles then lap them in, if lapped too much, you end up with a valve that has a dished face, that seals perfect at 70 degrees, but when running at maybe 800-1000 degrees expands and now fits quite poorly.

I would suggest blue the valve, and check the seating, cut the valve seat at in the same setup as when you ream the valve guide, so the seat and guide are concentric. Keep the valve seats narrow.