Glowgeek

I got a very low run time os fs26 in a trade. It was heavily gummed with castor, like the rest of the engines I got in that trade, so I used my standard cleaning process. Antifreeze in a crock at 170° for 12 hours and everything cleaned up to like new condition. So far so good. I reassembled the engine and found an exhaust valve leak. Took the head off and recleaned mating surfaces, reassembled and it still leaked. Touched up the valve seats by lightly lapping with polishing compound and tested for valve leaks using a rubber cup over the combustion chamber and a mityvac pump. No leaks so great, I think I've won the battle but no, once reassembled it's leaking again, just as bad as before lapping.

Here's the thing, the valve begins leaking as soon as the head bolts begin tightening. The more I tighten them the worse it leaks. I found no evidence of cracks or burrs on the mating surfaces of the case, liner or head.

I'm lost, has this happened to any of you? Any thoughts on what might be going on?

706jim

Sounds like the head might not be flat and distorts when the bolts are tightened. I can't remember if there is a flat or trepanned (grooved) mating surface. If flat, you could sand it on a flat piece of glass or other surface. It trepanned it would be a lathe job and a very awkward piece to hold.

Glowgeek

Thank for the reply Jim. The surfaces are flat and designed/assembled like many 2 strokes. The liner rim rests on the top of the cylinder, the head rests on top of the liner rim and the 4 head bolts are located just outside of the perimeter of the liner rim. Also note that the bottom of the head only comes in contact with the liner, there is a gap beyond that. The head bolt location would definitely lead to a cantilevering effect when tightened and even though this may not cause ill effect on 2 stroke I fear that the cantilevering effect is causing distortion of the valve seat on this 4 stroke.

I'll check for a crack in the head again as I find it hard to believe that OS wouldn't have taken all this into consideration during design.

706jim

Another possibility is a small void in the casting internally. Would be difficult to detect. BTW I use the same username on RC groups but the system won't recognize my password. And the password reset link doesn't work. Not much action on this site.