Cylinder Leak Down Testing - Revisited

[creia]

As some of you may already know from one of my previous postings, I am a newcomer to doing this testing. I believe that I have the technique down pretty good, however, I am noticing that the results (% leakdown) on the same engine seems to vary, sometimes significantly so, when conducting multiple tests. I can get results of 10-15%, then disconnect the hoses and rotate the engine a few cycles, hook everything back up then get 50% leakdown. If I rotate again and take another reading it is back to 10-15%. I do not believe it is a faulty tester as I have 2 different testers (different brands) and they will match each other in results. Is this typical with these testers? If it matters, when I get leakdown results (leakage) , it is at the cylinder (from the oil fill) and not at the valves.
As always, any help is appreciated.
Michael

[38racing]

I bought a unit with a quick connect and found it screwed up my reading as a leakage point.

[KE4AVB]

I suspect that the variance might be ring flutter (side play of the piston grooves). otherwords if the piston was traveling up, the ring are somewhere near the bottom of the grooves but as you air pressure the piton moves downward and rings moves up temporarily unsealing them as air leaks around the rings. The faster things moves less noticeable it is. Just a working theory so take it with a grain of salt. BGS is probably a lot more experience at this testing than I.

Also watch the tester base settings as if it like the one I got the setting can change due to regulator leakage and give a false good reading. The HF one here has a tendency to creep if you take a reading at the start then change if it stays connected. Otherwords I take the reading quickly and then disconnect, reset if needed, and retest as needed. But of course I am doing top to bottom of the stroke testing on OHV engines too. Found several Tecumseh engine here that giving problems to have lower cylinder problems by using this modified leakage test.

[bgsengine]

what KE4 said. I would suspect the procedure. You are actually backing your regulator off to zero before connecting air hose, right? and you are using a consistent input pressure of CLEAN air (like any air tool, dirty wet air from an air compressor - more often with older ones - can affect air tool performance) Each time you do the test , start with regulator zeroed down , hook up shop air and find your set point, then hook up tester hose (which is already hooked to engine) to tester gauge set, THEN prep engine and open tester flow valve fully. KE is right you can get some ring flutter, but I doubt it would account for THAT much leakage rate difference. I more suspect it may be variations in how the testing is done.... You'll almost always get different results if your procedure to hook up and test is different (I.E. run test , leave gauge hooked up to shop air or some such and turn engine over, and then try testing again, yep, different results.. I had the same experience... so learned quickly to do the FULL setup process from step one to set up the tester each time, no short cuts.) and oh , yeah obviously you do lock down the engine at the same position? if you cannot duplicate results of a test on same engine using same step by step procedure, your engine you are working with may very well have a bad spot which you happened to find by chance.. But again, I'd only do a leakdown test as a diagnostic tool, if it passed the first time, I would not see any need to try and repeat the test.

[KE4AVB]

And one that operates at a lower pressure (the HF one @12 PSI in) vs the standard 100 psi ones you do a modified leak down test on OHV engine when cripple the valve train.