Leakdown test- What is acceptable?

[sww914]

What's a go/no go limit for a leakdown on a cold engine with a few miles on it since a top end rebuild?
I'm looking at buying a used 2.7. I know that less that 5% is good on a warm engine, but it's out of the car and I can't warm it up.

[cstreit]

I've consistently seen only a few percent difference on cold vs. hot numbers on good higher mileage engines. However... I do occasionally see some engines where the hot leakdown was significantly improved... 5% difference. YMMV.

[john walker's workshop]

hard to say, because proper leakdowns are done on hot engines that have thermally expanded all their parts and have oil on the rings. 3-4 hot, probably double on cold. below 8 maybe.

[2.7RACER]

First let me preface this that I have little experience measuring lots of engines for leak down or compression.
That is to say I understand the purpose of each test, but I do not have the experience of actually measuring a large number of engines and developing a go, no-go feel.
That said it struck me about the variables using a leak down test on a warm vs cold engine.
It would seem to me the leakage past valves would not be significantly different warm or cold. Assuming the valves are fully closed when the test is done.
Therefore it leaves leakage past the rings or possibly the seal at the head which could vary depending upon the temperature of the engine.
If the rings are suspect, a small amount of oil sprayed into the cylinder and a retest should tell you if the leakage is rings, since leakage past worn rings would be reduced after being sprayed with oil, whereas a head leakage or valves would not respond to oil sprayed into the cylinder.
So I would expect a older higher mileage engine to have low leakage at the valves, some leakage at the rings, no leakage at the heads.
I would also expect the ring leakage to be similar from cylinder to cylinder.
The question I have for those with a great number of engines tested is,
"In your experience, what percentage number from a leak-down test would tell you it's time to tear down the engine?"
Or do you need the compression numbers along with a leak-down test to feel confident?
I'm guessing over 25% leak-down. which comes up significantly with the oil sprayed into the cylinder.
Comments please,

[CliffBrown]

This is what the FAA has to say.

In the United States, FAA Advisory Circular [1] 43.13-1B [2] specifications, chapter eight, state that engines up to 1,000 cubic inch displacement require a 0.040-inch orifice diameter, 0.250 inch long, 60-degree approach angle. The input pressure is set for 80 PSI, and 60 PSI minimum cylinder pressure is the accepted standard.

[bluebullet]

Did a valve adjustment on an 86 today, and leakdown was in the 2% range cold and cylinder 3 was 6%, but I think that was due to the compressor was charging up the air system in the shop, and it was throwing the leakdown tester off.

[engguy]

Cliff, that is okay with a test set with a calibrated orfice. The one I have if I set the input pressure at 80psi, I could have a huge leak and still read close to zero leakage. These china made things are not the same unit.

[hcoles]

Cliff Brown above....says re. FAA/etc.
Has anyone found a more complete spec. for a leak down tester?
Here is what I'm thinking... I/we need a spec. that would list a given flow rate of air at room temp. and pressure going through whatever type of orfice should have a listed pressure drop across the orfice. This then leaves it up to the constructor to make the device that meets the spec. I also assume that if the pressure gauge reads 72psig downstream of the orfice and 80psig upstream of the orfice then the leakdown percent is 10%. Is this correct? The problem for me is that I haven't been able to find a spec. for making/testing/calibrating a legal leakdown tester.

[antares]

. . . Good luck as this is all part of the continuing dark art found in porschedum.

[Alan L]

Originally Posted by engguy

Cliff, that is okay with a test set with a calibrated orfice. The one I have if I set the input pressure at 80psi, I could have a huge leak and still read close to zero leakage. These china made things are not the same unit.

I agree with this. I made one up to the 40Thou orifice spec (this is the only standard you will find and any 'calibrated' tester will be this spec). But note the spec is for engines up to 1,000 Cu in. A 3 .0 Sc is 186 Cu in or there abouts. This has the effect of hiding any leak issues until significant - the 40 thou orifice is capable of pouring enough air in to deal to a leaking 15 l engine. So it is putting in more than enough for a 3 l engine. I adjusted mine to 20 thou from memory, in order to get more precise numbers for my track machine. If I'm getting say 5% on one cyl,then on the 40 thou spec it probably barely registers (and looks OK - whereas it may have a fault - eg broken ring). But the 40 thou spec will still pick up a poor cylinder if it has a fault - it just may be a slight bit misleading as to what the significance of the issue is. But there is no auto industry standard for this, so we either live with the only standard we have, or adjust accordingly to what you need it for.
My 2c worth.
Alan

[afterburn 549]

Originally Posted by CliffBrown

This is what the FAA has to say.

In the United States, FAA Advisory Circular [1] 43.13-1B [2] specifications, chapter eight, state that engines up to 1,000 cubic inch displacement require a 0.040-inch orifice diameter, 0.250 inch long, 60-degree approach angle. The input pressure is set for 80 PSI, and 60 PSI minimum cylinder pressure is the accepted standard.

I do not trust much what aviation standards are for recip engines as "their technology is in the dark ages. ( Due to miles of paper work to change one freekin bolt )