leak down vs compression test?

[robert walsh]

I have paid for both at different points in my life, I have no idea how a leak down test is preformed or what it means, except i got some #'s and they said they were good. Compression test, I understand and as such understand what the #'s mean. Why is one test preferred over the other and which is preferred on a porsche (air cooled). PLEASE explain how a leak down test is done. Thanks

[Peter Zimmermann]

1. Place the engine on TDC for the cylinder to be tested.
2. Remove that cylinder's spark plug - and only that cylinder's plug.
3. Screw the leak down tester's connection hose into the spark plug hole (this is better done on a cool engine).
4. Connect the tester to the hose.
5. Connect a regulated air supply set at 100 psi to the tester.
6. Read the gauge. Some testers use two gauges; one measures air in, one measures air held in the chamber.
7. Remove the tester and hose, manually turn the engine to the next cylinder - repeat.

http://forums.rennlist.com/rennforums/showthread.php?t=189606&highlight=PPI+by+Peter+Zimmermann

[Zeke]

Well, PZ explains the procedure, but not why one would prefer a leak down to a compression test. The CT will tell you how well the cylinder is functioning, but not pin point the problem if it's not. The LD will tell you if the rings or either the exhaust valve or intake valve is bad. It will also detect a head sealing problem. Now, if you have more than one problem on a particular cylinder, it becomes harder to tell, but it will still direct you to one source more so than something that is in good condition.

Let me briefly elaborate: A bad intake will hiss through the intake manifold and the exhaust will his thru the exhaust system. Bad rings can be heard "leaking" by listening to the case breather. And a leaking head can be heard from underneath or observed by spraying a liquid on the mating surfaces. By these observations one could reasonably decide if a top end job would fix up the problem vs. a rebuild including pistons and cylinders.

There's probably more.... but that's my take.

[2.70Racer]

Easy to remember, If your engine has good compression numbers, no need for a leak down test.
Bad compression numbers, then do a leakdown to help pin point the problem before paying for a teardown and repair.
Compression test always first, leakdown only if needed.
Not so tough, eh.

[robert walsh]

Thanks I understand it now, can a leak down test be done on a liquid cooled engine? If so why is it not more common?

[Dantilla]

Yes, a leak-down test works on a liquid cooled engine. I have a leak down tester a friend found at a pawn shop. It came with adapters for different spark plug sizes.

[Zeke]

A leak down can be performed on just about any engine with a spark plug and I suspect there are diesel adapters. In fact, I just googled it and they do diesels all the time. Even 2-strokes.

[litning]

I'm about to buy a Porsche engine that has been sitting in a car under a roof, but otherwise exposed. I requested a compression test and a leakdown test. I got good compression (with a bell housing attached but the car is not running) and 5 good leakdown numbers 0,0,0,5,5 and a leakdown of 15 - 20% on a cylinder that had a great compression number. The first try of the leakdown on that cylinder was as good as the others, but on several additional tries the leakdown was the 15-20% number. It's kinda hard to make a purchase decision due to that cylinder but I'd appreciate any ideas/thoughts on the matter. The price would be good for a very good engine, but if I come up with a bad cylinder it's gonna burn up any advantage of swapping out my 2.7 for the 3.2. I'd end up spending as much as a rebuilt engine. Help!
Litning

[2.70Racer]

Litning,
Good compression readings and poor leakdown readings at the same cylinder are not compatible and indicate one thing.
Likely the leakdown test was performed wrong or the old loose piece of carbon is keeping a valve from fully closing.

A compression test requires a crankable engine. Reliable compression tests are performed on a warm engine. Not possible with a barn engine not running.

For a barn engine not running or crankable, a leakdown test is all you can do to test the integrity of the cylinders.

First you mention the compression test was fine for all cylinders.
Were the numbers all around 140 or better?
If so no further need for a leakdown test.

Second a leakdown of zero is suspicious, even new engines leak a percent or more.

Further a high leakdown measured at a cylinder that initially read good compression indicates one test or the other was not done properly or was faked.
My suspicion is the numbers reported to you are not correct. They simply don't match up. It is not possible for a cylinder compression to be good and a leakdown test for the same cylinder to be poor.
A compression test is simple. A compression test done wrong will result in low numbers.
A leakdown test requires some skill.
Repeating here again, a leakdown is called for if you first get low compression readings. The leakdown will pinpoint where a problem may be.
A leakdown test where the low readings are not reported as to where to potential problem is has no more value than a compression test.
A leakdown test is to SHOW WHERE THE LEAK IS.

[Gunter]

Originally Posted by Peter Zimmermann

1. Place the engine on TDC for the cylinder to be tested.
2. Remove that cylinder's spark plug - and only that cylinder's plug.
3. Screw the leak down tester's connection hose into the spark plug hole (this is better done on a cool engine).
4. Connect the tester to the hose.
5. Connect a regulated air supply set at 100 psi to the tester.
6. Read the gauge. Some testers use two gauges; one measures air in, one measures air held in the chamber.

O.K. 100 PSI in but over what period of time is any loss of pressure noted?
How long does the cylinder have to hold the pressure?

7. Remove the tester and hose, manually turn the engine to the next cylinder - repeat.

http://forums.rennlist.com/rennforums/showthread.php?t=189606&highlight=PPI+by+Peter+Zimmermann

Is the air supply a constant 100 PSI and the gauge shows a continuous loss of pressure in the cylinder?

Like one would have to set the output on an air compressor to a constant 100 PSI?

[2.70Racer]

My leakdown tester has an input regulator. The air compressor puts out 110-120 PSI.
I block the out put of the L/D tester, then set the input pressure at 100 PSI as shown on the first gauge and second gauge.
Then connect the output hose to the cylinder at the sparkplug hole.
With the cylinder at TDC compression, (both intake and exhaust valves closed).
Then open the air flow to the cylinder.
Maintain 100 psi at the first input air gauge. It may be necessary to readjust the input regulator to bring the input gauge to 100 psi.
100 psi input pressure merely makes the L/D math simple. 95 or 105 will work, just not simple math.
The second gauge will measure the air pressure minus the leakage in the cylinder.
A 100 reading at the input with say a 90 reading at the output gauge, means 10% leakage.
Yes the air is flowing from the air compressor during the test.
The idea is there is a small orfice in the tester manifold between the input and output gauges. If the cylinder leakage is small, the output gauge will drop very lttle compared to the input gauge. The greater the leakage, the lower the output gauge will read.
With 100 psi input air, the percentage of leakage is simply the amount below 100 at the output gauge. For example a output reading of 85psi vs 100 psi at the input gaugge means 15% leakage. Air in the exhaust is a leaking exhaust valve. In the intake, is an intake valve. Air in the crankcase is flowing past the rings. none of the above can mean air past the head gasket. Most engines will show some leakage, 2% to 5%. Generally past the rings.
This test is best done with a warm engine. A cold engine will generally read lower due to leakage past the rings.
Low compression readings should be backed up with a L/D test. Good compression readings, no L/D test required. Low compression readings should correlate with a low L/D reading.
The leakdown tells you where the problem is. A compression test cannot.

[litning]

Thanks 2.70Racer - I dont mean to hijack this thread, but we all learn something from these messages. I had an independent shop check out the engine. His tests indicated good leak down at lower pressures (he starts from 0 then up to 100) but the leakdown fell off at higher pressures to 15-20% leakdown. he felt that the leak was from the rings. Remember that he said his first test gave him good leakdown, but when he repeated the test that cylinder wasnt testing very well. All cylinders have been tested several times now, with the same results (0-2% on 1-3, 5% on 4&6, and 15 - 20% on #5). My independent (I pay him) has a good local reputation 3000 miles away, but he did say that he had checked the compression and it was good - maybe there is another way to check compression??? Boresight of the #5 cylinder didnt reveal any scoring. The head studs looked good and everything under the covers looks clean - I'm leaning toward buying the engine (54K miles) and "rolling the dice". I dont know if a poor leakdown cylinder can "fix itself" but we are hoping

[longhornchris04]

Did the leakdown test provide where the leak was likely from? I know if the engine has not been run hard then carbon can build up on valves/seats and cause poor sealing.

If thats the case then the engine can "fix itself" by getting it good and hot to clean out the carbon... assuming that it isn't a significan oil leak. I believe this is refered to as an "italian tune-up" - put some good gas in (well, you should always put good gas in), throw in a bottle of fuel cleaner, warm the engine up then go push hard for a while to ensure any carbon gets burned out.

Of course, if it is the piston rings, this won't help much. Still, not a bad plan.

Out of curiosity, what's the year on the engine you are looking at?

[dad911]

I have not heard of getting different results at different pressures.

at 15-20%, he should have heard the air rushing out a valve or crankcase.

I tend to question the procedure the shop used.

[litning]

It could be the tester (man or machine) but he only had a problem with the #5 cylinder, and he eventually decided that it was thru the rings. It is an '87 engine. I decided to go ahead and buy it, because the price was right and it was pretty low mileage. it's a bit of a crap shoot, but we'll check it out locally when it arrives. At worst, I've gotta do a top end which would raise my investment up to the cost of a rebuilt engine and I'd probably have to sell it. I'm hoping for the best, where I would get many years of use (at the rate Im driving this '77) before it needs anything else. This 3.2 swap is for fun; I'm the original owner of this '77 Targa and it's got just 82K miles on it. When my clutch leg finally gives out I'll have a good looking 32+ year old Targa and two engines - one fast and the other original!