Compression vs Leakdown Tests

[T.W.R]

Recently had this article arrive in the Ktech newsletter and though it worthwhile posting here to shed a bit of light on Compression & Leakdown tests

Pt1) The Compression Test

Everyone remembers how to do a compression test: screw in the tester, disable the ignition system to prevent sparks and system damage, hold the throttle wide open, and press the starter button (or kick). The cylinder take in air and compresses it, the tester traps it and the maximum is reached when the gauge holds as much pressure as the engine can produce.

Not too difficult, right? The weakness of the compression test however is that throttle position, engine ambient temperature - even battery condition - can make its results vary considerably. What's worse, a compression test checks too many components at the same time. A poor reading can indicate so many things , it's hard to tell which engine part is at fault without doing other tests. One of those tests is of course the classic squirt oil in the cylinder. If a subsequent test yields a higher reading, the cylinder or piston rings are indicated as the problem area, and not the valves.

[T.W.R]

Pt 2 The Leakdown Test

A leakdown test is very different however, and here's how to do it.

Raise the piston in the cylinder to be tested to TDC compression. Plug your tester into an air compressor line and adjust the regulator to get a "0" reading. Screw the hose into the spark plug hole, and then connect the tool to the hose. If the crankshaft turns, or you hear air rushing out of an obviously open valve, the cylinder wasn't exactly on TDC compression. Try again. When you get it right, the piston will stay put and the tool will indicate the amount of air that is escaping from around the rings, valves, and head gasket of that cylinder.

You can see that a leakdown test works differently from a compression test. Air is pumped into the cylinder from an outside source, instead of by the cylinder from within, and the gauge reads the percentage that escapes, not what accumulates. This eliminates the afore-mentioned variables, and as a bonus makes it easy to pinpoint the source of the leakage by simply listening to the engine while the test is still underway.
High readings accompanied by hissing in the carburettor indicate burnt, tight, or carboned-up intake valves. The same thing in a muffler points towards exhaust valves. A breeze coming from the dipstick hole indicates a worn cylinder or worn piston rings. And, air escaping from an adjacent sparkplug hole or from an open radiator pinpoints a blown head gasket.

Sometimes a high reading may be the result of carbon on the valves, and not valve or valve seat wear. It's easy to verify this. Remove the valve cover and with a hammer and drift, carefully tap on the valve followers for that cylinder, watching the gauge as you do so. This will often loosen carbon from around the valve and the reading will drop to a reasonable level.
In that case, the leakage is only temporary, and a good hard ride will fix it. If on the other hand shaking loose the carbon doesn't lower a high leakage reading, well, time to get the engine fixed.

[T.W.R]

Pt 3 But what is a high reading?

All cylinders leak a little, and large ones leak more. Production engines in good condition commonly leak 5% or less, but the limit is accepted as 10% before engine work is needed. If you're testing a big bore single or twin, get help holding the crankshaft because there won't be enough friction in the engine to resist the compressed air, even exactly at TDC, and the crankshaft will turn under pressure. Just as importantly, for cylinders over 100mm in diameter, allow a bit more leakdown, up to 15% instead of the usual 10%, before condemning it.

But consider a different way of using your leakdown tester. Let's say you have a four cylinder street bike that produces good compression readings, but smokes out the exhaust. All cylinders leakdown less than 10%, but one of the sparkplugs is oil-fouled, and the same cylinder oil-fouls a replacement sparkplug put in as a double-check. So you know the problem is limited to one cylinder, but what is the problem? Not very likely that just one valve guide is cracked, is it?
Grab your leakdown tester, and lower the pressure setting on the instrument about halfway down so that not so much force is going in to the cylinder that the crankshaft wants to rotate. Now repeat the leakdown test in your suspect cylinder, only this time lower the piston down in the bore a little. If the gauge suddenly reads 60% leakdown when the piston is halfway down the bore, it obviously indicates cylinder bore damage.
When you pull the engine apart you find that the gudgeon circlip has come out of it's groove and machined a trench into the cylinder wall. You already know why the high pressure leakdown test didn't find this problem, because the piston was at TDC, well above the damage. But why didn't a compression test find it? Similarly, the compression test didn't show it because there was still plenty of cylinder area above the damaged area in which to build adequate pressure during a compression test.
The fact is, one of the drawbacks of a compression test is that very little movement of the piston is required to build maximum cylinder pressure. But a low pressure leakdown test picked up the problem instantly.

[Motu]

It pays to verify TDC and lock the engine before you apply the air - one of my mechanics lost the ends of a couple of fingers when turning the engine with a leak down tester.He didn't have it on TDC and was bringing it up by turning the water pump pulley,it flicked over and caught his fingers in the multirib belt.Not likely on a bike,but caution anyway - lock it and don't touch unless you release the air.

A compression test is still quicker,and still first tool - also good for picking up incorrect cam timing,a very high cranking compression is an indication of advanced inlet timing.Another very useful test is Cold Crank Vacuum - you need suck before you can blow.I used this a lot in the gas conversion days,unlike petrol which is heavier than air,gas needs to be drawn in.But a very useful test for hard starting.

[roadracingoldfart]

Interesting But... if you do a comp test and the reading is within specs (already researched) then you dont need to bother with the potentially damaging leakdown test.

Dont just plug in to the air and crank the regulator open to 120 psi is very important. Yes some idiots do it , even long time mechanics doing it for the first time.
Recently at my work a 12 yr quald tech was about to blast 120 psi into an
XR8 just to hold a valve seated so he could replace the broken valve spring.