Do-it-yourselfers are not usually aware of the many intricacies involved in professional engine diagnostics. Most are familiar with the cylinder compression tester, but fewer with the cylinder leakdown tester, and fewer yet with how the two tools should be used together.
As a reminder, you will recall that the compression tester takes in air pressure developed by a cranking engine and holds it as a maximum. The correct technique is to hold the throttle wide open, as this is how the engine gets all the air it can. Whether the engine is hot or cold makes negligible difference. And, putting oil in the cylinder is not the best technique either.
But back to our two tools. It is not correct to say the leakdown tester obsoletes the compression tester. For, while the leakdown tester can find things the compression tester cannot, the reverse is also true -- the compression tester can find things the leakdown tester cannot. You simply have to know when to use each, and ideally you will use both. Take for example, an engine that runs poorly. A leakdown test indicates below 10 percent, which by itself you have to call good. A compression test however shows a problem, as the result is only 100 psi. How can this be? Well, remember, the one test -- cylinder compression -- measures the engine's ability to produce its own pressure, which relies on the camshaft, not just the cylinder, valves and gaskets. On the other hand, the other test -- cylinder leakdown -- measures the engine's ability to hold pressure brought in from the outside, which really involves only the cylinder, valves, and gaskets, not the camshaft. They're very different tests. So what would cause an engine to hold pressure brought in from outside, but not develop its own? Remember, the camshaft is the difference. Cam wear will be on the list, but more importantly, cam timing. Now, this poorly running engine, did it gradually run worse or all of a sudden? If gradually, you're looking at a cam timing issue due to cam chain stretch, or less likely, a very worn cam lobe(s). If sudden, it is likely jumped time or very recently misassembled out of time. Cam wear and incorrect cam timing will not be detected by a leakdown test, but a compression test will detect it. Cool, huh?
Now let's go the other way. Let's say we have an engine whose compression test is low-to-fair, at say 140 psi, and whose leakdown is bad, i.e. 20 percent. The lower than normal compression we'll keep in the back of our mind. The camshaft and its timing aren't problems, and even the cylinder, rings and valves can't be too bad or the cylinder wouldn't make the compression it does. The next step is to repeat the cylinder leakdown test and thus pinpoint what little weakness, one not really obvious, is both producing low-ish cylinder compression and higher than acceptable cylinder leakdown. With the leakdown test underway, remove the air filter element and listen at the airbox. Then put an ear to the exhaust pipes. Follow these by unscrewing the crankcase oil filler or dipstick and listening there too. Rushing air sounds at any of these points indicate sealing issues, with intake valves, exhaust valves, or cylinder and/or rings, respectively, most often the valves, and most often exhaust, as although both can suffer wear, the exhaust is usually the worst. The most likely kind of wear in this case, and one that gives mechanics a hard time because it causes the valve to sometimes seal well and sometimes not, is what is called valve recession, an erosion of the valve sealing face caused by rapid wear. This is possible in all four-stroke engines but was the most prevelant on early to mid 1980 model Japanese bikes.
But how could a cylinder seal well when its valves don't? First, the already-mentioned tendancy for recessed valves to seal intermittently is one reason. But also, remember there are some drawbacks to the cylinder compression test. A compression test's pressure is cumulative, so small defects can go undetected. And, only a very short movement of the piston is needed to build reasonable pressure in the cylinder. Spinning the engine faster and longer can make up for a lot in the compression test, making it not always conclusive. I once found a cylinder with a hole in it (see my article) that despite the hole built adequate compression. A leakdown test was good on it too. To read about how I found the problem, go here. :-)