Testing an engine for compression in the cylinders often yields information you might want to have on hand when buying or selling a racecar. Buyers often want compression numbers. Sellers would be wise to have them at the ready. Solid numbers can help demonstrate the general overall health of the engine, whereas numbers that are off might help you avoid acquiring someone else’s headache.
A compression test is a decent way to assess the sealing functions of an engine’s valves and valve seats, piston rings and cylinder head gaskets. You can dive more in depth to pinpoint where a cylinder or cylinders might be leaking with a leak-down test, and we’ll take that on in a later edition of Speed News, but for now, let’s focus on the compression test.
It’s a pretty easy test to perform if you have the proper tool, which in the days of Harbor Freight and Amazon, is more affordable than ever. Don’t go too cheap on the tool, though.
The procedure is pretty straightforward. First, you need to bring the engine up to operating temperature. When the engine has been warmed appropriately, you must find a way to shut off fuel delivery. In a carbureted car, it’s not that big of a deal because the accelerator pump squirts fuel only when you depress the gas pedal, but injected cars will keep pulsing fuel every other revolution. The easiest solution is to pull the fuel pump fuse.
Next you have to find a way to shut off the ignition. With coil-on-plug ignition systems like the one shown in the photos, that’s pretty simple because you have to remove the coils to be able to remove the plugs. There are a number of ways to quell any spark, from unplugging the coil from the distributor or coil packs or disconnecting the ECU. The simplest way is usually best.
It’s also a good idea to have a battery charger on hand, and to have fully charged the battery first. All the cranking you have to do can drain the battery a bit, and if you have to test eight cylinders, you might run the battery low.
After you have removed all the spark plugs, thread the appropriate compression test hose into the No. 1 spark plug hole and connect the hose to the compression gauge. Now you’re ready to perform the test on cylinder No. 1. As you are cranking the engine over, you must have the throttle at wide open before you engage the starter. Performing the test with the throttle at idle position will yield low numbers and an inaccurate reading.
This is easy on cars with a throttle cable. If you have a drive-by-wire car, make sure the throttle blade operates with the starter engaged, the fuel pump fuse out and the ignition disabled. If it doesn’t, you’ll need to find a way to prop it wide open.
When you crank the engine over, count the revolutions and use the same number for each cylinder to create equal results. For this test, we rolled the engine over four times and you can hear that in the video toward the end of this story. What is equally as important as the numbers themselves is the variance among all four or six or eight. Or five if you have an Audi! The general rule is that you want all the numbers to be within 10 percent of one another. The numbers from a fresh Mazda MZR we used for this test were within 1 percent of one another.
Another way to perform the test is to crank the engine till the needle stops rising and use that rotation count for all cylinders. That likely will yield higher numbers overall, but again, test them all the same way and recognize that the variance among all cylinders is equally important.
The numbers that appear on the gauge will differ depending on the compression ratio of the engine and — on high-compression engines — the amount of overlap between the intake and exhaust cam lobes. The chart below is a good guide for measured compression numbers that should be present based on compression ratio:
8:1 to 8.5:1 compression: 150-170 psi per cylinder
8.5:1 to 9.5:1 compression: 170-210 psi per cylinder
9.5:1 to 11:1 compression: 210-275 psi per cylinder
11:1 or higher compression: 250-plus psi per cylinder
If you come across a cylinder that is well out of spec, spray some oil — about a teaspoon — in the cylinders and try the test again. If your numbers improve, that usually indicates worn piston rings. It not, it could indicate a leaky valve or valve seat.
As valid as a compression test is, even the Society of Automotive Engineers acknowledges it’s not the be-all, end-all test of an engine’s overall health over time. An SAE research paper showed that compression tests done on two fleets of vehicles during periods of mileage accumulation showed that, “The relative overlap in the magnitude of leakage changes and measurement uncertainty shows that the leak-down and compression tests are not useful in their present forms for monitoring incremental changes in engine leakage.”
In other words, a compression test is a decent “snapshot” of an engine’s health, but for monitoring small changes in engine performance over time, it’s still not terribly precise. With that acknowledgement, it becomes all the more important to perform it properly when you do check it.
Be sure to perform the compression test with the throttle wide open. Be sure drive-by-wire systems will operate with the starter engaged and the fuel pump and ignition systems disabled.
Keep a piece of paper handy so you can write down the numbers as you test each cylinder.
Thread the connecting hose into the spark plug hole. Hand tight is fine because it has an O ring seal at the base of its threads.
Once you have the compression gauge attached to the connecting hose, you are ready to perform the test. Remember to press the gas pedal all the way down to achieve wide-open throttle.
Now repeat the procedure for cylinders 2, 3 and 4. Five-, six- and eight-cylinder engines take more time and effort, but it’s important to measure the compression in all cylinders so you can compare them. The numbers should fall within 10 percent of one another.
The numbers on this engine fell within 1 percent of one another, which is good because the engine had less than two hours of run time.
VIDEO
Here’s a brief synopsis of how to do a compression test on a four-cylinder engine. Video shows the action of the compression gauge and the sounds the engine makes while testing.
Sorry, but the numbers are not within 1%. You have to compare the lowest and the highest — that’s 103 and 108. It’s 2.5%. You were looking at how close to the average, not how close to each other.
Thanks for the correction. Math was never my best subject. BB