I got a message last week from my friend Craig Fitzgerald, asking for help with his 1965 Corvair Monza. As president of the New England Motor Press Association and editor-in-chief at BestRide.com, Craig has his hands in a number of automotive-related pies. But not surprisingly, we have the most fun when we’re just two regular car guys. This was one of those times.
Like me, Craig is a Hagerty customer and a serial acquirer of affordable interesting cars. A few years ago, he bought the very solid, running Monza for $1500. It didn’t need much other than light tinkering to be a knockaround weekend driver. Last spring, I helped him troubleshoot an intermittent no-spark problem, which, to no one’s surprise, turned out to be pitted points. Having never laid my hands on a Corvair before, I found it massively cool. The engine is a boxer-style air-cooled six, laid out similarly to a Volkswagen or a Porsche pancake engine. A pair of aluminum heads flank three air-cooled cylinders on each side. But distinct from the VW or Porsche are the twists and turns that the fan belt takes. They’re almost Mobius strip-like.
Craig’s message last week, unfortunately, was that, when he was changing his oil, a piece of a piston ring came out of the oil pan. Although the car seemed to be running OK, obviously this wasn’t good. We swapped messages on what could be done to try and find which piston the ring came from. He had a compression tester. I offered to come over with a leakdown tester.
Compression testers and leakdown testers are like a stethoscope and thermometer, providing the one-two punch in determining engine health. Most folks reading this are probably familiar with the use of a compression tester. It’s simply a pressure gauge with a check valve, attached to a hose that you screw into the spark plug holes. You first prevent the engine from starting by disabling the ignition, screw the compression tester into each plug hole, crank the starter so the engine turns over five or six times, read the compression, and then move onto the next cylinder.
It’s important to keep in mind that compression readings are proportional to how quickly the engine spins. If you have a weak battery, the compression readings will be artificially low, so it’s important to conduct the test with a fully-charged battery. In addition, I typically remove all the spark plugs at once so the engine can spin freely and isn’t fighting against the compression in the other cylinders. Lastly, on vintage cars with points-based ignition, even if you’ve removed all the plugs at once to prevent the engine from starting, it’s good practice to disconnect the power lead from the coil (the side labeled “+” or “15”), so the coil isn’t being fired without a path to ground. And if the car has an electric fuel pump that runs as soon as the key is turned, it’s a good idea to disable it so that the cylinders aren’t being flooded with fuel during the compression test.
The compression tester gives you cylinder-by-cylinder compression and lets you find if any cylinders are low, but it’s the leakdown tester that helps you find why compression is low. A leakdown tester is simply a pair of pressure gauges, an input connection to an air compressor, and an output connection that threads into the spark plug holes. What you’re doing is using the air compressor to pump air into each of the cylinders, and reading the inlet and outlet pressures on their gauges. If you want the leakdown numbers, you take the ratio of the outlet to inlet pressure, express that as a percentage, subtract that from 100% percent, and the result is the percentage leakdown. For example, if the inlet pressure is 100 psi and the outlet is 90, the ratio is 90/100—or 90 percent—which means that there’s 10- percent leakdown (in other words, 10 percent of the pressure is leaking).
You may read that, in leakdown tests of freshly-rebuilt motors, the percentages should be in the low single digits. Perhaps this is true, but my experience is primarily in leakdown-testing of well-used motors, in which the numbers are often far worse, like 20 or 30 percent. When I initially encountered this, I was horrified, but after talking with a number of mechanics who work on vintage cars, this appears to be not at all unusual with old motors, and is simply a sign of age and use, not of imminent demise. Similar to compression, what you want to see are even numbers across all cylinders. I’m philosophical about these sort of things, as there are many human-applied numbers, such as weight and waist size, that increase as we get older. Why should our cars be any different?
But if, for example, 20 percent of the pressure is leaking, where is it going? This is the beauty of the leakdown test. Forget the percentages. If you’re engine has low compression in a cylinder, the way to conduct the test is to get that cylinder to top dead center (we’ll get to that in a moment), screw the tester’s hose into that cylinder, fire up the compressor, get about 50 psi or more into the cylinder, shut off the compressor, and then listen and probe. If you hear nothing, then no air is leaking (0 percent leakdown). But you never hear nothing. Basically, air can leak out in five places:
Past the piston rings, between the piston and the cylinder wall, into the crankcase, and out the path of least resistance, which is usually the oil fill cap or the vent that goes to the PCV valve.
Past the closed intake valve(s) and into the intake manifold.
Past the closed exhaust valve(s) and out the exhaust.
Through a cracked cylinder head into an oil or coolant passageway.
Through a blown head gasket into an oil or coolant passageway.
Regarding listening and probing, what you do is take a section of hose, something like a garden hose a few feet long, put one end next to your ear, withdraw the dipstick, and place the other end of the hose near a crankcase vent such as the valve cap or the tube connecting to the PCV. Don’t use the oil dipstick tube if the engine has oil in it, as the oil will likely prevent air from escaping up the dipstick tube. If you hear air, then pressure is leaking past the rings, into the crankcase, and out the vent. In practice, you always hear some amount of air here, because especially in an old car, pressure is always leaking past the rings. What you want is for it to be about the same for all cylinders.
Next, place the hose at the intake manifold. In a carbureted car, you can put it at the top of the carb’s throat. In a fuel-injected car, you can put it in the throttle body. If you hear air escaping there, you have a leaky intake valve. Try to be as certain as you can that you’re hearing air from the intake manifold and not residual air that’s coming up from the crankcase. In a primitive car, there should be no connection. In a post-1970s car, there is usually a positive crankcase ventilation (PCV) valve that may connect to the air cleaner but not to the intake manifold. But in a more modern car, there may be small hoses that connect to electronically-controlled valves on the manifold itself.
Finally, go to the back of the car and put the hose in the tailpipe. If you hear air there, you probably have a leaky exhaust valve.
Now, if instead of hearing air, you hear bubbles, you cry. Case in point: I once did a compression test on a BMW 2002tii I was about to sell. Three cylinders had 150 psi compression, but to my stunned surprise, I found one cylinder had only 70. I did a leakdown test on the low cylinder, and immediately heard bubbling. I opened up the radiator cap to hear and see a steady stream of bubbles in the coolant. This pointed to either a blown head gasket or a cracked head. I pulled the head and the head gasket looked fine. I didn’t see any cracks in the head, but I took it to a machine shop, and when they tanked and cleaned it, two fine cracks became visible in the combustion chamber that had the low compression. Not a good day.
Now, in the above list, I say it can leak through a closed intake valve or exhaust valve, so there’s a crucial step here that I’ve left out: When you conduct the leakdown test on a cylinder, the valves in that cylinder must be closed, which means that piston in that cylinder MUST be at Top Dead Center (TDC), the top of its compression stroke. If it’s not, the valves may be open, and the test is useless. Thus, this simple-sounding leakdown test hinges on being able to rotate the engine so that, one cylinder at a time, the piston in that cylinder is at top dead center.
In this way, doing a leakdown test is a bit like adjusting valves—you set the engine at TDC for #1 cylinder, do what you need to do on that cylinder, then rotate the engine so that the next cylinder in the firing order is at TDC, do what you need to do there, rotate the engine again, etc.
So let’s walk through this again, concentrating on engine rotation. First, note that when pressure is introduced into a cylinder that’s at TDC, it can push the piston down and cause the engine to rotate. You don’t want the engine to rotate, since if it does, it moves the piston in the cylinder you’re trying to test out of position, which can open up a valve. So the test is best done with the drive wheels on the ground, the car in gear, the emergency brake engaged, and the wheels chocked.
1) Rotate the engine to TDC for #1 cylinder. All engines have a mark on the crankshaft pulley for TDC, and a pointer or marker on the engine block where that pulley mark should line up. If your car has an old-school distributor, you can trace the plug wire back from #1 cylinder to the distributor, see which terminal it is on the cap, pull off the cap, use a Sharpie to mark where that terminal is on the rim of the distributor (there’s often a small embossed mark there anyway), and rotate the engine until the rotor points to that mark. Note that the standard direction of engine rotation, with certain rare exceptions like Corvairs and early Hondas, is clockwise as viewed looking at the crankshaft pulley, and that most distributors, with certain rare exceptions, rotate clockwise.
2) With the engine at TDC for #1, screw the hose for the leakdown tester into the spark plug hole. Loosen the knob on the leakdown tester to evacuate any pressure already in it. Connect the hose to the tester. Tighten the knob to let air into the cylinder. If you’re really trying to get actual leakdown percentages, it’s best to pressurize to 100 psi as it makes the math easy, but if you’re doing the seat-of-the-pants listening thing, any pressure above, say, 50 psi is fine.
3) Shut off the compressor so you can hear air. Listen as I described above. Put one end of the rubber hose to your ear and move the other end from the crankcase vent tube, to the intake, to the exhaust, and then open the radiator and check it for bubbles.
4) Disconnect the tester from the hose, and unscrew the hose from the spark plug hole.
5) Rotate the next cylinder to TDC. This is the challenging part, as #1 cylinder is the only one with a TDC mark on the crankshaft. What you need to do is follow the firing order. It’s usually stamped on the valve cover or engine shroud. For example, if you have a four-cylinder engine, the firing order is almost certainly 1-3-4-2, so rotate the engine 90 degrees in the direction of rotation (again, for most cars except Corvairs and early Hondas, clockwise as viewed from the front of the crankshaft pulley) to put cylinder #3 at TDC.
5a) Method #1 of Rotating to Next TDC: If you have a four-cylinder engine, you can do a pretty good job of judging 90 degrees of rotation by simply looking at the rotor in the distributor. You can even mark it with a Sharpie on the rim of the distributor. Go straight across from the mark you made for where the rotor was at TDC, and that’s 180 degrees. Half the distance, and that’s 90. You can make four marks on the rim if you like. Rotate the engine CW (most cars), the distributor will rotate CW (most cars), and when the rotor is at 90 degrees after the TDC mark for #1, the next cylinder in the firing order should be at TDC.
5b) Method #2 of Rotating to Next TDC: If you don’t have a four or a six-cylinder engine, judging rotation by looking at the distributor is more difficult. If you can view the camshaft lobes and rocker shafts, there is a very accurate method of finding TDC of cylinders by looking at the valve overlap of their “companion cylinder,” but this requires taking the valve cover or covers off. Instead, there’s a very simple method: Take a thin wooden dowel, or a chopstick, or a piece of a coat hanger, put it down the spark plug hole, feel for the top of the piston, and have a friend rotate the engine in its running direction (again, almost always CW as viewed from the crankshaft pulley) as you directly feel for when the piston is at the top. Typically you’ll go past it and feel the piston go back down, then have your friend reverse direction and bring it back up.
So, with that tutorial, back to the Rob and Craig show. We conducted a compression test on Craig’s Corvair, and it revealed between 140 and 185 psi compression in five of the six cylinders, and cylinder #2, on the left bank, had zero. Did that mean #2 was the one that was missing a piece of its piston ring? Well, that’s what the leakdown test is for. At least in theory.
We first leakdown tested #2, rotating the engine through its 1-4-5-2-3-6 firing order to get there (and remembering that the Corvair is one of a handful of cars where the engine’s running direction is counter-clockwise, not clockwise), and using a short piece of a coat hanger to verify TDC. It showed a lot of leakage past the rings, but to our surprise, the hissing gun of zero compression was, instead, a massive amount of air into the intake manifold. Diagnosis: Sure looks like #2 has zero compression due to a bad intake valve.
We then circled back and leakdown-tested all cylinders, and found other cylinders that also appeared to have leaky valves. In particular, cylinder #5, the one in the right bank with 140 psi compression, clearly failed the hose-in-the-exhaust test.
So… did we definitively find out which piston the piece of ring in the oil pan had come from? Not really. But we did get Craig two crucial pieces of information. The first was that, with zero compression in one cylinder, any driving of the car is likely to result in catastrophic engine damage sooner rather than later. The second was that, if he addressed only the issue of zero compression in #2 cylinder, he’d likely be left with an engine with other leaky valves. At a minimum, he should pull both heads and have the valves done. If the heads come off, Craig can rotate the engine and look inside the cylinder bores. I’d expect that the piston that broke a ring would’ve left a mark on the cylinder walls.
The challenge with inspecting the pistons directly is that to do that, you need to unbolt the rod end caps from the crankshaft and slide the pistons out the front. Then, to get the pistons and rings back into the cylinder, you need to compress the rings. Once you’re doing these things, you’ve slid down the slippery slope from a valve job to at least some sort of an engine rebuild.
Not bad for two guys on a Wednesday night.
Rob Siegel has been writing the column The Hack Mechanic™ for BMW CCA Roundel Magazine for 30 years. His new book, Ran When Parked: How I Road-Tripped a Decade-Dead BMW 2002tii a Thousand Miles Back Home, and How You Can, Too, is available here on Amazon. In addition, he is the author of Memoirs of a Hack Mechanic and The Hack Mechanic™ Guide to European Automotive Electrical Systems. Both are available from Bentley Publishers and Amazon. Or you can order personally inscribed copies through Rob’s website: www.robsiegel.com.