Main.TrailIndexPage | Engine | Compression and compression testing

Compression and compression testing

This section is part of the Engine section.

The following is the content from the old Yahoo documents on the site. It needs to be structured and edited in the correct sections of the entire document. After moving particular content to its correct place in the manual, please delete it here.

What would the proper compression be on my '66 230 SL? I have found it in my service manual, but it's metric and I don't know the conversion.

Frank Mallory says: I think they use atmospheres, which are approximately 15 psi. The absolute reading isn't very important: it will be lower in mile-high Denver than here in sea-level Maryland; also, many gauges are inaccurate. The important thing is uniformity of readings from cylinder to cylinder and from time to time, under the same conditions with the same gauge.

Will Samples says: The metric unit of measure for pressure is 1 kg/cm (with the cm squared but I do not know how to do it on the computer) and this is equal to 14.22 psi. Normal compression for 230 SL is 11-12 kg/cm or 156-170 psi. Minimum is 9 kg/cm or 128 psi. Measure compression with normal operating temps (70-80 C) and open throttle. Crank engine at least 8 times. Deviations between cylinders should not be greater than 1.5 kg/cm or 21 psi. The above is from the Technical Data Manual for 1966. My own thoughts are that a deviation of 21 psi between cylinders is too much. I was taught 10%. Compression above 170 psi is probably due to carbon buildup in the cylinder. You did not ask but Cylinder Leakage was listed as well. Permissible loss is 25%.

Will and Frank, thanks for your timely answers. As I recall (8 months ago) my compression was in the 130-135 psi range across all cylinders but one which was about 120 psi or so. This is in Ohio at about 800 ft elevation. Is low compression a critical problem? Will it cause other problems? What does it indicate as weak points in the engine? I hear lots of talk about "leak down" tests. What are they in comparison to a compression test?

Will says: You need 3 things to make a gasoline engine work, compression, fuel, ignition. Without enough compression the engine will not function. Low compression causes you to crank the engine longer when trying to start it. As you crank the engine it slowly builds pressure inside and will eventually fire up. I cannot think of other problems it will lead to because once you have lost compression, you will not have a working engine, so nothing else can break. What does it indicate as weak points in the engine and how is the leak down test related? There are various possibilities for low compression and the leak down test pinpoints them fairly well. Your engine is an airpump and the leak down test mesures the engine's ability to hold air. The combustion chamber has a theoretical limit of holding 100% of the air/fuel mixture that is drawn into it and compressed prior to being ignited by the spark plug. The leak down test measures with a flow meter the actual amount of air the combustion chamber will hold. Weak points will be noted by the sound of rushing air. So, if the rings on the pistons are weak, listening at the oil filler cap while doing the leak down test will allow you to hear the sound of escaping air. A bad exhaust valve will allow air to rush out the tail pipe so you stick your ear to the tail pipe and you hear the noise of air passing by a bad exhaust valve. The compression test uses a gauge to measure the build up of pressure inside each cylinder as you crank the engine. The cylinder leakage test uses compressed air that is forced into the combustion chamber. Both testers are screwed into the sparkplug hole.

To check the condition of the rings, ask your mechanic to do a leak down test. This test measures the ability of the cylinders to maintain compression, thus testing the integrity of the rings and valves. The cylinders with low compression MAY have poor rings, but check the valves first. I had the valves done on mine, hoping that would cure the problem, but it was terminal; 3 of the cylinders had more than 50% leakage cold and about 70% hot - bad rings! I had a complete overhaul done. Now it runs like it just came from the factory (better, says my mechanic). If it isn't using oil, the rings are probably good for many more miles. Just as a point of reference, a complete overhaul -including rebuilding of the injector pump- should cost about $5 - 7.5K.

I think this has been explained before, but I'm not exactly clear on it. You say a leakdown test checks the rings, but then you say it tests the rings and valves. What exactly does a compression test tell you and what exactly does a leak down test tell you? I know you need to use both to get complete information.

A compression test tells you the compression provided by the piston on the compression stroke. It is a relatively instantaneous measurement, but if you keep the meter on the port you can see how the compression decreases over time. If it decreases too quickly, something's not right; either leaky valves or rings. The leak down test quantifies the compression fall-off as a percentage over time and (usually) at cold and hot engine temperatures. It is a better indication of engine condition and is sometimes used to predict remaining engine "life." If your rings are bad, you'll usually see two things:

  1. low compression
  2. high (or higher than usual) oil comsumption

You may also see white smoke from the exhaust in extreme cases. If there is no smoke and the compression tests are good, then the engine should be fine. If you still have doubts, check the valve guides and valves (if you see a puff of white smoke when you leave a stop light - after decelerating) it could be valve guides or seals. A valve job is easier and MUCH cheaper than a ring job. A skilled mechanic can usually determine if it's the valves or rings. One more thing. The 2.8 litre engine consumes oil normally. If you check the owner's manual (PN: 113 584 15 96), page 26, a new, "broken-in" engine will consume from 0.15 to 0.25 litre per 100 kilometers (66 miles) or 0.6 to 1 litre per 400 kilometers (about 250 miles - 1 tankfull of gas). If you're within this range, the engine is operating within MB specs. My newly overhauled engine is on the low end of this range, but if I drive for a long distance, at high speed and in South Texas weather (100 degrees +) it will consume more.

These engines actually inject a measured amount of engine lubricating oil, through a small oil line with an orifice, into the fuel manifold at the back of the fuel injector pump. Gasoline, unlike diesel, is too dry to properly lubricate the Bosch fuel injector plungers, so the oil added to the gasoline accomplishes this. This is why these engines have a 2 quart low mark on the dip stick. It's not uncommon to see white exhaust smoke on starting, when the mixture is rich.

In a car this age, anything above 120 psi is probably okay - but more importantly is the variance. You should not get any more than 10% variance between cylinders. I'd start by keeping things simple. First, check all valve clearances, you can lose compression through the valves. Once this is done, check the compression again, and do it a few times to be sure it's not your compression tester that is buggered. There is a procedure for testing valve problems, which is also a possibility here. Pour some oil down the spark plug holes just before you take the reading. If the reading changes, then you've sealed the problem with the oil (i.e. your rings are gone), but if it doesn't then it's more likely to be your valves - which means a rebuilt head (much cheaper than the alternative).

Dan says: good compression is about 165 psi. Anything under 130 is actually getting low and showing some wear. Since the bottom end usually wears at about the same general rate smaller variations are leaking valves in most cases. Two low ones side by side is a burned head gasket. Really big ones are burned valves or broken rings which require a rebuild. When you get down into the 100 to 110 range she's all worn out. Sorry.

All 3 SL engines had about the same compression ratio of 9.5 or so. How did the successively larger engines maintain this ratio? For example, from the 230 to the 250, did the 250 have larger combustion chambers to compensate for the increased displacement, or were different shaped pistons used?

Pete Lesler: all of these engines used flat top pistons. The compression ratio can be altered by changing the volume of the combustion chamber in the head as well. Or by modifying the height of the piston at top dead center, making the combustion chamber larger. I believe it was accomplished using the former method, not the latter.

Dan Caron on a ‘load test’: that would be where you keep your foot on the brake and apply some throttle at the same time. Bad rod bearings should rattle like a pea in a tin cup. Not very scientific but it works.

A load test is when you place the car in gear and place your foot on the brake pedal slowly letting out the clutch just as if you are going to stall the engine (but don't let it die give it more gas but don't slip the clutch too much. what you are trying do is to place a large load on the motor so as instigate a main bearing knock if you have one. this is a test that takes three feet in a standard but can be done with two. What you will hear if you have a problem is a faint to loud knock as the load is increased if this is the case a total rebuild is your best option. if you don't hear anything then that is a great sign of the condition of the bearings around the crankshaft. It is very easy to do on an automatic.

The real trick on compression is to know how to read it. 1st stroke pressure and then total pressure, all taken with the throttle wide open. The first stroke should all be within 10% and you can see the variance is about 8% for the total compression. First stroke tells you about the rings and total tells you about the valves. If the valves are bad you then check with the vacuum meter to confirm. Here you are looking for fluctuation from a normal, I don't care what the number is, only the way the needle on the gauge acts. Or you can use the piece of paper trick at the exhaust pipe to check for a bad exhaust valve but not for a bad intake valve.

Dan Caron: 175 psi would be about optimum - not much better than this unless there was a lot of deposits in the cylinders and then I'd think that the wear that allowed these deposits would lower the reading anyway. Leak down would be low , probably under 25% and maybe as low as 10%. A CO of 2% is actually pretty good. Why would you think that a low reading is a bad thing? Leave it alone. These cars often run rich and on a rebuilt pump it tends to be a case of not knowing about how they were originally done. It seems the testing oil was of a different viscosity back then and if set up using modern test oil ( you have to really) the pump will run rich on the fuels used today. The trick is to set them on the lean side of the rebuild test sheets to compensate for this.

Static compression test how-to

(There is also an engine running dynamic compression test you can do):

  1. Warm up the engine
  2. Take out all the spark plugs
  3. Disconnect the power to the coil/distributor to prevent spark
  4. Remove the fuse for the fuel pump (#4) to shut off any fuel going to the CSV. You don't want fuel spraying out of there while doing compression testing or any other operation where you are spinning the engine over without any ignition. All of that fuel will eventually wash down the cylinder walls and you won't have a very accurate compression test. If the CSV happens to be leaking you will really have a lot of fuel in the cylinders.
  5. Take off the air cleaner/filter
  6. I hook up the battery charger to lessen any variance in cranking speed
  7. Have a second turn the engine over at WOT
  8. Record three major items for each cylinder multiple times (the average will tend toward the norm)
    1. Pressure on the 1st compression stroke
    2. Number of compression strokes to maximum pressure
    3. Maximum compression pressure

When we do a compression test we can obtain five to ten data points on each cylinder. A data point is generated each time the engine goes through a compression stroke, and the total number of strokes to reach maximum compression. I use all of the data I can record because there is more information there. Most folks only use the final data point, i.e. total compression. The assumed constants are engine temperature, engine cranking speed, air temperature, air resistance.

Variables on each cylinder are:

  • valves do the seal
  • rings do they seal
  • cam does it let the engine breath

So what does the data gathered during the compression test tell us.

  1. The compression on the 1st stroke shows us how well the cylinder is breathing and should be within about 10% for each cylinder. A low cylinder(s) is not breathing correctly, a function of the valve seal, valve clearance or cam wear.
  2. The number of strokes to maximum compression tells us how well the engine is breathing (assuming the compression is within 10%) should be plus or minus one. If not then there is a breathing problem, a function of valve clearance and cam wear.
  3. Maximum compression shows the sealing properties of the rings and valve seals. A squirt or two from the oil can in a cylinder, let the engine set a while can then eliminate the ring variable.

If A or B first check the valve adjustment and re-test. If the symptom persists then check the cam for ware. The cam is most likely not lifting the valve enough to have good air flow but enough so that the total compression pressure is ok. Valve seats should be good and only the cam needs to be taken care of.

If C and you have eliminated the rings, i.e. you have put oil in the cylinder the the compression is still low, use a vacuum gauge or do the exhaust pipe paper test to look for a burnt valve. If the paper is sucked into the pipe and then blows back out, then the valves are burnt for sure as the sucking is the engine sucking back by the exhaust valve.

This along with a good vacuum gauge can tell you about anything you want to know about the health of your engine. When you buy a car, do a base line and compare data at each turn-up. This will help you understand what is going on in your engine. Hope this helps. I did this quickly without the help of my references (most are in storage in the US while I am in Germany) so I may have left something out. Please add to it and make it better.
< Overhauling – things to know | Main.TrailIndexPage | Performance Modifications >