The previous article on compression ratio is relevant here and should ideally be read first if you are in any doubt as to what compression ratio is exactly. When the piston rises from BDC to TDC on the compression stroke it compresses the air in the cylinder and raises its pressure. The higher the compression ratio of the engine the greater the pressure at TDC. If the rings, valves and head gasket are sealing properly we can expect a particular engine to generate a specific pressure in the cylinder at TDC. By connecting a pressure tester into the spark plug hole and cranking the engine over we can measure this pressure. If the pressure is lower than expected it shows that there is an internal problem with the engine.
At cranking speed there is plenty of time for air to enter the cylinder. We can expect that when the piston reaches BDC the cylinder will be full of air at atmospheric pressure - which is approximately 14.7 psi. As the piston rises, this trapped air will be compressed and the pressure in the cylinder rises. If the engine has a compression ratio of 10 to 1 then all the trapped air will be compressed into one tenth of its initial volume. We might expect that the final pressure at TDC would then be 147 psi. This ignores a number of factors about how both gases and engines work.
When a gas is compressed its temperature rises - this is a basic fact we are told by the Universal Gas Laws. This increase in temperature leads to a further increase in pressure over and above what we would expect from consideration of the compression ratio alone. If the compression ratio is high enough, the temperature generated in the cylinder can ignite the fuel without any need for a separate spark. This is how Diesel engines work. In an average petrol engine the effect of temperature is to generate about twice the pressure we would otherwise expect. In that case maybe we should expect an engine with a 10:1 compression ratio to generate about 300 psi at TDC instead of 147 psi? There are more factors to consider though.
Inlet valves do not close at BDC but well after this, when the piston is already part of the way up the bore. This is done to help trap more air when the engine is turning at normal running speeds. There is not enough time at running speed for the air already drawn into the cylinder to turn round and go back out of the inlet valve before it shuts so more air gets trapped than if the valve closed at BDC. At cranking speed of only a few hundred RPMs though, everything is happening very slowly in comparison. There is plenty of time for air drawn in at BDC to get pushed back out of the valve before it shuts. The static compression ratio is therefore not a good measure of how much air is actually being trapped and compressed. The longer the cam duration and therefore the later after BDC the inlet valve closes the less air gets trapped at cranking speed. For a given compression ratio we can therefore expect engines with race or rally cams to generate lower cranking pressures than engines with road cams.
The gauge and the tube leading to it also have some internal volume. In effect by screwing the gauge into the spark plug hole we are increasing the volume of the combustion chamber by several CCs. This reduces the compression ratio below the static measured figure and further reduces the pressure that the engine will generate during the compression test.
For an average road engine we don't need to calculate the effect of all the above factors because the expected cranking pressure will be quoted by the manufacturer. Haynes manuals and other tuning guides also quote the figures. If the data is not available you can make a good estimate of the expected cranking pressure from a healthy engine just from the compression ratio. You can expect the cranking pressure in psi for a road engine with a standard cam to be about 17 to 20 times the value of the compression ratio. So our engine with a CR of 10:1 should generate between 170 and 200 psi. The better the condition of the engine the closer the figure will be towards the top end of this range. If the engine has been modified by fitting a different cam then you need to make some allowances. Fast road cams will reduce the expected figure by 10 psi or so. The best thing to do is take a set of compression readings as soon as the engine has been built so that you know what it generates in good condition. For competition engines it's even more important and I recommend to all my customers that they do a compression test after every race to make sure that nothing inside the engine is going wrong. The vast majority of them take no notice of course because it involves a bit of work. Finding a potential problem before it leads to an engine blow up can save a lot of time and money in the long run though.
The cranking pressure can also be used as guide to setting the compression ratio to suit competition camshafts. The longer the cam duration the higher the CR needs to be to make it work properly. As explained above, the later inlet valve closing of the long duration cam will reduce the cranking pressure but when you raise the CR this compensates and the cranking pressure comes up again. My rule of thumb for normally aspirated competition engines is I don't like to see less than 200 psi cranking pressure regardless of cam duration if the CR has been matched to the cam correctly. Anything less than 180 psi and the cam won't be working at its best, especially at low rpm.
An equation I use for estimating cranking pressure from static compression ratio (CR) for engines in good condition with standard road cams is as follows.
Cranking pressure (psi) = CR^1.3 x 14.7 x 2/3
CR^1.3 is the Compression Ratio raised to the power 1.3. On pocket calculators you can do this with the x to the y function.
For modified engines, or more specifically engines with long duration cams you need to then apply the following approximate factors.
Fast road cam - multiply by 0.95
Rally cam - multiply by 0.90
Full race cam - multiply by 0.85
Example. A rally engine with 11:1 static compression ratio.
11^1.3 = 22.58
22.58 x 14.7 x 2/3 x 0.90 = 199 psi
With an engine in good condition you should expect cranking figures similar to those in the table below. Obviously some of the boxes are never going to be encountered in practise i.e. a 7:1 race engine or a 13:1 standard one but I'll leave them in there for the sake of completeness.
Some cheap gauges just have a rubber bung at the end which you are meant to push against the top of the spark plug hole by hand. Avoid these like the plague. Without a remote starter switch you can't both do this and reach the ignition key at the same time - they also don't always seal properly. Most gauges come with a flexible hose about a foot long and a threaded end which screws into the plug hole. There will also be adaptors to suit different plug threads. Any gauge of this type should suffice although expensive ones like Snap-On can be expected to be more accurate and longer lasting. It isn't a tool you use often though and should last a lifetime in normal use anyway. Gunson do a perfectly adequate gauge which you can get from Halfords or tool suppliers.
Not all gauges are marked in psi - some read in Bars. It is commonly thought that 1 Bar is the same as normal atmospheric pressure of 14.7 psi but this isn't true. A Bar is 100,000 Newtons per square metre which equates to 14.5 psi. I prefer gauges reading in psi but either type will do as long as you can translate the reading into the expected pressure.
If possible i.e. if the engine still runs then run it until it reaches normal operating temperature. Tests done on a cold engine usually show slightly lower readings, perhaps 10 psi lower depending on wear although it should not affect a proper diagnosis and all cylinders ought at least to be affected equally so it's no big deal. Remove the HT leads and take out all the spark plugs. The ignition system MUST then be disabled. If this isn't done it will continue to generate high tension voltages into the HT leads which will have nowhere to go with the plugs out. These high voltages will find another route to earth and can damage the ignition system or even the car's ECU. Unplug the low tension connections to the coil or to the distributor. If you aren't sure how to disable the ignition system on your own car then check with a dealer before you start. It is also good practise to unplug the fuel injectors or disable the fuel pump, especially on cars fitted with a catalytic convertor. This prevents unburned fuel getting into the exhaust system during the test and contaminating the cat.
Screw the gauge into cylinder 1 and rest it somewhere you can see the dial while you crank the engine. Open the throttle fully either by pressing the accelerator or wedging the linkage open under the bonnet. If the throttle isn't open then air can't get into the cylinders in the first place and the readings will obviously be far too low and pointless to evaluate anything properly with. Crank the engine until the gauge stops rising and count the revolutions while you do so. It should normally take no more than 10 engine revolutions (5 compression cycles) to get a full reading. You can count the cycles by watching the gauge too - each jump of the needle is one compression stroke. Write down the final reading and also make a mental note of how quickly the gauge rose on the first few cycles. Then just repeat for the other cylinders. Make sure that each cylinder reaches its highest reading after the same number of engine revolutions. If all readings are good then the test can end there. If any cylinders are low then a "wet" test can be done. This involves squirting a few ccs of oil , no more than a teaspoon full, into the cylinder and repeating the test. The oil will help seal bad rings and increase the reading but won't affect it if the problem lies in the valves or head gasket.
An engine in good condition should have readings within the specified range (preferably at the upper end of it) and with all cylinders within 10 % of each other. A perfect engine might have almost identical readings on all cylinders to within even just 1 or 2 psi - it is certainly possible to achieve this on a really well blueprinted competition engine. A good cylinder will reach about 2/3 of its final reading on the first compression cycle and reach the full reading after only 3 to 5 cycles. If the rings are worn you often see a gauge rising in smaller jumps of 20 to 30 psi per cycle rather than one big initial jump and also taking more revolutions to reach a peak reading.
If all cylinders show similar psi but are below the minimum figure then this usually indicates excessive ring and bore wear due to high mileage. A slipped cambelt can give a similar result though. Very low readings can be caused by bent valves.
Very low but similar readings on two adjacent cylinders is often caused by a head gasket which has blown between those two cylinders.
One cylinder low means a bit more detective work. If the wet test improved things back to a normal reading then the problem lies in the rings or bores. If not then its usually either valves or gasket not sealing properly. To an extent I fail to see the point in worrying overmuch about exactly where the problem lies. Most of the time the cylinder head is going to have to come off so you might as well do that first and see what shows up. It may of course not be just one problem though and there is no point in spotting the bent valve and refitting the head but failing to realise that there was a broken ring as well. A leak down test can help here.
One problem that might not lie inside the cylinder itself is a worn camshaft. If a lobe is badly enough rounded off then that valve won't open fully, or at all, and if no air can get into the cylinder then of course it's not going to generate much of a cranking pressure. The CVH engine is rather prone to this fault because of its recognised cam wear problem - especially on cylinder number 4. A collapsed hydraulic lifter that isn't pumping up properly can also lead to a low reading - or very occasionally an engine with solid lifters can spit out a valve adjustment shim - the Peugeot 205 8 valve engine is prone to this if over revved but you'll probably work out the reason anyway from the consequent appalling rattle.
This involves pumping high pressure air into the cylinder and noting how quickly it leaks out on a gauge. The leakage rate should be very low if the rings etc are in good condition. Most garages have leak down testers and it might be worth having this done if the compression test shows a problem but you aren't sure where it lies. During the leak down test you can also listen for where the air is escaping and this can pinpoint the problem to a valve, gasket or ring fault. If the leakdown test shows no fault but the cranking pressures are low then it must be a cam lobe or cam timing type of issue.
Doing a compression test as a first step whenever there's an engine running fault can save a lot of time in the long run. Plenty of people have struggled in vain to find an ignition timing or carb fault to try and cure a rough idle or lack of power when there was nothing wrong with those items in the first place. Nowadays with complex engine management systems and electronic everything, it's easy to fall into the trap of assuming an engine fault must be due to something horrendously complicated in a little black box somewhere. As often as not the answer is much more simple than that. Mechanical things go wrong much more often then electronic ones after all. The motto is "when in doubt, check the basics first and save the panicking for later".
Back to main menu page