How To Test Your Compression

[GTsRasta]

Compliments of Smaay:

http://www.newcelica.org/forums/showthread.php?t=227280

Originally Posted by Smaay

I have been seeing alot of questions about compression and how to test, so here is a writeup i did.

Compression Test by Smaay

1. Open hood



2. Remove engine plastics with 5mm Allen Wrench.



3. Plastics removed



4. Reach behind engine to unplug power steering sensor






5. Remove the four 10mm bolts holding coils.



6. Unplug coils



7. Remove the two 10mm nuts holding wiring harness.



8. Lift wiring harness out of the way.





9. Pull all four coils.





Coils removed exposing sparkplugs


9.5 unplug the 4 injectors so they dont squirt.



10. Using 5/8 spark plug socket and suitable extension (6-8 in.) remove and inspect sparkplugs. They should be a copper color like pictured.



Image of compression tester purchased at Autozone/Pep Boys/Kragen/Napa/etc.



11. Screw compression tester into cylinder 1 hand tight.





12. Hold down gas and clutch at the same time, crank engine over about 6 times, until needle stops jumping on compression tester

13. Record results.




14. Release air pressure from tester and repeat for each cylinder.





Straight from the Toyota Manual (BGB)

Compression pressure:
1ZZ-FE 1,500 kPa (15.3 kgf/cm2, 218 psi)
2ZZ-GE 1,400 kPa (14.3 kgf/cm2, 203 psi) or more
Minimum pressure:
1ZZ-FE 1,000 kPa (10.2 kgf/cm2, 145 psi)
2ZZ-GE 1,000 kPa (10.2 kgf/cm2, 145 psi)
Difference between each cylinder:
1ZZ-FE 100 kPa (1.0 kgf/cm2, 15 psi) or less
2ZZ-GE 110 kPa (1.1 kgf/cm2, 16 psi) or less

15. After test complete, put car back together in reverse order.

16. DONE! Have a beer to celebrate good results, or sulk in sorrows.


edit: this will throw a CEL for misfires. there is a fuse you can pull to prevent it but im not sure which one it is. but just unhook your negative terminal on the battery for 5 minutes and it will reset the CEL. Also the car might have a hard time starting, dont worry, just give it a few more cranks. It will start

Enjoy.

[jtj]

Im betting those coils have something to do with his CEL/misfire codes! Looks like they have been sitting in some water...

if you have a lab scope or graphing multimeter and an amp probe/clamp you can do an overall compression check in a few minutes without removing any parts. just graph the current ramp for the starter during cranking (pull fuse for fuel pump to prevent starting) with the clamp around the battery cable at the starter and sync the trigger off of number one cylinder .it wont tell you specific psi but all ramps will be even on a sound motor and if you have one ramp thats alot lower than the others, you can count in the firing order off of your sync to find which cylinder it is and do a real test on that one cyl. If your engine isnt running bad or other mechanical problems then this prob wont be much use to you, but it has saved me a ton of time on engines that are fairly hard to reach the plugs on and show signs of compression issues, since i didnt have to go looking at 11 other cylinders first (murphy's law will ensure the low cylinder is the last one checked, always!). This current ramp graphing method is also good if you have a low amp probe and need to check injectors when chasing rough running issues with no codes, since the electrical properties of coil windings and thier signature when open/closing/sticking are the same no matter who makes the injector.

[codymac]

Originally Posted by jtj

if you have a lab scope or graphing multimeter and an amp probe/clamp you can do an overall compression check in a few minutes without removing any parts.

Wow... you just changed my opinion that most modern tech.s are merely parts replacers.

[silvElise]

of course jtj is the man!!

[TimMullen]

One thing that was skipped in the instructions.

When you are cranking the engine, open the throttle fully. It does make a difference in the pressure readings that you will get, and it will come up to full pressure much quicker (with less drain on the battery).

As noted above, I'd pull the fuse for the injectors rather than disconnecting the wring for them.

Also, check the engine manual for the car that you are going to do a compression test on. On some of them (for instance the "wasted spark system used on early Miatas), cranking the engine with the spark plugs disconnected will often burn out the coils requiring them to be replaced. That shouldn't be a problem in this case since the individual coils are removed (and on the Miata the proper way to do the test is to disconnect the separate coils).

[pianomaniac]

Originally Posted by TimMullen

One thing that was skipped in the instructions.

When you are cranking the engine, open the throttle fully. It does make a difference in the pressure readings that you will get, and it will come up to full pressure much quicker (with less drain on the battery).

Tim, I believe he covered that in step #12.

"12. Hold down gas and clutch at the same time, crank engine over about 6 times, until needle stops jumping on compression tester"

[TimMullen]

Originally Posted by pianomaniac

Tim, I believe he covered that in step #12.

"12. Hold down gas and clutch at the same time, crank engine over about 6 times, until needle stops jumping on compression tester"

Doh...

I saw the hold down the clutch - missed the hold down the gas part.

Sorry. Nevermind...

[Dave]

The fuse is R7 BTW in the engine bay boxes, the only 10A. It will also disable the coils.

[jtj]

Originally Posted by codymac

Wow... you just changed my opinion that most modern tech.s are merely parts replacers.

Unfortunately many of them are. Its sad the number of people who claim to be techs that can only pull a code and guess at a part with no further diag involved, and actually make a living doing it that way. I stay busy with cars from other shops that still arent fixed. The common one is lean codes and kwik car did a injector flush, put an air filter, oxygen sensors and a mass air flow meter on it then sent them to the dealer because the code still returned (and they got enough of the customers money I guess). Each one you can hear the vacum leak with the hood shut while standing beside the car! Every time I go to factory training for Jag, I am amazed at the backyard bumkin who is there (every class has at least one) that thinks an obd2 code is all you need to fix a car. that and the fact he actually convinced a dealership to let him touch $100k cars is pretty amazing too...

Another quick test you can do with many cars (although I have never tried it with a Lotus) is a cylinder balance test with a graphing multimeter, scope or a scan tool that will show a pid graph for a component and freeze frame it.. I use this on jags to verify injector faults/repairs, but its prtetty generalized for any car that can be forced into open loop with a wide band o2. put the car into open loop, usually pulling the number 1 injector plug while it is running will do this on nippondenso based systems. while the dtc is set, the ecu will ignore 02 input for fuel map/injection and use a preset base line. the 02 will still read the ratio however because it doesnt know its being ignored. dfepending on the tool used you can get an actual reading or a voltage in the case of a scope backprobed into the sensors harness . then, while running the o2 voltage graph over enough time to see all cylinders on that bank in one screen shot, pull one injector plug at a time long enough to see a viewable waveform then plug it back in long enough to return to the engine baseline ratio before unplugging the next one. then you can compare exactly how much fuel each injector is dumping in relation to the others on that bank (especially easy when a "picture" of them all is laid side by side to compare.) wide bands are extremely accurate and this will show something like a dirty injector nozzle that otherwise might not show in other tests and without dissasembling the injector rail etc. it can also verify faulty 02 sensors if you have one thats possibly stuck/dead/lazy and want to see an actual value change at the sensor, although a scope is better for that due to refresh rates etc.

[andykeck]

Originally Posted by jtj

Im betting those coils have something to do with his CEL/misfire codes! Looks like they have been sitting in some water...

No kidding! Those look nearly as bad as mine finally got when I was running on two cylinders.

[Devious]

Maybe I missed it, but I didn't see battery voltage and engine temp mentioned. Both can alter your readings slightly.

Leak down tests will also help diagnose issues that compression tests may miss.

[ZAMMY1]

RASTA: WELL DONE!
JTJ: GOOD TIPS ON DIAGNOSIS!

auto techs are no longer 'greasemonkeys'.
many are highly trained and capable.
when it gets down and dirty on a subtle problem there's no substitute for training.
sam

[jtj]

Originally Posted by jtj

Another quick test you can do with many cars (although I have never tried it with a Lotus) is a cylinder balance test with a graphing multimeter, scope or a scan tool that will show a pid graph for a component and freeze frame it.. I use this on jags to verify injector faults/repairs, but its prtetty generalized for any car that can be forced into open loop with a wide band o2. put the car into open loop, usually pulling the number 1 injector plug while it is running will do this on nippondenso based systems. while the dtc is set, the ecu will ignore 02 input for fuel map/injection and use a preset base line. the 02 will still read the ratio however because it doesnt know its being ignored. dfepending on the tool used you can get an actual reading or a voltage in the case of a scope backprobed into the sensors harness . then, while running the o2 voltage graph over enough time to see all cylinders on that bank in one screen shot, pull one injector plug at a time long enough to see a viewable waveform then plug it back in long enough to return to the engine baseline ratio before unplugging the next one. then you can compare exactly how much fuel each injector is dumping in relation to the others on that bank (especially easy when a "picture" of them all is laid side by side to compare.) wide bands are extremely accurate and this will show something like a dirty injector nozzle that otherwise might not show in other tests and without dissasembling the injector rail etc. it can also verify faulty 02 sensors if you have one thats possibly stuck/dead/lazy and want to see an actual value change at the sensor, although a scope is better for that due to refresh rates etc.

i actually had a chance to try this out on a lotus the other day, and unless there is another way to force it into open loop for the duration of that key cycle then this will only work on a scope that can show a time display of 5 minutes or so if your a slow mover/no assistant.you could do it one frame at a time and print the waveform/pattern for each cylinder and compare later if your using a lower end scope/meter, but i would prefer them to be on the same screenshot for better comparison. unplugging an injector will take about 20 to 30 secs to put car in open loop, plugging it back in takes less than 10secs for ecu to return to closed loop (based on lotus scan data anyway, although its prob more instant and the tools refresh rate is slowing it down).its doable but not really as easy or practical as it is on other cars. just thought i would let you all know...

[Dmbrown34]

Originally Posted by jtj

Im betting those coils have something to do with his CEL/misfire codes! Looks like they have been sitting in some water...

if you have a lab scope or graphing multimeter and an amp probe/clamp you can do an overall compression check in a few minutes without removing any parts. just graph the current ramp for the starter during cranking (pull fuse for fuel pump to prevent starting) with the clamp around the battery cable at the starter and sync the trigger off of number one cylinder .it wont tell you specific psi but all ramps will be even on a sound motor and if you have one ramp thats alot lower than the others, you can count in the firing order off of your sync to find which cylinder it is and do a real test on that one cyl. If your engine isnt running bad or other mechanical problems then this prob wont be much use to you, but it has saved me a ton of time on engines that are fairly hard to reach the plugs on and show signs of compression issues, since i didnt have to go looking at 11 other cylinders first (murphy's law will ensure the low cylinder is the last one checked, always!). This current ramp graphing method is also good if you have a low amp probe and need to check injectors when chasing rough running issues with no codes, since the electrical properties of coil windings and thier signature when open/closing/sticking are the same no matter who makes the injector.

While this is good stuff...a graphing multimeter runs over 1500 bucks...

[darkSol]

TurboPhil and I just did a compression test on my engine...

140, 140, 140, 140 - like clockwork.

Funny thing is that it pulls like a freight train, doesn't burn any oil and the plugs looked fine (well a bit of detonation but no blown off tips or anything like that).

WTF!

[Dave]

Originally Posted by darkSol

TurboPhil and I just did a compression test on my engine...

140, 140, 140, 140 - like clockwork.

Funny thing is that it pulls like a freight train, doesn't burn any oil and the plugs looked fine (well a bit of detonation but no blown off tips or anything like that).

WTF!

140 LBS? Try a different gauge, seems too coincidental for it to be that low and so linear on all 4.

[turbophil]

I thought maybe the gauge was off too, so I hooked the gauge to my air compressor. The compression tester agreed within a psi of the pressure I had in the air compressor leading us to believe the tester is accurate... Perhaps we're over-looking something here, or perhaps Dark's compression rings are a bit toasty from running so lean for so long If the latter is the case, it's fortunate forced induction tends to mask low compression As long as there' no oil blowing, we just need to up the boost--- once we can tune for it

Best,

Phil

[TimMullen]

Did you hold the throttle wide open while you cranked the engine? If you don't, the engine can't pull in enough air, and the compression readings will be low.

Is your car a "Throttle By Wire?" If so, does the ECU keep the throttle closed under the cranking conditions?

The fact that all the cylinders are reading consistent usually indicates that things are well, and the readings are low for some reason.

[turbophil]

Originally Posted by TimMullen

Did you hold the throttle wide open while you cranked the engine? If you don't, the engine can't pull in enough air, and the compression readings will be low.

Is your car a "Throttle By Wire?" If so, does the ECU keep the throttle closed under the cranking conditions?

The fact that all the cylinders are reading consistent usually indicates that things are well, and the readings are low for some reason.

05 car and yes, the throttle was wide open... Tried both ways, thrttle closed and open- readings were the same... which is intuitive since all the other plugs were out and letting air into the plenum---An open spark plug hole creates a huge vacuum leak..

Best,

phil

[jim-clayton]

My Milton has a value screwed into the end of the hose. It has to be a special valve. A normal tire valve will give you low readings.

could be, no way it's 140