Hi everyone,
My 84 944 is out of commission at the moment, and I recently did compression and leak down tests on it to try to assess the health of the engine. For the last 6 months or so when it was driving, it was running very rich with no clear indication as to why. It has rebuilt injectors and all new vacuum lines, and a fresh O2 sensor. The DME was also replaced with a rebuilt unit six or seven years ago. When the old DME was failing it was also running rich, to the point that the cat got red hot a few times.
I'm hoping that you guys might be able to give me some insight as to what's going on with this engine.
Compression test results by cylinder:
1: 170psi
2: 195psi
3: 142psi
4: 168psi
Adding oil to the cylinders increased the compression values to 195, 217, 215, and 210psi, respectively.
Leakdown, input vs cylinder pressure:
1: 90 to 88psi, with a decent bit of air coming out of the oil filler
2: 90 to 90psi, a small amount of air coming out of the oil filler
3: Initially 90 to 80psi, with a lot of air coming out of the exhaust, then I moved the crank very slightly, hoping that it wasn't at tdc. That resulted in 90 to 15psi, with a ton of air coming out of cylinder 4. There's obviously a problem here (cracked block?). I moved it back and forth a few times, getting the same result.
4: 90 to 86psi, air coming out of the oil filler
A few years ago the head gasket disintegrated and dropped a bunch of gasket material into cylinder 1, and it’s got some scoring on the cylinder wall. Cylinders 2 and 3 are also scored, but not quite as badly. I haven’t removed the head to feel how bad the scoring is, but it doesn’t look great. When that happened I had the head completely rebuilt with fresh guides and a new valve to replace a bent valve on cylinder 1.
I know that some leakage past the rings is to be expected, and I assume the scoring isn’t helping that. Based on what I’ve read, the leak down seems to be surprisingly little, other than in cylinder 3 of course.
Is it likely that the block is cracked, letting air into cylinder 4 from 3? Why wouldn’t the air go the other way, from 4 into 3?
Any help would be greatly appreciated.
Compression/Leakdown Question
-
Tom
- Site Admin
- Posts: 8928
- Joined: Fri Jun 25, 2021 2:04 pm
- Location: Silicon Valley, CA
- Has thanked: 932 times
- Been thanked: 3994 times
- Contact:
Curious is you ever checked to see if the Catalytic Converter might be clogged up? That can make the engine run rich, and make the cat overheat like you were seeing. Kind of a chicken and egg thing -- is the cat running hot because the car is running rich, or is the car running rich because the cat is clogged/overheating and not flowing. Both are possible... Was the car down on power when it ran?
On the leak down, how sure are you that #3 was at TDC on the compression stroke when you ran the test? If it was off a bit and valves were open on both #3 and #4, that might explain the pressure loss and cross-flow; and why it didn't flow back from 4 to 3 when you tested 4. Anything's possible, but that seems more likely than a cracked block, especially if it was running. Bad head gasket or valves also come to mind, but also seems unlikely given the history and symptoms.
Hard to assess how bad those marks are in the cylinders. I want to say I've seen worse, though it's hard to get perspective on those shots. I'd just say these motors can run pretty well with a surprising amount of bore score...
On the leak down, how sure are you that #3 was at TDC on the compression stroke when you ran the test? If it was off a bit and valves were open on both #3 and #4, that might explain the pressure loss and cross-flow; and why it didn't flow back from 4 to 3 when you tested 4. Anything's possible, but that seems more likely than a cracked block, especially if it was running. Bad head gasket or valves also come to mind, but also seems unlikely given the history and symptoms.
Hard to assess how bad those marks are in the cylinders. I want to say I've seen worse, though it's hard to get perspective on those shots. I'd just say these motors can run pretty well with a surprising amount of bore score...
-
dr bob
- Moderator
- Posts: 622
- Joined: Thu Jul 08, 2021 9:30 pm
- Location: Central Oregon
- Has thanked: 249 times
- Been thanked: 245 times
Following a little on Tom's points --
I generally use leakdown on an engine that can't be run to temperature for a compression test. Else a compression test is preferred.
-- Both tests rely on correct cam timing. Less important for leakdown, but the cam timing affects cylinder filling a lot so will affect the compression readings.
-- For leakdown the cylinder under test must be at TDC on the compression stroke. There's some room without risking the test, but not a lot or the air in the cylinder will roll the crankshaft. If I hear valve leakage, I look hard at whether I'm at TDC on compression, as TDC on end of exhaust and start of intake has both valves open some as flow functions overlap. That includes flow sounds to other cylinders when valves are open on both. It's easy to lose track. Piston at bottom may offer more cylinder leakage besides the valve-open condition.
-- For both tests, all plugs are out. That lets you listen in adjacent cylinders for cross leaks under leakdown, and reduces the starter load while spinning the crank for the compression test.
-- Compression test has engine at operating temperature, testing performed with throttle held wide open, both injection and ignition disabled (pull fuses is usually enough).
Oil in the rings will help seal leakage at the cylinder wall. More valuable during compression testing than leakdown, as the lower cylinder pressure for leakdown testing allow oil to mask bigger leakage issues. Listening at the oil filler tells you what you need to know about cylinder leakage without adding any extra oil.
I generally use leakdown on an engine that can't be run to temperature for a compression test. Else a compression test is preferred.
-- Both tests rely on correct cam timing. Less important for leakdown, but the cam timing affects cylinder filling a lot so will affect the compression readings.
-- For leakdown the cylinder under test must be at TDC on the compression stroke. There's some room without risking the test, but not a lot or the air in the cylinder will roll the crankshaft. If I hear valve leakage, I look hard at whether I'm at TDC on compression, as TDC on end of exhaust and start of intake has both valves open some as flow functions overlap. That includes flow sounds to other cylinders when valves are open on both. It's easy to lose track. Piston at bottom may offer more cylinder leakage besides the valve-open condition.
-- For both tests, all plugs are out. That lets you listen in adjacent cylinders for cross leaks under leakdown, and reduces the starter load while spinning the crank for the compression test.
-- Compression test has engine at operating temperature, testing performed with throttle held wide open, both injection and ignition disabled (pull fuses is usually enough).
Oil in the rings will help seal leakage at the cylinder wall. More valuable during compression testing than leakdown, as the lower cylinder pressure for leakdown testing allow oil to mask bigger leakage issues. Listening at the oil filler tells you what you need to know about cylinder leakage without adding any extra oil.
dr bob
1989 928 S4, black with cashmere/black inside
SoCal 928 Group Cofounder
928 Owner's Club Charter Member
Former Ex Bend Yacht Club Commodore Emeritus
Free Advice and Commentary. Use At Your Own Risk!
1989 928 S4, black with cashmere/black inside
SoCal 928 Group Cofounder
928 Owner's Club Charter Member
Former Ex Bend Yacht Club Commodore Emeritus
Free Advice and Commentary. Use At Your Own Risk!
-
dr bob
- Moderator
- Posts: 622
- Joined: Thu Jul 08, 2021 9:30 pm
- Location: Central Oregon
- Has thanked: 249 times
- Been thanked: 245 times
FWIW, I've grown into the habit of connecting a separate starter button at the starter solenoid for compression checks. Key stays off. I can crank the starter from the engine bay and hold the gauge, listen for noises, making this a single-person exercise. With key off, no ignition or fuel.
dr bob
1989 928 S4, black with cashmere/black inside
SoCal 928 Group Cofounder
928 Owner's Club Charter Member
Former Ex Bend Yacht Club Commodore Emeritus
Free Advice and Commentary. Use At Your Own Risk!
1989 928 S4, black with cashmere/black inside
SoCal 928 Group Cofounder
928 Owner's Club Charter Member
Former Ex Bend Yacht Club Commodore Emeritus
Free Advice and Commentary. Use At Your Own Risk!
@Tom - I had thought about the cat being clogged, maybe it would be worth running the engine with just the headers to see if that makes a difference. Honestly I don't know if it was down on power, but it never felt particularly strong.
I didn't even consider if each cylinder was on the compression or exhaust stroke, but that makes a ton of sense. I may have just gotten lucky with the other cylinders. I wasn't losing oil or coolant (nor did they seem to be mixing), so a cracked block doesn't seem likely, but I couldn't think of another explanation. I will definitely retest. Is there a way to know if each cylinder is on the compression stroke? Or would I just test, and if the results are bad, give it a rotation and test again?
At what point is scoring bad enough that it's not worth replacing the rings and running the engine? I'd imagine that my fingernail would catch on some of them, but I'll have to wait to check until I remove the head for one reason or another.
@dr bob - Really good info, thanks for sharing all of that. As far as I'm aware, cam timing is correct. I did both tests after the engine had warmed up, but I'm sure it was cooling down as I did them.
I did not do this compression test with the throttle open, I forgot about that step. I had previously done one with the throttle open and got these results: 155, 135, 165, 177psi (1/3/4/2). It seems like cylinder 3 has an issue, which is potentially corroborated by the leakdown test, but that might be out the window if I was on the exhaust stroke.
I'll have to perform both tests again with the proper procedures.
Thank you for the help.
I didn't even consider if each cylinder was on the compression or exhaust stroke, but that makes a ton of sense. I may have just gotten lucky with the other cylinders. I wasn't losing oil or coolant (nor did they seem to be mixing), so a cracked block doesn't seem likely, but I couldn't think of another explanation. I will definitely retest. Is there a way to know if each cylinder is on the compression stroke? Or would I just test, and if the results are bad, give it a rotation and test again?
At what point is scoring bad enough that it's not worth replacing the rings and running the engine? I'd imagine that my fingernail would catch on some of them, but I'll have to wait to check until I remove the head for one reason or another.
@dr bob - Really good info, thanks for sharing all of that. As far as I'm aware, cam timing is correct. I did both tests after the engine had warmed up, but I'm sure it was cooling down as I did them.
I did not do this compression test with the throttle open, I forgot about that step. I had previously done one with the throttle open and got these results: 155, 135, 165, 177psi (1/3/4/2). It seems like cylinder 3 has an issue, which is potentially corroborated by the leakdown test, but that might be out the window if I was on the exhaust stroke.
I'll have to perform both tests again with the proper procedures.
Thank you for the help.
-
Tom
- Site Admin
- Posts: 8928
- Joined: Fri Jun 25, 2021 2:04 pm
- Location: Silicon Valley, CA
- Has thanked: 932 times
- Been thanked: 3994 times
- Contact:
Yeah, that probably explains the odds results. To get each cylinder to TDC Compression, set the #1 to TDC by lining up the flywheel mark in the bellhousing. Then just follow the firing order for the other 3 cylinders. The order is 1-3-4-2, and you can't forget it because it's cast right into the top of the cam tower. Check number 1, then rotate the motor exactly 180 degrees so that #3 is at TDC Compression, then 180 degrees and check #4, then another 180 to check #2. Make sense?boing wrote: Sun Jul 20, 2025 4:45 pm @Tom - I had thought about the cat being clogged, maybe it would be worth running the engine with just the headers to see if that makes a difference. Honestly I don't know if it was down on power, but it never felt particularly strong.
I didn't even consider if each cylinder was on the compression or exhaust stroke, but that makes a ton of sense. I may have just gotten lucky with the other cylinders. I wasn't losing oil or coolant (nor did they seem to be mixing), so a cracked block doesn't seem likely, but I couldn't think of another explanation. I will definitely retest. Is there a way to know if each cylinder is on the compression stroke? Or would I just test, and if the results are bad, give it a rotation and test again?
At what point is scoring bad enough that it's not worth replacing the rings and running the engine? I'd imagine that my fingernail would catch on some of them, but I'll have to wait to check until I remove the head for one reason or another.
@dr bob - Really good info, thanks for sharing all of that. As far as I'm aware, cam timing is correct. I did both tests after the engine had warmed up, but I'm sure it was cooling down as I did them.
I did not do this compression test with the throttle open, I forgot about that step. I had previously done one with the throttle open and got these results: 155, 135, 165, 177psi (1/3/4/2). It seems like cylinder 3 has an issue, which is potentially corroborated by the leakdown test, but that might be out the window if I was on the exhaust stroke.
I'll have to perform both tests again with the proper procedures.
Thank you for the help.
-
walfreyydo
- Posts: 134
- Joined: Wed Dec 11, 2024 7:28 am
- Location: Wisconsin
- Has thanked: 28 times
- Been thanked: 57 times
- Contact:
Cylinder 3 appears to have issues, if the leakdown had #3 air hissing strongest out the exhaust, I would suspect an issue with the exhaust valves sealing (either worn/pitted seats/faces, or slightly bent valve). Could also be rings/cylinder walls, but you would hear more hissing out the oil cap (sound test is not always the best indicator since its a bit subjective) as you did with the other cylinders. Air sound out the oil cap indicates healthy valves and the air escaping out the rings and into the crankcase (normal as long as the pressure loss is less than 10%).
Does the car noticeably burn any oil? That would definitely lend itself to rings/cylinder walls (but could also be valve guides - especially exhaust) and would indicate the need for a rebuild of either the head or the bottom end.
If no burning oil (see or smell or loss of oil) then I would look at rebuilding the head (resurface, cut/grind valve seats/faces, new guides, new valve seals) - a much cheaper/easier solution than rebuilding the bottom end.
As a general rule:
Leakdown tests should not leak more than 10% of the pressure input (if you did the test at 100psi, you should be able to hold 90psi or higher) and should be done at TDC. Leakdowns are superior as they can help you indicate which component is the cause of low compression (valves or rings).
Compression tests should not have more than 10% variation across all cylinders. They dont help you identify the cause of poor compression like a leakdown can. I have always done mine dry with the throttle plate open (but have also done them with the throttle closed and have seen no difference in the numbers, at least on my car). I do not recommend doing a compression test "wet" since thats not the normal state of the engine when it is running and will do nothing but mask poor compression.
Finding TDC for each cylinder can be done with a long object (wooden dowel or similar), rotating the engine until it is at the apex of its height, and then installing your leakdown tester.
Does the car noticeably burn any oil? That would definitely lend itself to rings/cylinder walls (but could also be valve guides - especially exhaust) and would indicate the need for a rebuild of either the head or the bottom end.
If no burning oil (see or smell or loss of oil) then I would look at rebuilding the head (resurface, cut/grind valve seats/faces, new guides, new valve seals) - a much cheaper/easier solution than rebuilding the bottom end.
As a general rule:
Leakdown tests should not leak more than 10% of the pressure input (if you did the test at 100psi, you should be able to hold 90psi or higher) and should be done at TDC. Leakdowns are superior as they can help you indicate which component is the cause of low compression (valves or rings).
Compression tests should not have more than 10% variation across all cylinders. They dont help you identify the cause of poor compression like a leakdown can. I have always done mine dry with the throttle plate open (but have also done them with the throttle closed and have seen no difference in the numbers, at least on my car). I do not recommend doing a compression test "wet" since thats not the normal state of the engine when it is running and will do nothing but mask poor compression.
Finding TDC for each cylinder can be done with a long object (wooden dowel or similar), rotating the engine until it is at the apex of its height, and then installing your leakdown tester.
89 S2 Variocam, Megasquirt DIYPNP
Garage
Garage
I'm back with some new results from further testing, having taken everyone's input into consideration.
Compression: cylinder 2 195psi, cylinder 4 190psi, cylinder 3 150psi, cylinder 1 190psi
Cylinder 3 leak down: 90psi to 74psi. By moving the crank back and forth very slightly, and then one full rotation, I was able to get air to come out of all of the cylinders at one point or another.
So obviously cylinder three has an issue, and these results point to a problem with the exhaust valve if I'm not mistaken? I'm wondering if the rich running might have caused carbon to build up on the exhaust valve, which resulted in it not sealing well?
I also ran the car with the exhaust detached behind the headers, and after it warmed up a bit the gassy exhaust smell was gone. It also seemed to rev far more freely like this. So it seems like the cat is probably clogged.
Thanks everyone for the help!
Compression: cylinder 2 195psi, cylinder 4 190psi, cylinder 3 150psi, cylinder 1 190psi
Cylinder 3 leak down: 90psi to 74psi. By moving the crank back and forth very slightly, and then one full rotation, I was able to get air to come out of all of the cylinders at one point or another.
So obviously cylinder three has an issue, and these results point to a problem with the exhaust valve if I'm not mistaken? I'm wondering if the rich running might have caused carbon to build up on the exhaust valve, which resulted in it not sealing well?
I also ran the car with the exhaust detached behind the headers, and after it warmed up a bit the gassy exhaust smell was gone. It also seemed to rev far more freely like this. So it seems like the cat is probably clogged.
Thanks everyone for the help!