Need advice - Leak-down great, Compression test not so much
 

Hello Porsche engine experts,

Baseline - plugs look good, low oil consumption, starts with no smoke, runs fine.
Maybe 40k miles after top end rebuild, rings were replaced and cylinders properly honed by Jerry Woods shop. The cylinders are Nikasil. Cross hatching very visible at the time of re-ringing.


Well I need expert advice.

Here is what I did...
Engine was warmed up - hot. At the beginning a bit hot to work in there.
Yesterday I did a compression test and then a leakdown test.
The compression was low e.g. 90 across a few cylinders on different banks. I got a different (new) gauge - same result. The "hold" valve at the gauge on my gauge was leaking.
I got my leak-down devices and did the tests. I tested at 80 psi and all the second readings were very close to 77 or 78. I heard no air moving. Taken at TDC and crank held with transmission and brake. Crank didn't move.

Now what do I try other than to take the engine apart? PEA cleaner?

Thanks,
-Henry

Calling Henry Schmidt? Henry this is Henry. :-) I've posted on a mod you made to chain tensioner part. You made the pivot bushing longer. That was a while back. Thanks.

Just throwing this out there because it happens.Did you rotate the engine to be at TDC for each cylinder?

Yes, rotate to TDC per the marks on the pulley and watching the rotor. I can hear the air rush out just prior to TDC - this confirms I'm on the compression stroke.

This brings up a question.... maybe do a leak down at other than TDC in an attempt to find the location of the compression leak. I see this discussed here and there.

Thanks for the question.

Sounds like it runs fine. Any other issues other than low compression on the gauge? Was the throttle open when you tested compression?

Car runs fine. Throttle open but we tried it both ways and didn't make much difference. The question is what to do now. Repeat compression test, repeat after running it hard, try some sort of ring cleaning juice, etc. , etc. I can take the cat off and put on the test pipe if running something through it will possibly help buy may cause cat damage.

If you can hear the air rush out just prior to TDC, that sounds more like the overlap part of the cycle, not the compression stroke.

If the compression is low, try fogging some oil in the cylinder then retesting. If the compression comes up, the rings aren't sealing. But if the leakdown was ok, you probably already know that. 78/80 on the leakdown is fine.

Incorrect cam timing can throw off compression readings, too. How confident are you that it's correct?

Update after better sealing of the Stomski attachment and using a better gauge valve.

Cylinder 1 - 135psi (had a bit of oil put in yesterday)
Cylinder 2 - 125psi
Cylinder 3 - 131psi
Cylinder 4 - 132psi (had a bit of oil put in yesteday)
Cylinder 5 - 122psi
Cylinder 6 - 129psi

Viewed 2 cylinders with borescope some carbon is noted at the bottom on the cylinder walls. Places where we looked on the cylinder still had cross hatching.

@dannobee - Places I read indicate the air rushing out is TDC on the firing stroke. I'm pretty sure of the cam timing - I used a dial indicator when I built the top end and followed the Wayne book. I put oil in two cylinders yesterday and the numbers came way up. Today the numbers on those cylinders are down a bit.

I'm thinking to get advise from Henry or ? on something to run in it to clean some amount of carbon and then retest to see what happens. Maybe leave well enough alone at this point.

Those numbers were taken this morning on a cold engine - I should have noted that.

Can you tell if air is leaking by the rings or through the valves? Maybe just double check your valves and adjust if needed.

The valves will be on dwell during a leak down test. Adjustment won't change that.

They should be if they are adjusted correctly, but if they are too tight won’t that prevent them from closing completely?

Well I guess yes but during the 20 years I've been adjusting the valves on this car the gap hasn't changed much and not near zero lash. I guess it could of happened but I adjusted the valves about 2 months ago and there was one valve that was maybe 1/2 mil tight.

Probably not the issue then. I have nothing further. I’ll page Henry for you, wait here.

Thanks much.

My ears were ringing. I'll give it a try. If the cylinders were "properly honed" years ago you may be dealing with excesses wear. Back in the day, honing Nikasil was a little more of a mystery than it is now. RA factor was at times ignored.
If the RA factor of the cylinder was in excess of 10, lets say 12-18, with 40,000 miles on a set of rings, the rings are worn out.
If you're dealing with excessive carbon, you may have stuck rings. The way to clear the carbon is to introduce water into the fuel mixture. With carbs, MFI or ITBs that is pretty easy. Poor a small amount of water [1 or 2 ounces] into each cylinder. The steam created by the water will break the carbon free. CAUTION! TOO MUCH WATER WILL HYDROPHILIC THE ENGINE, POSSIBLY BENDING THE RODS!
If that frightens you [why wouldn't ?] you could try some Supertec Bearing Life.
It's designed to reduce friction on contact parts but it has been known to free up sticky rings.
For now that's all I have. Good luck.

Henry, Thanks!
I'll ask Jerry Woods re. RA that they might have used.
For the water..... pour in water. Crank engine to get excess out to avoid hydro lock.
Then install plugs and go for it? Is that the process? I have Motronic.
Supertec Bearing Life - where does this go .. I guess the crankcase? I didn't see this on your website.

Henry.... maybe spray water in the air cleaner housing with the engine running? This should prevent hydrolock. Thanks.

If the rings were worn out - would that show up on the leak-down test?

Introduce the water while the engine is running (fast idle). A little bit at a time. You might end up running a pint or two over a few minutes time. Just a little at a time. The more "new school" way is using GM Top Engine Cleaner as specified on the label and toward the end, let it back down to an idle and let it run rich enough until it dies. Then let it sit for a couple of hours and let the stuff work its magic.

With respect to leak down testing, check which cylinder the ignition rotor is pointing to and look at the crank pulley to set that cylinder to TDC. Use 100psi and the second gauge will read the percentage directly. There should be no big whoosh of air leaking anywhere.

In general, compression tests on 911 engines is inaccurate.
When you remove the the spark plugs, carbon gets dislodged and settles on the exhaust valve. If there is no detectable leakage from the leak-down test, the compression test is almost meaningless.
What are you trying to accomplish?

GM Top Cleaner sounds like the safest way to go and I assume will not give the cat or the O2 sensor a bad time.
I don't have an issue with using 100 psi, I used 80 psi and can convert to % or do over then next time I test after GM TEC. One issue with using 100 is the needles are near the "end of readings" on the gauge and that's not the best accuracy. Yes I used the rotor pointing method and watched the crank pulley mark to align with the mark on the fan housing.
Thanks very much for the comments and helping me along.
Stay tuned... I'm going over to Jerry Woods tomorrow and ask what RA they might have targeted.

Ah, okay I just saw this. I'm getting all my ducks lined up for selling the car. Many buyers want to see both tests. Note that the compression numbers were... I would call fairly close or consistent - this might imply the carbon from the plug areas fell in six places in the same manner. I want to have compression numbers that will not scare people off from buying the car or not scare them too far.
The leak-down only dropped 1 or maybe 2 psi from 80 for all 6 cylinders. I heard no air moving.

I appreciate all the comments and help. As noted below (I guess now above) I'm going to try GM and then maybe a water intro. at fast idle. I might just put in some GM and then drive up HWY 280 in 3rd gear and load the rings on and off at a high engine speed. Then do a compression test again to see if there are any changes.

Thanks again.

In case a picture of the plugs give any additional clues.

http://forums.pelicanparts.com/uploa...1632095859.jpg

Sounds to me like a normal good running engine. Nothing to fix.

I would never reply to this sort of post, but this is gone past reality.

You have even said, the oil consumption is good and the car runs fine. So, what are you going to do, pull the engine apart to do what?

Don't bother Jerry with what RA finish he used? After 40K miles you can bet its not the same. Get on with what is important in life.

Your mission is commendable for the new owner but...... have you ever heard the saying,

" Just like an old lady with a new table"""

While your compression numbers might not rise to the comfort level of some prospective buyers, what seems more important is the consistency across the board. Take the car out for a hard drive as planned, maybe squirt in the GM juice first, and see what you've got. If the numbers are still consistent, I don't see a problem.

+1

Those plugs look consistent and pretty darn good. I concur with the "just drive it" group.

So to everyone saying "there's nothing wrong, just drive it," you'd be comfortable spending $70k+ on a 3.2 Carrera with "low" compression numbers? He did say he was selling the car and didn't want those numbers to scare folks away.

How did you do this?

Here is what you should do:

1. The engine needs to at operating temperature.
2. The coil should not fire and plug wires should be off the spark plugs.
3. The throttle should be fully open.

Now, remove one plug and screw in hose and attach compression tester. Turn the engine over for 5 to 8 complete revolutions until the value on the gauge stabilizes. Write down the value.

Repeat for all cylinders.

the test
 

Those compression readings are fine.Leakdown is good.Where is the problem?Different compression testers give different values.Snap-On tester reads 10% less than a Mac-Tool tester.It is just a number.The only number that does matter is the leakdown.That covers everything.Chill out.Fred

The leakdown is fine, that says it all.

I reassembled things yesterday and drove it. Ran fine.
Good point on testing with a hot engine. I know to do that and will do it on the retest.
Google could not find any writing that said to take out one plug at a time. On engines with head gaskets and water it was mentioned to take out all plugs so that a blown gasket would not affect readings. In any case it only affects the cranking speed. We tried open and closed throttle, no change but we used open throttle.

Seems odd. Maybe you don’t have full throttle at the throttle body.

Porsche specifies only leak down test is valid for determining engine condition.

okay, good point - I didn't know that. I could have the compression test values and if asked by prospective buyers will provide, if not well they didn't ask.

My compression numbers, at 5,000 feet, have always been low to what others have found. I leaked my stock US 82 3.0 down, and got ~0 (the two gauges don't quite line up with no pressure on them, but when I hear no noise at all I can buy 0) on all but one cylinder, which was at 2%. Since I have a pop off valve, I could get my listening hose right into the #2 intake, which is where the faint noise was. Unlike most times I have done a leak down, I heard no sound in the oil tank (from rings) - usually there is a little there. Exhausts were ~0. I call that good.

The compression numbers were in the low 100s. I put a wrench into the pop off valve, to simulate WOT. I was surprised to note just about no difference in the compression readings.

Car ran fine racing.

Walt, thanks for the info.
Same as what I found re. open throttle or not. Maybe air is getting past the ICV.
I was thinking more on what Winders said. One plug out at a time might reduce the chance of carbon getting stuck in the exhaust valves or keep more heat in.
Right now I solved the problem of the AC stopped working so I'm putting things back together.

@hcoles

giving 135 PSI to a combustion chamber for a leak down test is far too much for a precise test result. By this most test results are between 2% and 5% which looks good on the first view.

Most quality leak down test Gauge/Units like from Hazet do internally limit the input pressure to approx. 4 Bar where by this a wider reading bandwith of the actuall pressure loss is possible.

Give em about 58 PSI and re check the leak down value.
As shown below everything down to 20% is acceptable as long as all cyclinders are almost matching

https://d2q8g8w0j97itz.cloudfront.ne...808/4795-1.jpg

20% leak down is not acceptable for a 911 air-cooled engine…..