Rough running problem

[BigBen]

I have recently acquired a budget 2003 Boxster S. The car came with the following problem:

i) After running for 30 miles or so the car starts running poorly and ultimately the check engine light come on
ii) Resetting the error results in the car running smoothly again
iii) The 30 miles can be done 3 miles a day over 10 days or all at onced
iv) The ECU 'thinks' the car is running lean on bank 1 so adjusts the fuel trim to full rich which is the source of the problem.

I have tried replacing the lambda sensor as that is what the computer reports as the problem, but I think that is a symptom rather than the cause and it goes out of range due to the rich running.

The symptoms all suggest a vacuum leak somewhere on bank 1 downstream of the MAF, does anyone know of any potential leak hotspots?

Any other suggestions, no hurry I won't have time to try much for a couple of weeks

Ben

[Steve H]

Most likely would either be an air leak causing it to run weak or something electronic pulling the fueling down.

You can look for air leaks by spraying carb cleaner round the inlet system at idle and listen for changes in the engine note but if it was an air leak big enough to flag a fault while driving I'd expect it to be pretty poor at idle all the time.

Next favourite would be if the airflow meter was reading a bit low, this is particularly common on Bosch MAFs from around that era, they were fitted to some Porkers, possibly yours. This is much harder to prove without substituting the part and either running it or observing how it affects the fuel trims.


All that said and at risk of sounding critical, your description of what is happening suggests that you're a bit out of your depth and should probably get some qualified help.

[cd1957]

Hi what fault code are you getting.

Chris

[BigBen]

Steve H said:

Most likely would either be an air leak causing it to run weak or something electronic pulling the fueling down.

You can look for air leaks by spraying carb cleaner round the inlet system at idle and listen for changes in the engine note but if it was an air leak big enough to flag a fault while driving I'd expect it to be pretty poor at idle all the time.

Next favourite would be if the airflow meter was reading a bit low, this is particularly common on Bosch MAFs from around that era, they were fitted to some Porkers, possibly yours. This is much harder to prove without substituting the part and either running it or observing how it affects the fuel trims.


All that said and at risk of sounding critical, your description of what is happening suggests that you're a bit out of your depth and should probably get some qualified help.

Yes it does seem like an air leak but as you say I would expect that to cause an idle problem all the time. Either spraying carb cleaner or getting a garage to run a smoke test are possible options for finding an air leak

The problem is only apparent on Bank 1 but the airflow meter serves both banks so it is unlikely to be that. That said it hasn't been changed so replacing it probably isn't a terrible idea.

On your final point I was trying to avoid writing too long a post on the subject so my level of understanding may not have jumped out from the screen, but I am fairly confident in my skillz Getting help would spoil all the fun satisfaction of fixing an odd problem such as this one.

cd1957 said:

Hi what fault code are you getting.

Chris

Fault code P0140, p2236, P0171 which are all 02 sensor related, but per my OP I believe these are a result of some other problem rather than O2 related.

[Steve H]

So what are the fuel trims running at on bank 1&2?

[Steve H]

I wouldn't bother with the smoke test unless there are any pipes that are hard to check between the MAF and throttle body, anything after that will be under heavy vac at idle and should be easy to find with carb cleaner.

[BigBen]

Steve H said:

I wouldn't bother with the smoke test unless there are any pipes that are hard to check between the MAF and throttle body, anything after that will be under heavy vac at idle and should be easy to find with carb cleaner.

My only reason for thinking of a smoke test was it must be a very small leak to have such a slow effect on fuel trim, anything bigger would surely cause rough running at all times.

Steve H said:

So what are the fuel trims running at on bank 1&2?

I don't have the car to hand today, will check this evening and report back. The bank 2 Lambdas behave exactly as they should so suspect there will be quite a difference between the two.

[Steve H]

BigBen said:

My only reason for thinking of a smoke test was it must be a very small leak to have such a slow effect on fuel trim, anything bigger would surely cause rough running at all times.
.

I was looking at it the other way, if it's flagging after being driven along (as opposed to doing it at idle) a small manifold leak is very unlikely to be the cause as they tend to show up mostly with the butterfly closed and with high vac in the manifold.

[BigBen]

Steve H said:

BigBen said:

My only reason for thinking of a smoke test was it must be a very small leak to have such a slow effect on fuel trim, anything bigger would surely cause rough running at all times.
.

I was looking at it the other way, if it's flagging after being driven along (as opposed to doing it at idle) a small manifold leak is very unlikely to be the cause as they tend to show up mostly with the butterfly closed and with high vac in the manifold.

I have been thinking along the same lines, have not experimentally proved this yet but I have the impression that the time between error codes is longer if driving compared to sitting at idle which indeed suggests the higher vacuum at idle accentuates the problem.

I have just had a look at fuel trims following reset:

Bank 1 short term heads for ~ +25% with the long term at 0% (presumably this would change after, well, a long term!)
Bank 2 short term oscillates within +/-3% of 0% with long term also at 0%

Pre reset I logged too many parameters and the app ran out of graph colours so can't really see what is going on, will try again later with only the short and long term values for each bank logging.

[Steve H]

Just a shot in the dark, thinking about what can affect one bank but not the other, any chance that the cam timing has skipped or been set wrong.........?

[BigBen]

Steve H said:

Just a shot in the dark, thinking about what can affect one bank but not the other, any chance that the cam timing has skipped or been set wrong.........?

That was actually one of the first things I looked at on Durametric as an initial though was the Variocam (or whatever it is called) solenoid being dead but it all seemed to check out. The car has not had cam work so it won't have been set wrong.

A few more observations from this evening. I went on a lenthy ish run and sure enough the long term fuel trim on bank one rose to match the short term at around 25% (as one would expect). I did a reset at the end of this run then left the car for an hour.

On the next start up the bank 1 fuel trim was at more like -20% but in the short 1 mile drive from the gym returned to +25%. This makes sense as I believe fuel trim is set from the Lambda sensor which was reading +0.9V at the start of the journey then +0.1V at the end, e.g. was initially reporting rich running then went to lean.

I though I was logging both trips but had the logger set up incorrectly, will try again tomorrow or possibly in a minute.

Ben

[BigBen]

Couldn't wait so here are some logs of a drive following an ECU reset:

As described Bank 2 behaves as one would expect and Bank 1 is all over the shop.



Next added RPM to the graph and for improved clarity remove Bank 2 data. It is difficult to tell as the logging rate is probably quite slow but it looks a bit like lower RPM results in the ratio heading North, but inconclusive. Worth re-runnign with fewer parameters being logged to get a higher data rate so improve resolution (this will be tomorrow as I need to go to bed)



Next graph shows Bank 1 Lambda Voltages and RPM. Harder to convince myself that there is any correlation between the two in this graph (and therefore possibly shows the ratio / rpm graph is a red herring given ratio is derived from Lambda)



Final graph shows Bank 1 Lambdas and Throttle opening



I think the next step is to run logging again but with Bank 2 logging turned off to get a higher rate on Bank 1 / RPM / Throttle to see if there is any correlation.

Ben

[Steve H]

I'm not sure how this is leading you to a solution TBH. We know it's running weak on one bank, most of the remaining causes aren't going to reveal themselves on diag kit.

[BigBen]

Steve H said:

I'm not sure how this is leading you to a solution TBH. We know it's running weak on one bank, most of the remaining causes aren't going to reveal themselves on diag kit.

Agree, I am going to do one more test of RPM vs fuel trim as follows:

i) run at idle for 30 secs
ii) 2000 RPM 30 secs
iii) 3000 RPM 30 secs

If the fuel trim is 'worse' at lower RPM then it will give more certainty that it is a vacuum problem.

The other reason for doing lots of diagnostics at the moment is I don't have time or access to my garage / lift until the 17th ish and want to be doing something. As you know the engine isn't massively easy to access on the Boxster!

Ben

ETA job lot of carb cleaner arrives today.....

Edited by BigBen on Tuesday 6th March 09:15

[Steve H]

[cd1957]

Hi just looked at Porsche diagnostic codes and this is what has come up;

p0140 No oxygen activity found
p2236 Heater circuit fault
p0171 adaption range below limit

Chris

[BigBen]

I could not wait so took the afternoon off today to spend a bit of time on the problem. I carried out two tests:

i) I sprayed carb cleaner around the engine bay aimed at the intake air path for bank 1. There was no change to the way the engine ran at all.

ii) There is a one way valve that feeds into the intake on the bank 1 side, I guess breathers from both crank cases (two connections into a Y output into the valve). I removed this valve and confirmed its one way function was still working. I then tried running with this valve reversed to remove the chance of some leak upstream of this being the cause. It wasn't!

Whilst under the car I also noticed that the second (post CAT) O2 sensor had not been changed for ages so this may be worth doing next, however my understanding was the second sensor does not actually feed into the fuelling control loop so is unlikely to be a factor here.

At the moment not sure what to try next.

Ben

[Steve H]

Yep, rear sensor is just for cat monitoring really.

Do the two sides each have their own fuel pressure regulator or are they piped together?

Could be worth swapping the injectors from one side to another. No reason why three would all suddenly get restricted and if it was just one you'd expect to hear a single cylinder miss but it's worth a look.

I'm trying to think what else will only affect one side of the engine and I keep coming back to timing as one of the few remaining items.

[BigBen]

Steve H said:

Yep, rear sensor is just for cat monitoring really.

Do the two sides each have their own fuel pressure regulator or are they piped together?

Could be worth swapping the injectors from one side to another. No reason why three would all suddenly get restricted and if it was just one you'd expect to hear a single cylinder miss but it's worth a look.

I'm trying to think what else will only affect one side of the engine and I keep coming back to timing as one of the few remaining items.

Don't know about the fuel pressure regulation. However the exhaust smells like a petrol station so don't think it is detecting a lean condition due to lack of fuel! I can hear a ticking injector but it is on bank 2. If I let it run long enough with the long term fuel trim at 25% you do get a misfire on bank 1, not enough to hear it beyond the rough running of the engine by that point.

Other details are the engine is only rough at idle, at higher revs it remains OK even when the problem has occured.

The previous owner says he had an exhaust leak test done, as an exhaust leak pre-O2 sensor would also produce a false lean condition. I have no evidence of this actually being done so could try again.

I plotted the bank 1 and bank 2 cams the other day and they tracked each other well so I think the timing is OK.

[Steve H]

It isn't just misfiring on one cylinder is it?

That would give low O2 sensor readings which would indicate weak mixture on OBD.