Xantia V6 broken exhaust (update: and rough low rpm running)

[larppaxyz]

It's simply not possible for the belt to cause your problems.

Probably not, but we will soon find out

But it's not that far fetched idea, with chain engines symptoms are misfiring, rough idle and other timing issues.

[Mandrake]

It's simply not possible for the belt to cause your problems.

Not to mention I can't see how a stretched belt would cause intermittent symptoms. The day after my recent post saying the performance was terrible the performance was semi-normal again, at least for the first 10 minutes of driving, then it started sagging again.

Now that I've also ruled out the fuel pump I'm pretty convinced I'm looking at an ignition related problem.

[addo]

What if you jerry-rigged an interruptor switch to one pair of coils? Then you could see if it truly reflects your symptoms.

Yes, you'd have to cut the harness to do this.

But it's not that far fetched idea, with chain engines symptoms are misfiring, rough idle and other timing issues.

It is, in the context of this engine design. Once you see how the belt is routed and tensioned, it's easy to appreciate how it's all but impossible for this to occur.

[Mandrake]

What if you jerry-rigged an interruptor switch to one pair of coils? Then you could see if it truly reflects your symptoms.

Yes, you'd have to cut the harness to do this.

I'm really reluctant to cut the harness to do this test, as I don't want to be trying to make reliable, air tight joins at a point where there is hardly any spare wire to play with. I'm not sure what it would prove anyway as I'm not looking for a dead miss (which is what I'd simulate) but a misfire under load which as you know is really hard to reproduce when the car isn't under load driving or on a rolling road.

Although I'm toying with the idea of a temporary ECU swap to rule out the ECU causing ignition primary side issues I've decided that I need to do something a bit more basic first:

If I've been running with a faulty coil pack induced under load misfire for the last few months chances are good I have some seriously fouled plugs again. I was hoping they'd clear themselves with a bit of Italian driving but it doesn't seem to have happened. Last time the plugs got really fouled even after I did the plug leads/coil etc it still ran badly until I did the plugs. So it could be that I've fixed the original cause of the misfire (coil pack) but the fouled plugs themselves are still causing a misfire.

The second point is that I've been a bit lazy and re-used the inlet manifold gaskets at least once each the last few times I've had the manifold off - I've had the manifold off many times for various reasons and got sick of buying new gaskets, and I've been careful not to break the gaskets.

However I've recently realised that the gasket that is on there now may have been reused twice - so there is a good chance its compressed (it did look a bit compressed) and no longer sealing perfectly, so I could have a mixture imbalance between different cylinders that leads to a lean misfire on one or more cylinders and/or the ECU correcting the mixture for the lean cylinders causing others to run rich and foul the plugs.

So I think new plugs and a new inlet manifold gasket are in order. I'm going to do that some time next week and see where we stand after that.

[Mandrake]

Picked up three inlet manifold gaskets yesterday, yes three.

One to fit immediately and two spare ones so that I don't find myself forced to re-use the gasket, let alone re-use it twice. Given that I'm still potentially hunting a problem the manifold may need to come off again later so having a couple of spares on hand is wise I think.

Do you guys fit the gasket dry ? Or smear a little bit of grease or sealant on it ? On different occasions I've fitted them both dry and with a slight smear of grease but can't say I've noticed any difference, unless advised to the contrary I think I'll fit it dry this time.

Still waiting for the spark plugs to arrive. Can anyone think of anything else to check while the manifold is out that could potentially be causing my issues ? Other than checking for oil in the wells of course. Air leaks on the injector to inlet manifold mounting perhaps ?

[addo]

Check every sensor possible, by measuring at the ECU plug.

[Mandrake]

Check every sensor possible, by measuring at the ECU plug.

Have already done so in the past. Can't find any problems with sensor inputs to the ECU.

I've even checked the crankshaft sensor and knock sensor signals on a scope at the ECU connector with a running engine.

The crank sensor output looks perfect and has plenty of output, (many volts p-p) the knock sensor has a relatively low output in the 10's of millivolts, (which I think is normal) but shows the expected damped oscillation when the block or inlet manifold is whacked with a tyre lever, and a small amount of random "noise" on a running engine.

Oxygen sensor measures ok, TPS and MAP sensors measure fine, coolant/air temperature sensors measure fine. Have I forgotten anything ?

On the output side, coil primaries are good (0.5 ohms) and I've also tested the wiring from the ECU connector to the coil pack for any sign of intermittent or high resistance connection and there is none, with less than 0.1 ohms in the wiring loom and no intermittent issues.

I haven't checked the continuity from ECU connector to injectors recently, each injector is 12 ohms so the parallel pairs should measure 6 ohms, I'll check that next time I'm under the bonnet.

One thing I haven't tested mainly because its hard to get at or test is the canister purge solenoid that is connected by a long hose to the bottom of the butterfly. The Lexia reports a "believable" duty cycle for the solenoid with various conditions (minimal opening at idle, much higher opening duty cycle under wide throttle) but I haven't made any attempt to confirm that the canister purge solenoid is actually working and isn't stuck open or closed, or that there isn't a vacuum leak on the hose.

What would be the symptoms of a stuck open purge solenoid ? Or is it unlikely to cause what I see ? If it was introducing a vacuum leak due to the hose then I would have though it would only upset idling and low throttle running but not cause major power loss at wide throttle ?

What would be the symptoms if the hose from the purge solenoid to the canister had an air leak or has fallen off/split ?

Anyone know how to test the functioning of the canister purge solenoid and verify there are no leaks on the hoses given that its well hidden inside the front left wing and therefore nearly impossible to get at ?

[addo]

If the tube is broken it will act like a vacuum leak. Smoother running at a decent clip. Failed operation means permanently closed; no real effect other than slight richness at idle. The orifice is quite small.

It's hard to hear with the 50Hz actuation test because it's wrapped in soundproofing foam.

[Mandrake]

If the tube is broken it will act like a vacuum leak. Smoother running at a decent clip. Failed operation means permanently closed; no real effect other than slight richness at idle. The orifice is quite small.

An electrical failure would mean it fails closed yes, but it could mechanically fail open. Are you sure failing closed would cause it to run rich though ?

When the valve is opened (which tends to happen mostly when the throttle is open) it draws the petrol vapour from the canister and thus actually makes the mixture richer. I'm sure I've read in one of the MP7.0 docs that the ECU has to compensate for the richer mixture when the solenoid is open, and automatically reduces the injection time slightly.

I doubt its a source of major power loss but I'm wondering if a fault in this system could cause poor mixture regulation under light load changing conditions. Lets say there was an air leak between the solenoid and the charcoal canister due to a faulty hose, when you accelerate and the ECU opens the solenoid it will be expecting the mixture to richen slightly due to the petrol vapour but instead it would just get a vacuum leak instead so it would lean out.The mixture swinging the opposite way to what the ECU expects could cause a stumble until the oxygen sensor loop has time to compensate.

I am quite often seeing little momentary "stumbles" when opening the throttle lately. Sometimes from a standing start with a modest throttle opening there will be a momentary stumble where the revs dip before they start rising. It could be misfire related of course but its something to think about.

I suppose I could try temporarily blanking off the spigot on the butterfly housing that it goes to see what effect it has, although there is nothing to say that it would run optimally with it blanked off if the ECU is expecting a mixture change due to the petrol vapour. It might run better than if there is a vacuum leak due to a faulty hose though.

[addo]

Blank it off at the throttle body and see what happens to your injector dwells. The carbon canister can/does also draw in air.

It can't fail open mechanically; the valve is opened by a wave voltage and there's a (proportionally) strong spring that closes it immediately the driving waveform ceases. If it were obstructed by debris that somehow entered the airline and created themselves into a suitable shape which held the valve open without blocking the metering orifice, I for one would be extremely surprised. It's coming close to the "tornado in a scrapyard" analogy.

[Mandrake]

Well, this car sure loves throwing me curve balls!

I had a play with the canister purge system today after fixing the cruise control - I removed the hose from the throttle butterfly and tried sucking on it - no air leak when the electrovalve is off.

I then used the Lexia actuator tests to operate the canister purge electrovalve - when the electrovalve opens I can then easily suck through the hose although its hard to judge how much restriction there is - it doesn't seem like much restriction at all. There is a very faint barely noticeable taste of petrol, no liquid, just a bit of vapour.

So what we know:

* The electrovalve works and is controllable by the ECU
* The hose from the throttle butterfly to the electrovalve is air tight.
* There is at least some petrol vapour coming through the hose when the valve opens, but not much.

What we don't know:

* Is the hose from the electrovalve to the canister also air tight or is it introducing a vacuum leak when it opens ? There seems to be very little if any restriction, I would have expected the charcoal canister to offer significant resistance.
* Is the canister intact, and is the line from the petrol tank to the canister also air tight ?
* When the electrovalve opens where does the air come from that the engine sucks through ? Does it just pass from the fuel tank through the charcoal canister and through the electrovalve ?

Regardless, I blanked off the spigot on the throttle butterfly temporarily and went for a long drive. Here's the curve ball - it seems to have made a major difference.

There was an obvious improvement right away, and it got better and better for the first 15 minutes or so of driving before stabilising. I wouldn't say it was perfect but it was driving with a real sense of urgency and responsiveness both at light throttle and wide throttle that it hasn't recently.

The odd behaviour it sometimes used to have where you would try to accelerate and it would momentarily respond but then go flat for a few seconds and then pick up again seemed to be gone, and the idle seems a lot more consistent as well.

In total we were out driving around for a couple of hours looking at houses, both motorway and urban speeds and for most of that time you would never have thought there was anything wrong with the performance of the car...

Now we all know just how intermittent and contrary this car is, so I'm not jumping to any conclusions, and I am still going to do the spark plugs, manifold gasket, and oil and filter change, hopefully all this week as I have the week off.

Other confounding variables are that I only drove the car for a couple of miles after blanking off the spigot before I filled the petrol tank from nearly empty - so I can't rule out a change in fuel quality as a contributor, although it did already seem a lot better before I filled up, and previous fill ups at the same garage with the same fuel had not seen the same improvement.

Also to get access to blank the spigot off I had to remove the air filter box - and past experience tells me that when I disturb the air filter box and surroundings it sometimes runs better for a while, for unknown reasons. (I didn't disconnect or reset the ECU though)

However it really has got me wondering... what if the hose from the electrovalve to the charcoal canister has split or fallen off ? At idle the electrovalve is only open about 5-10% so a leak there wouldn't affect idle much, but as soon as you put your foot down a decent amount the electrovalve opens 60-80% duty cycle or so, which would then introduce a vacuum leak and lean out the mixture. That could potentially be the cause of the lean oxygen sensor reading I've sometimes seen on wide throttle.

Another thing is that the purge solenoid doesn't always operate - it goes through various modes of operation where it usually opens when you open the throttle, but not always, depending on coolant and air intake temperature and other factors, which could explain some of the intermittent contrary nature of the symptoms.

Has anyone ever had the charcoal canister out or looked closely at it when they've had the bumper off ? I suppose getting at the hose from the electrovalve to the canister is nearly impossible in-situ ?

While I was doing the cruise control today I could see the bottom of what I believe to be the charcoal canister - with a large tube at the bottom that doesn't seem to go anywhere, but the top looks very inaccessible..

I'm inclined to leave it blocked off for a while and see how the engine continues to behave... can anyone see any problems with doing this ?

[addo]

Have you had a look with the blinker out?

[Mandrake]

No. Judging by looking at it from beneath I think its a long way back from the blinker though.

[addo]

Not really.

[Mandrake]

Ok guys, I might be onto something here

Ever since I've blanked off the spigot on the lower butterfly that goes to the canister purge electrovale the car has been running MUCH better. Not perfect but massively better than it has been recently.

The car has loads of low down pull now and will easily accelerate from 1500 which it wouldn't before. The weird flat spot that would appear randomly under acceleration during pickup and go away after a few seconds seems to be gone. Idle seems nice and stable now too.

At high RPM there is good power but it MAY still be dropping down to 5 cylinders under wide throttle judging by the fact that the engine is still quite "raspy" sounding under wide throttle, and doesn't have as much wide throttle performance as I remember at its best, but under low to medium throttle it seems to be smooth with plenty of power and responsiveness now with no random variations in performance.

Assuming its not yet another false alarm (!) it seems like there might be a vacuum leak somewhere on the canister side of the electrovalve - in which case as soon as the ECU opens the electrovalve during acceleration it inadvertently introduces a large vacuum leak that leans the mixture.

If it was a constant vacuum leak like a manifold leak the ECU would eventually adapt and compensate for it through oxygen sensor feedback but because the "leak" keeps varying in response to the electrovalve activation percentage (which itself varies on a number of sensor inputs) the ECU sees it as a random mixture variation that it can't explain or adapt to.

So the engine ends up running too lean in some conditions and too rich in others. When its too lean it probably either starts knocking and triggers the knock sensor to retard the timing (loosing a lot of power) or misfires, causing an even larger loss of power. When its too rich it may be fouling the spark plugs.

I only have a quick kludge consisting of some clear thin walled hose with a bolt shoved in the end which doesn't really fit the spigot properly, so I'm going to fit some blanked off 7mm ID fuel hose to make it a bit more permanent so I can leave it blanked off for quite a while to monitor its performance.

The only thing I'm not sure about is whether I should leave the hose from the electrovalve disconnected or whether I should blank that off too ? If I left it off then the petrol vapours from the canister would escape into the air although I don't think that's a bit deal as there is nothing to monitor the vapour recirculation system on this ECU, and a working vapour recirculation system is not an MOT requirement for a non-monitored system ?

If I was to block it off permanently would it cause pressurisation of the fuel tank, or is there a relief valve somewhere ? Bearing in mind that the electrovalve only ever opens when the engine is running anyway. Would blocking it off permanently damage the charcoal canister over time by over saturating it with petrol vapour that never gets drawn off to be burnt ? Or would permanently blocking it be harmless ?

It's been around three or four weeks since i did that and it's been much better. Not perfect still, but there is no 3500rpm lag appearing yet.

Right now my car is waiting for new belt(s), tensioners, water pump and most importantly new engine mount rubbers. Cost ~1000 euros inc. labour.

How's your car running lately ? If you're still having problems with low rpm performance and pickup could you try blanking off the right hand lower spigot on the throttle butterfly as described above ? I'd be interested to see what difference if any you notice.