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

[stevieb]

I'm still encouraged by the fact that stevieb swapping his injectors did seem to solve the problem for him - maybe he was a bit luckier than me with his second hand injectors and got a better set than I did. As I mentioned earlier 3 of mine were well and truly stuck closed and unresponsive when I first started testing them before fitting them, one of them took over half an hour of patient thumping with a screw driver (while mounted in the rail) with the injector tester running on constant pulse...

It's true that my injector swap did give huge improvements, but by far the biggest improvement in consistency of performance has come from replacing the fuel pump, and I often wonder how the original injectors would've performed had I replaced the pump first. It wouldn't have been perfect, considering I had a noisy injector causing its own issues, but I'm curious how healthy the other injectors were especially after last year's other various bits of work.

[stevieb]

So, a simplistic view is that the ECU is reverting to a more retarded ignition state, only after a period of time of hot engine/long running? Short journeys where the engine bay etc. don't get fully up to temperate don't trigger the change at the ECU?

But have you tried a long journey immediately after the ECU reset and does it cause a return to the problem on the subsequent start-up? Is it the long journey/hot engine bay, or is it simply the miles being covered that's triggering the change in the ECU.

We know the ECU learns over time, so if it's true that the new parameters don't kick-in until a given mileage, it would obviously take longer to achieve if you only did short journeys, giving the impression that the heat generated on longer journeys is related to the issue. I know I've read numerous times that the ECU needs around 200 miles to begin applying the changes based on its learning (it's probably a round figure in kilometres to be more precise), so is it that a single 50 mile journey after an ECU reset isn't enough to cause the problem, where 200 miles of only short/cool journeys could cause it? Just to try an eliminate heat as a potential contributory factor...

I guess if heat is the issue then it might be wax in the injectors causing problems, where the wax is softening with all that extra heat and impeding fuel flow (either by expanding slightly or by being pushed closer to the opening maybe), but not being softened enough on short/cold journeys.

I also have strong suspicions about the ICV in a warm engine bay. If the lubrication is lacking in the two bottom bearings there's a risk that it isn't reacting smoothly enough (and don't forget this device is more important at low rpm/low throttle opening), which might cause an inconsistent mixture and subsequently all sorts of spurious readings at downstream sensors.

I can't remember whether I asked before, but have you cured the vacuum leak you had? Can you stall the engine at tickover by blocking the ICV/air filter hose? If not, then could it be that the warm engine bay situation is encouraging this leak (hot air being thinner, and possibly the location of the leak becoming larger with expansion).

There are still a lot of variables that can be checked/cured for free before firing off those old injectors for cleaning

[Mandrake]

It's true that my injector swap did give huge improvements, but by far the biggest improvement in consistency of performance has come from replacing the fuel pump, and I often wonder how the original injectors would've performed had I replaced the pump first. It wouldn't have been perfect, considering I had a noisy injector causing its own issues, but I'm curious how healthy the other injectors were especially after last year's other various bits of work.

In my case I've replaced the fuel pump and the problem persists, so I have to keep looking.

So, a simplistic view is that the ECU is reverting to a more retarded ignition state, only after a period of time of hot engine/long running? Short journeys where the engine bay etc. don't get fully up to temperate don't trigger the change at the ECU?

But have you tried a long journey immediately after the ECU reset and does it cause a return to the problem on the subsequent start-up? Is it the long journey/hot engine bay, or is it simply the miles being covered that's triggering the change in the ECU.

Yes I've tried an ECU reset on an already warmed up engine and then been for a long drive and the problem usually returns during the drive, let alone the next time the car is started, although it does depend on driving style and conditions.

We know the ECU learns over time, so if it's true that the new parameters don't kick-in until a given mileage, it would obviously take longer to achieve if you only did short journeys, giving the impression that the heat generated on longer journeys is related to the issue. I know I've read numerous times that the ECU needs around 200 miles to begin applying the changes based on its learning (it's probably a round figure in kilometres to be more precise), so is it that a single 50 mile journey after an ECU reset isn't enough to cause the problem, where 200 miles of only short/cool journeys could cause it? Just to try an eliminate heat as a potential contributory factor...

I guess if heat is the issue then it might be wax in the injectors causing problems, where the wax is softening with all that extra heat and impeding fuel flow (either by expanding slightly or by being pushed closer to the opening maybe), but not being softened enough on short/cold journeys.

It's not mileage related per-se. I don't think heat has anything to do with how well the injectors are flowing at different temperatures - there are two temperature related factors as I see it:

1) A cold engine doesn't knock / pre-ignite nearly as easily, as heat is a major factor in the development of knocking. So it's only once the engine has warmed up a bit that it will knock and start triggering the knock sensor. Air intake temperature is a factor in knocking too - even on a fully warmed up engine if the inlet air is very cold due to sub zero ambient temperatures that tends to suppress knocking as well, (one reason why turbo's often use inter-coolers) and as I've reported before the car runs much better and more consistently in very cold ambient conditions even after the engine has warmed up.

2) The ECU richen's the mixture up a lot on a cold engine, which suppresses the tendency to knock, and if there was otherwise a problem with knocking from a lean mixture the lean mixture is not present while the ECU is still enriching the mixture for a cold start. For the first 2-3 minutes it actually runs in open loop mode and runs rich, then even after it switches to closed loop mode and starts using the O2 sensor every time you accelerate and go into open loop mode it's running much richer than a fully warmed up engine would be. (Hence the horrible fuel consumption on a cold engine)

This is quite visible on the O2 reading on the scope - for the first few minutes after a cold morning start the O2 reading is nice and rich when accelerating hard, its only once the engine has started to warm up quite a bit that it THEN starts to have trouble running lean at WOT.

What I also notice is that how I drive after an ECU reset affects whether performance drops off quickly again or not. If I leave it in sport mode so it's tending to hold a lower gear, and I'm driving with a light throttle in a lower gear at higher rpm it stays very very responsive and zippy with no apparent loss in performance even if its driven for a long time. I can briefly test it at a lower rpm and it pulls very well as long as I don't try to use the higher gears too much.

On the other hand if I leave the gearbox in normal mode so it tries to lug at low rpm in higher gears all the time it quickly deteriorates. I think this is simply because its only running lean below 2500 rpm at load - if I keep the rpm up (and at a lighter engine load with a lighter throttle) the engine doesn't knock and the timing doesn't get retarded - I'm keeping the engine's operating conditions out of the "danger area" where it's having a tendency to run lean and knock. The knock retard table can only learn values for rpm/load combinations which you subject the engine to. If you never drive at WOT at 2000 rpm it can't learn any bad values there.

I also have strong suspicions about the ICV in a warm engine bay. If the lubrication is lacking in the two bottom bearings there's a risk that it isn't reacting smoothly enough (and don't forget this device is more important at low rpm/low throttle opening), which might cause an inconsistent mixture and subsequently all sorts of spurious readings at downstream sensors.

I can't remember whether I asked before, but have you cured the vacuum leak you had? Can you stall the engine at tickover by blocking the ICV/air filter hose? If not, then could it be that the warm engine bay situation is encouraging this leak (hot air being thinner, and possibly the location of the leak becoming larger with expansion).

I'm not sure that I do have a vacuum leak. I'm pretty sure that when I checked after I'd swapped the injectors over (and I'd added the hylomar blue to the inlet gasket etc) that it DID stall when I blocked the ICV air inlet, but I'm not sure if the engine was fully warmed up at the time, so I might have to re-do that test.

However a small vacuum leak is not going to cause the engine to run lean at WOT - as the throttle plate is already wide open under those conditions... a vacuum leak will only affect idle and light throttle openings.

[stevieb]

I'm going to start playing dumb to see if we can get to the bottom of this. Too many sensor readings and theoretical ideas are muddying the waters I think.

So AFTER an ECU reset (which is the only fair time to form an opinion as the ECU is having less of an influence) the lean mixture is more common at low rpm/high loads, and if we disregard the sensor readings, you've established that keeping rpm higher and load lower on a journey (keeping away from that lean mixture zone) gets you further before the ECU begins retarding the ignition, which we presume is happening to counteract the knocking caused/exaggerated by the lean-burn.

BUT booting the throttle to increase the load (increase in demand for both fuel and air) increases the likelihood of lean-running (or insufficient enrichment) and the knocking that comes with it, triggering the ECU to revert to a more retarded ignition state.

As you've pointed out before, the ECU will happily retard the ignition, but not advance it back again without a lot of persuasion. So it could be that a short-term problem is enough to trigger this change in the ECU, even if it only occurs under very specific conditions and passes fairly quickly. It could come and go, but then you're stuck with the retarded ignition state.

But how long does a problem have to exist for before the ECU takes action? Especially a difficult to reverse action like switching ignition tables? I just can't see there being a problem across enough injectors to create enough knocking for the ECU to retard the ignition like that. And if it was happening before, with a completely different set of injectors...

I'd still be suspicious about air supply/vacuum/vacuum readings. My failed MAP sensor gave the feeling of very retarded ignition (I had no means of measuring it, just going on experience from carburettor days), so although switching to the original Bosch one was better, the new one was better still. I agree about vacuum leaks not affecting the engine when the throttle plate opens, but there IS a moment there where a low rpm leak might mean there's a reduced vacuum in the manifold - just for a moment. Is this the reading you're seeing at the oxygen sensor, or is it longer-lasting?

And did we establish at what rpm's the ICV is supposed to open/close? If the throttle plate is opened beyond a certain point, I'm sure the ICV is supposed to close - so the ECU has a known volume of air passing the throttle plate at the given TPS position. If the ICV isn't closing fully, or is worn and therefore leaking, could it be that a greater volume of air is finding its way into the manifold than the ECU is banking on at that throttle position? Especially if signals from the MAP sensor aren't a part of the equation in these particular circumstances to help the ECU make sense of what's happening.

As a daft thing to try, have you tried a plain resistor across the inlet air temp sensor terminals to kid it that the air temp is constant? See what effect that has on the readings at the oxygen sensor. The logic is if the sensor is reading on the cold-side when you start up (and you say the problem has been more obvious on warmer days), then could it be that the ECU is assuming the density of the air passing the throttle plate is greater than it is, and therefore it's pegging the fuel back a tiny bit - causing the lean supply? In cooler air it's possible that the mixture problem is just very close to causing problems, but stays the right side due, as you say, to the fact that things are running rich anyway. When the air density drops on a warmer day (I'm sure it drops suddenly above a certain temperature), and especially on a warm start, the ECU thinks it's colder out there than it actually is and is under-fuelling as a result. Random I know, but the resistor trick might help take out that particular variable and let the real-world air environmental changes make a difference.

[Mandrake]

So AFTER an ECU reset (which is the only fair time to form an opinion as the ECU is having less of an influence) the lean mixture is more common at low rpm/high loads, and if we disregard the sensor readings, you've established that keeping rpm higher and load lower on a journey (keeping away from that lean mixture zone) gets you further before the ECU begins retarding the ignition, which we presume is happening to counteract the knocking caused/exaggerated by the lean-burn.

BUT booting the throttle to increase the load (increase in demand for both fuel and air) increases the likelihood of lean-running (or insufficient enrichment) and the knocking that comes with it, triggering the ECU to revert to a more retarded ignition state.

As you've pointed out before, the ECU will happily retard the ignition, but not advance it back again without a lot of persuasion. So it could be that a short-term problem is enough to trigger this change in the ECU, even if it only occurs under very specific conditions and passes fairly quickly. It could come and go, but then you're stuck with the retarded ignition state.

Mostly agree with the above.

But how long does a problem have to exist for before the ECU takes action? Especially a difficult to reverse action like switching ignition tables?

Knock retard in a particular rpm/load cell is applied almost instantly (fraction of a second) as knocking is considered to be potentially dangerous if its actually detonation rather than knocking. After the knocking passes timing is crept forward again slowly, so long as you stay in that rpm/load cell with no knocking occurring. What "slowly" means for this ECU I don't know - it could be seconds, minutes, hours, but its certainly not immediate.

A lot of knocking over a wide range of load/rpm cells will cause excessive timing corrections, this will trigger a change of octane map where all the base timings are reduced. Once this happens it can take hundreds of miles to go back to a higher octane map, or a battery off reset.

It's normal that if you run an engine like this on low octane fuel that once you fill up with high octane again you could go through a whole tank and several hundred miles without it switching back to the high octane map by itself. (one thing that people that say higher octane makes no difference don't take into account) I always do a battery off reset if I go from low octane petrol to high octane - I did the same thing on my previous 2 litre Xantia and yes it made a difference.

I just can't see there being a problem across enough injectors to create enough knocking for the ECU to retard the ignition like that. And if it was happening before, with a completely different set of injectors...

As far as I know on this engine it only takes one cylinder knocking to knock the timing back for all of them, and potentially trigger an octane map change. (Some high performance engines like turbo charged Subaru's are able to individually adjust the timing on each cylinder to squeeze every last drop of performance - easier on a 4 cylinder anyway as there is no combustion stroke overlap like there is on a 6 cylinder) So if you had one injector that flowed a lot less than the others that cylinder would run leaner and start knocking or lean misfiring before the rest. So it's not as if all the injectors would have to be low in flow rates to cause the symptoms.

True, I have a completely different set of injectors, but remember that they're second hand, unknown mileage, and three of them were stuck when I first got them - and although they passed my balance checks I'm not able to test their absolute flow rates. I've been bitten enough times already by second hand parts that aren't any better than the ones I already have. If anything the lean running has in fact got worse with the second hand set of injectors.

I agree about vacuum leaks not affecting the engine when the throttle plate opens, but there IS a moment there where a low rpm leak might mean there's a reduced vacuum in the manifold - just for a moment. Is this the reading you're seeing at the oxygen sensor, or is it longer-lasting?

It's long lasting. If I hold the engine at the right conditions - say 2000 rpm at 60% throttle in 3rd gear up a hill it will run flat line lean on the O2 sensor indefinitely. As soon as I ease off the throttle a bit it will either go rich or go back to closed loop mode where it switches lean/rich rapidly as it should.

And did we establish at what rpm's the ICV is supposed to open/close? If the throttle plate is opened beyond a certain point, I'm sure the ICV is supposed to close - so the ECU has a known volume of air passing the throttle plate at the given TPS position. If the ICV isn't closing fully, or is worn and therefore leaking, could it be that a greater volume of air is finding its way into the manifold than the ECU is banking on at that throttle position? Especially if signals from the MAP sensor aren't a part of the equation in these particular circumstances to help the ECU make sense of what's happening.

The ICV is supposed to open when you open the throttle and stay open under wide throttle acceleration. The reason for that is that it acts as an anti-stall damper like the dashpot on a carburettor. Otherwise when you take your foot off the throttle suddenly you'd get a whiplash inducing lurch. It stays open when you take your foot off the throttle to smooth the deceleration, its also what causes the rpm to linger slightly at 1500 when you snap the throttle then release it.

Remember that once manifold pressure is at maximum further throttle butterfly opening makes no difference. At low rpm this happens at surprisingly low throttle openings. For example at 2000 rpm anything beyond about 30-40% throttle opening is in fact wide open throttle as far as manifold pressure and air flow is concerned.

As a daft thing to try, have you tried a plain resistor across the inlet air temp sensor terminals to kid it that the air temp is constant? See what effect that has on the readings at the oxygen sensor. The logic is if the sensor is reading on the cold-side when you start up (and you say the problem has been more obvious on warmer days), then could it be that the ECU is assuming the density of the air passing the throttle plate is greater than it is, and therefore it's pegging the fuel back a tiny bit - causing the lean supply? In cooler air it's possible that the mixture problem is just very close to causing problems, but stays the right side due, as you say, to the fact that things are running rich anyway. When the air density drops on a warmer day (I'm sure it drops suddenly above a certain temperature), and especially on a warm start, the ECU thinks it's colder out there than it actually is and is under-fuelling as a result. Random I know, but the resistor trick might help take out that particular variable and let the real-world air environmental changes make a difference.

You're back to front there - warm air is less dense so requires less fuel, so a higher IAT reading would reduce injector pulse width. Besides, I've tested the inlet air temp sensor and it is reading correctly.

[Mandrake]

A small update.

Yesterday we went on a bit of a long drive and the car was performing really poorly, very lacking in low torque - on the scope it was clearly running very lean indeed in open loop - even putting my foot down a bit in 4th at 60mph it was going lean at nearly 3000 rpm. On the trip back performance wasn't nearly as bad but still showing very lean. So I ran a few tests last night to double check some things and make sure I hadn't missed anything before I get my original injectors cleaned.

I put the fuel pressure gauge on the rail for the first time since I fitted the new pump just to verify things were ok - pressure with the vacuum hose off is 2.8 bars, a bit lower than it should be, but that's because I have the fuel pressure regulator that came with the S/H rail fitted and I noticed when I fitted it it was slightly lower than the one I had in before which was 2.9 bars, but higher than my original which was 2.7 bars. Presumably its just the spring in the regulators that goes a bit soft with age...

During full wide open throttle snaps the pressure did not budge off 2.8 bars at all, suggesting no problems with fuel delivery. With the vacuum hose refitted pressure was about 2.2 bars at idle. I then did some tests to see what pressure the pump could produce into a load - basically pinching off the return rubber hose that runs over the top of the rear timing cover with some pliers. With the hose completely pinched the pump was able to produce 6.4 bars, which is a good healthy figure. Of course I didn't do that for too long.

As I'm considering fitting a 3.5 bar regulator as well as getting the injectors cleaned (belt and suspenders! Just in case a freer flowing exhaust is a contributing factor to lean running) by adjusting my grip on the return line I tried increasing the fuel pressure to various figures between 3 and 6 bars to see how much compensation range the ECU has in closed loop mode - and found that at least at idle it was able to compensate even for 6 bars - within about 20 seconds the mixture adjusted and was switching back and forth across stoichiometric again, (also no fault codes logged) so a small increase to 3.5 bars should be no problem at all. This will make it run a little bit richer than normal (about 8%) in open loop mode but make no difference in closed loop mode where the O2 sensor takes control.

What was also interesting was that you can temporarily make the mixture very rich with NO apparent problem with the idle, (I was expecting it to stumble a bit with the rich mixture) but make it too lean and you can easily get it to stall. I could have it idling at 2.2 bars, (vacuum hose on) pinch the pipe to increase the pressure almost immediately to 5-6 bars and it would keep idling happily and smoothly as if nothing had changed even though the mixture must be at least 40% richer at twice the pressure. (Flow change = sqrt(new pressure/old pressure) ) Hold that high pressure for about 20 seconds until the ECU fully compensated the injector times (O2 starts switching again) and then release it suddenly making it go 40% lean and it just immediately stalls.

Or if you do it a bit less severely - say raise it to 4 bars, let it idle until the ECU has compensated then let go it will almost stall, surge, almost stall, surge, misfiring slightly then gradually come right as the mixture is richened up again by the ECU. What was interesting was the stall/surge behaviour in this condition is EXACTLY the same as what it used to do on a hot start sometimes... and I'm wondering if those almost stall / surge cycles that occasionally occur on a hot start are in fact a mixture that it is too lean until the O2 sensor warms up and makes a correction to the mixture.

Furthermore I've noticed on a hot start if I watch the O2 sensor for the first 30 seconds it SHOULD read slightly rich until it starts switching, (the ECU should err on the side of being rich before the O2 sensor has warmed up, the engine will happily run rich but it won't run properly lean) but as the sensor warms up you can clearly see that it is indicating LEAN running after a hot start before the O2 sensor starts switching - meaning that in open loop mode its even running lean at idle.

Everything points towards lean running in open loop mode. I'm satisfied that the fuel pump is working properly and that whilst slightly low in pressure the pressure regulator is working fine. I also checked the rail is holding static pressure when the pump is off and its not dropping at all now. (The non return valve in the new pump works properly and the regulator has no leaks)

The small difference in pressure between 2.8 and 3.0 bars is not enough to account for such a severe lean condition (not to mention the car WAS running normally on the original 2.7 bar regulator for the first 6 months!) so as far as I can see it has to be the injector flow rates, and once again I've just swapped one set of old suspect parts for another set of old suspect parts!

So I do plan to go ahead with the injector clean and if the 3.5 bar regulator I have my eye on will fit I'm going to fit that as well to give me a bit more safety margin. (Better to err slightly rich than lean, especially when trying to eliminate throttle snap hesitation) Fingers crossed...

[CitroJim]

Excellent Simon, more very good diagnosis. Reading this is like reading a dissertation. You're well on the way to a first class degree in ES9 Diagnostics with all this stuff if not a Doctorate...

I mean it, it's bloody good stuff. You really know your stuff now...

[stevieb]

You're back to front there - warm air is less dense so requires less fuel, so a higher IAT reading would reduce injector pulse width. Besides, I've tested the inlet air temp sensor and it is reading correctly.

You're right, of course. Too many theories in my head and not enough sleep...

It's funny you mention the hunting at tickover because I occasionally had the same huntingr with my old injectors, but not very often at startup - it happened most often when I pulled up after a hot-engine run (being stuck in traffic for a period without a chance to blow out any soot afterwards). It tended to occur more frequently when I was doing a lot of short journeys and the fuel in the tank was regularly low, so I guessed it was just that - it began to ease off when I returned to regularly doing over 100 miles a week, and disappeared completely for a while after the Terraclean session before gradually returning (probably coinciding with that injector starting to tick again) - but it's virtually gone completely now the injectors have been changed, with just a very faint flutter now and again (which may be nothing more than the cooling fans firing-up or the air con clutch engaging and me not hearing the click). So maybe you're right about the injectors - I just got lucky (luckier) with mine than you have with yours.

As you've said though, at least you have a spare set now. And if you're sending off the better set, there's more chance they'll be successfully cleaned - the concern I had with getting my original, unhealthy, six cleaned was that I'd send them off and be told one or more were beyond repair or couldn't be balanced - so I'd still have to buy some spares anyway and then either replace the first six knowing at least one would have to come out again then come back and swap the remainder, or be without the car a lot longer. I wasn't willing to risk being either. And then I had Terraclean done and the idea went in the bin...

[stevieb]

I've had a couple of thoughts (yeah yeah, I know... )

Firstly, if I were you Simon, if you're planning on swapping out the pressure regulator anyway, I'd be doing that first, before spending anything on getting the injectors cleaned. After all, it's a job that can be done without removing anything else (or at least very little), and if it provides a cure then you've saved yourself a lot of extra work and cost.

Secondly, and partly related to the first idea, I've been thinking back to when I swapped out my pump. Two days before the new one arrived I decided I'd make sure the old one was freed-off - I'd read that the locking ring can be a pig, so I thought I'd do a bit of prep for the new one arriving. I released the fuel supply hose (with a couple of rags handy) and made a mental note of the pressure there, albeit I have no idea of actual figures. When I came to change the pump I decided I didn't want quite so much spillage so unplugged the electrical connector and ran the engine for a bit, hoping this would reduce the pressure in the hose and limit the spillage. The pressure in the hose was barely any less than it had been before - and judging by how much fuel pooled in the well in the middle of the pump, I'd say there was a similar amount that escaped. Both disconnections were done soon after long journeys, so I doubt there was any time for pressure leakage in either case.

What does this tell us? Well the engine ran to a stall, yet there was still probably 1.5 to 2.0 bar of pressure in the fuel rail. So how much pressure do the injectors require as a minimum to function and allow the engine to run? If we assume the engine stalls at 1.5 bar (a very conservative estimate, I'd be more inclined to say it's nearer 2.0 bar - after all, yours is running 2.2 bar at tickover) against a spec of 3.0 bar, then you could argue we have a working range of 1.5 bar.

If we then look at your old regulator opening at 2.7 bar - 0.3 bar below spec - that 0.3 bar is 20% of the working range. Which I would've thought is quite a lot to lose. That percentage is obviously even higher if the working range is less than 1.5 bar, and surely the greater that percentage, the greater the chance of problems.

My third thought needs an answer. Would incomplete combustion cause a lean reading at the oxygen sensor? If so, could the debatable fuel supply pressure cause a bad spray pattern with the right/wrong set of injectors? Low pressure plus an injector opening fully will surely be more dribble than spray...? And conversely, better pressure on a slightly bunged-up injector might give a very good spray pattern, albeit short on overall volume.

I don't know about this last idea - I'm just hypothesising on the way fluids work under pressure, and wondering what effect different variations in pressure and restrictions to flow will have, but I thought I'd throw it out there.

[Mandrake]

Excellent Simon, more very good diagnosis. Reading this is like reading a dissertation. You're well on the way to a first class degree in ES9 Diagnostics with all this stuff if not a Doctorate...

I mean it, it's bloody good stuff. You really know your stuff now...

Thanks Jim, but lets wait and see if I can get it fixed first, then I'll celebrate.

It has been a very good learning experience though, as frustrating as it has been to be learning on a shoestring budget on my daily driver, and all the stumbles and missteps along the way... I'll come out of it in the end knowing far more about how to diagnose an engine than when I started! (which was basically nothing)

A lot of the difficulty has been a lack of diagnostic information available from the ancient ECU on this engine - a more modern engine (even an ES9J4S) and I could have got fuel trim data from it with any OBD-II scanner in minutes which would immediately have shown strong positive fuel trim numbers and pointed the way towards a lean running diagnosis. No such data is to be had on this ECU, even with a Lexia. So I've been learning and working in the dark at the same time.

I had seen occasional signs of lean running by watching the O2 voltage reading on the Lexia while driving but just looking at a number is very difficult to judge compared to looking at a graph. It's not until I had my tiny scope sitting in the corner of the instrument panel monitoring the O2 sensor all the time while driving that I realised just how pervasive the lean running is and the penny dropped.

I still don't know 100% for sure why it's running lean in open loop mode (after cleaning the injectors it will either be fixed or another possible cause will be ruled out) but I do know 100% for sure that it should NOT be running lean under these conditions and that the lean indication on the scope correlates with the observed lack of power.

I'm also gratified that after more than a year of speculating that the timing was getting retarded below 2500 rpm that I have actual proof on the scope. (wish I'd taken some screen shots - hard to do when you're driving though!)

I did a back of the envelope calculation this morning to try to get a sense of how much the timing is being retarded, and the difference in timing shown on the scope at 2000 rpm between running poorly and what it was immediately after a battery off reset was about 1ms. At 2000 rpm that works out to 12 degrees! (If I did my sums right) So resetting the ECU was temporarily bringing the timing forward by about 12 degrees at 2000 rpm, no wonder it makes a difference as 12 degrees is a HUGE shift in timing. Retard the timing by 12 degrees at low rpm and it will absolutely kill the performance. (and MPG too)

So again, I can't be 100% certain why the timing is being retarded so much but I am 100% sure that it should not be and that it will be causing a big loss of power. The fact that the lean running seems to occur almost exclusively below 2500 rpm AND the timing advance suddenly drops dramatically back below 2500 rpm is not a coincidence I think.

12 degrees could well be the maximum available timing retard available in the knock retard tables - so if you could somehow look at those tables you'd see lots of zeros filled in above 2500 and lots of -12 filled in below 2500 with a fairly abrupt transition at about 2500 rpm. (The change in timing on the scope as you go either side of 2500 is very rapid)

So I know why its running poorly, (lean and retarded) fingers crossed it is the injectors that is the root cause, and not something like a valve gear fault...

[Mandrake]

It's funny you mention the hunting at tickover because I occasionally had the same huntingr with my old injectors, but not very often at startup - it happened most often when I pulled up after a hot-engine run (being stuck in traffic for a period without a chance to blow out any soot afterwards). It tended to occur more frequently when I was doing a lot of short journeys and the fuel in the tank was regularly low, so I guessed it was just that - it began to ease off when I returned to regularly doing over 100 miles a week, and disappeared completely for a while after the Terraclean session before gradually returning (probably coinciding with that injector starting to tick again) - but it's virtually gone completely now the injectors have been changed, with just a very faint flutter now and again (which may be nothing more than the cooling fans firing-up or the air con clutch engaging and me not hearing the click). So maybe you're right about the injectors - I just got lucky (luckier) with mine than you have with yours.

I know the symptom you speak of when pulling to a stop with a hot engine - I used to see that too, where as I slowed down to stop at a junction the rpm would dip too low then surge to recover. It occasionally still does that so it will be interesting to see if that goes away.

As you've said though, at least you have a spare set now. And if you're sending off the better set, there's more chance they'll be successfully cleaned - the concern I had with getting my original, unhealthy, six cleaned was that I'd send them off and be told one or more were beyond repair or couldn't be balanced - so I'd still have to buy some spares anyway and then either replace the first six knowing at least one would have to come out again then come back and swap the remainder, or be without the car a lot longer. I wasn't willing to risk being either. And then I had Terraclean done and the idea went in the bin...

Yeah, I was reluctant to send my only injectors away for exactly the same reasons - I'm much happier sending them away when the car is still running on the other set.

[Mandrake]

Firstly, if I were you Simon, if you're planning on swapping out the pressure regulator anyway, I'd be doing that first, before spending anything on getting the injectors cleaned. After all, it's a job that can be done without removing anything else (or at least very little), and if it provides a cure then you've saved yourself a lot of extra work and cost.

Well, changing the pressure regulator requires 80% of the work of changing the injectors - the manifold has to come out, the front fuel rail has to be lifted up, and its very difficult getting the regulator out with the rail only lifted up a bit - especially if its stuck in firmly as it was last time. Then there's the risk of breaking the breather tube again... I'd rather just do the job once and fit the regulator and injectors together. Belt and suspenders.

Secondly, and partly related to the first idea, I've been thinking back to when I swapped out my pump. Two days before the new one arrived I decided I'd make sure the old one was freed-off - I'd read that the locking ring can be a pig, so I thought I'd do a bit of prep for the new one arriving. I released the fuel supply hose (with a couple of rags handy) and made a mental note of the pressure there, albeit I have no idea of actual figures. When I came to change the pump I decided I didn't want quite so much spillage so unplugged the electrical connector and ran the engine for a bit, hoping this would reduce the pressure in the hose and limit the spillage. The pressure in the hose was barely any less than it had been before - and judging by how much fuel pooled in the well in the middle of the pump, I'd say there was a similar amount that escaped. Both disconnections were done soon after long journeys, so I doubt there was any time for pressure leakage in either case.

What does this tell us? Well the engine ran to a stall, yet there was still probably 1.5 to 2.0 bar of pressure in the fuel rail. So how much pressure do the injectors require as a minimum to function and allow the engine to run? If we assume the engine stalls at 1.5 bar (a very conservative estimate, I'd be more inclined to say it's nearer 2.0 bar - after all, yours is running 2.2 bar at tickover) against a spec of 3.0 bar, then you could argue we have a working range of 1.5 bar.

I can halt your line of speculation right there, because I have actual measurements. Running the engine to a stall by disabling the fuel pump is what I usually do before I do any work on the fuel system, and then usually crank the engine a few times as well. The engine will keep idling to a surprisingly low pressure - there is typically about 0.2 bars left in the rail after it stalls. Remember though that as soon as the pressure starts dropping it immediately goes lean so the ECU starts constantly trying to crank up the injection time to compensate - so it becomes a race between the falling fuel pressure and increasing injection time.

So no, there is no-where near 1.5 bars left after it stalls. After cranking a few times the gauge will read close to zero but there is still pressure in the system as I found out the hard way when I changed the fuel filter the other day and got a face full of petrol that kept spraying like a fire hydrant for 5 seconds even after I'd run the pressure down with the pump disabled and cranked it a few times. So even if the pressure is very low you'll still get a lot of petrol spurting out, probably because the fuel lines are so soft and have a lot of give.

If we then look at your old regulator opening at 2.7 bar - 0.3 bar below spec - that 0.3 bar is 20% of the working range. Which I would've thought is quite a lot to lose. That percentage is obviously even higher if the working range is less than 1.5 bar, and surely the greater that percentage, the greater the chance of problems.

The problem with your theory is that there isn't a "working range", these type of injectors don't have a minimum or maximum pressure that they'll operate at, its just the flow will vary with pressure and the ECU is calibrated for a certain pressure - all other things being equal. (There are some injectors that need a minimum pressure to open as they use the pressure to open them, but not these ones)

2.7 bars sounds quite low but that's only 10% low in pressure, and flow is roughly proportional to the square root of pressure change (not proportional to linear change) so that works out to a volume reduction of about 5.4% which is noticeable but not huge on its own, if everything else was operating perfectly. (Remember the car WAS running seemingly normally on the 2.7 bar regulator at first) +/- 10% is usually considered to be the acceptable range for fuel trim on modern ECU's, anything outside that and you're starting to look at trouble and beyond about 15-20% fault codes will get set, at least on more modern systems.

However flow reduction due to low pressure and due to restricted injectors is of course additive - if the pressure is a bit low and the injectors are a bit restricted the cumulative effect could be enough to cause trouble. When I changed my regulator from 2.7 bars to the 2.9 bar one with my original injectors it did run noticeably and more consistently better but it did NOT fix the problem entirely, it was a clue pointing in the right direction but not a cure by itself.

If I had to guess I would say that the net (injectors and fuel pressure) shortfall in flow rates would have to be in the order of 15 probably 20% to cause the severity of symptoms I'm seeing, with only 5% of that attributable to fuel pressure alone. So lets say that the injectors have a 15% flow deficiency due to deposit build up, and going from the currently fitted 2.8 bar regulator to 3.5 bar gives a further 11% flow increase, that's quite a big change in total and I'd be very surprised if it didn't make a huge difference if not solve the problem entirely.

If the injectors were restored to 100% normal flow the extra 8% flow from the higher pressure (3.0 to 3.5) would not be harmful and would be within acceptable spec, but if the injectors are still a wee bit marginal even after cleaning that extra 8% could be the difference between minor symptoms and no symptoms at all. That's the way I look at it.

My third thought needs an answer. Would incomplete combustion cause a lean reading at the oxygen sensor? If so, could the debatable fuel supply pressure cause a bad spray pattern with the right/wrong set of injectors? Low pressure plus an injector opening fully will surely be more dribble than spray...? And conversely, better pressure on a slightly bunged-up injector might give a very good spray pattern, albeit short on overall volume.

I don't know about this last idea - I'm just hypothesising on the way fluids work under pressure, and wondering what effect different variations in pressure and restrictions to flow will have, but I thought I'd throw it out there.

Yes from what I've read poor spray patterns will cause poor atomisation of the spray which can cause incomplete combustion and high HC readings, although I'm not sure if that would give a high O2 reading or not. I've also read that higher fuel pressure does tend to improve the spray pattern and atomise the fuel better so that may be a small side benefit. (Higher rail pressure would also reduce the tendency for the rail to boil - I am still noticing the rail boiling when the car sits after a long run, even though the fuel pressure isn't dropping)

I really do think, hope, that cleaning the injectors and bumping up the fuel pressure slightly will fix the remaining symptoms, including the high HC and CO readings. I guess we'll know in a few weeks...

[stevieb]

Well, changing the pressure regulator requires 80% of the work of changing the injectors - the manifold has to come out, the front fuel rail has to be lifted up, and its very difficult getting the regulator out with the rail only lifted up a bit - especially if its stuck in firmly as it was last time. Then there's the risk of breaking the breather tube again... I'd rather just do the job once and fit the regulator and injectors together. Belt and suspenders.

Bah, and there was me thinking it might be an easy job, seeing as the regulator is sat there staring me in the face when I open up the bonnet

I can halt your line of speculation right there, because I have actual measurements. Running the engine to a stall by disabling the fuel pump is what I usually do before I do any work on the fuel system, and then usually crank the engine a few times as well. The engine will keep idling to a surprisingly low pressure - there is typically about 0.2 bars left in the rail after it stalls. Remember though that as soon as the pressure starts dropping it immediately goes lean so the ECU starts constantly trying to crank up the injection time to compensate - so it becomes a race between the falling fuel pressure and increasing injection time.

So no, there is no-where near 1.5 bars left after it stalls. After cranking a few times the gauge will read close to zero but there is still pressure in the system as I found out the hard way when I changed the fuel filter the other day and got a face full of petrol that kept spraying like a fire hydrant for 5 seconds even after I'd run the pressure down with the pump disabled and cranked it a few times. So even if the pressure is very low you'll still get a lot of petrol spurting out, probably because the fuel lines are so soft and have a lot of give.

This brings up two questions. How the hell does the engine run right down to 0.2 bar? I wouldn't have thought that (or at least any less than 1.0 bar) was sufficient to push past an injector filter, let alone create a spray that could be ignited - and if as you say, the injectors are opening longer to allow more of this fuel to pass through them, then there's even less chance of a spray forming - it'd be more like a dribble, surely? And a dribble won't combust. My fuel hose had a LOT more than 0.2 bar resting in it when I disconnected my fuel pump - I would guess in old money it would've been around 15-20psi, but with no means of measuring it I can't be certain. Of course the clue here might be the fact that I'm talking about the pump and your readings are at the rail - is it possible you've got a blockage or a kink in the supply hose? I wouldn't imagine the hoses are prone to too much expansion either, as most of the length is stiff-walled plastic, but that makes them a lot more at risk of getting kinked or pinched.

The second question is a simple one. If the ECU opens the injectors for longer to compensate for perceived lean running (based on O2 readings), then why isn't yours? If all six injectors are seemingly working fine after an ECU reset (and I'm sure you'd have a few more erroneous readings on your scope right across the rpm range if they were inconsistent in any way), then why would they gradually decide they don't want to open that little bit longer to counter lean-running when it happens? I'd say it's more likely the injectors are doing their bit, but something else is amiss that deteriorates with use/heat/whatever and the injectors reach a point where they can do no more.

The problem with your theory is that there isn't a "working range", these type of injectors don't have a minimum or maximum pressure that they'll operate at, its just the flow will vary with pressure and the ECU is calibrated for a certain pressure - all other things being equal. (There are some injectors that need a minimum pressure to open as they use the pressure to open them, but not these ones)

2.7 bars sounds quite low but that's only 10% low in pressure, and flow is roughly proportional to the square root of pressure change (not proportional to linear change) so that works out to a volume reduction of about 5.4% which is noticeable but not huge on its own, if everything else was operating perfectly. (Remember the car WAS running seemingly normally on the 2.7 bar regulator at first) +/- 10% is usually considered to be the acceptable range for fuel trim on modern ECU's, anything outside that and you're starting to look at trouble and beyond about 15-20% fault codes will get set, at least on more modern systems.

However flow reduction due to low pressure and due to restricted injectors is of course additive - if the pressure is a bit low and the injectors are a bit restricted the cumulative effect could be enough to cause trouble. When I changed my regulator from 2.7 bars to the 2.9 bar one with my original injectors it did run noticeably and more consistently better but it did NOT fix the problem entirely, it was a clue pointing in the right direction but not a cure by itself.

If I had to guess I would say that the net (injectors and fuel pressure) shortfall in flow rates would have to be in the order of 15 probably 20% to cause the severity of symptoms I'm seeing, with only 5% of that attributable to fuel pressure alone. So lets say that the injectors have a 15% flow deficiency due to deposit build up, and going from the currently fitted 2.8 bar regulator to 3.5 bar gives a further 11% flow increase, that's quite a big change in total and I'd be very surprised if it didn't make a huge difference if not solve the problem entirely.

If the injectors were restored to 100% normal flow the extra 8% flow from the higher pressure (3.0 to 3.5) would not be harmful and would be within acceptable spec, but if the injectors are still a wee bit marginal even after cleaning that extra 8% could be the difference between minor symptoms and no symptoms at all. That's the way I look at it.

Yes from what I've read poor spray patterns will cause poor atomisation of the spray which can cause incomplete combustion and high HC readings, although I'm not sure if that would give a high O2 reading or not. I've also read that higher fuel pressure does tend to improve the spray pattern and atomise the fuel better so that may be a small side benefit. (Higher rail pressure would also reduce the tendency for the rail to boil - I am still noticing the rail boiling when the car sits after a long run, even though the fuel pressure isn't dropping)

I really do think, hope, that cleaning the injectors and bumping up the fuel pressure slightly will fix the remaining symptoms, including the high HC and CO readings. I guess we'll know in a few weeks...

I still believe the injectors have an unofficial "working range" (the pressure required to be effective as injectors rather than leaky taps), but we'll have to differ on that one. I've kinda covered the other bits above, but I've had a few other thoughts that might add or hinder the process...

Coming back to the knocking issue. Running 98RON on a 95RON fuel map will reduce the likelihood of knocking, yes? But would 98RON on a 95RON map also change the O2 reading at the sensor through the combustion being different? Assuming the answer to the latter is either no, or the O2 readings would reduce, then the question is do you get further on 98RON fuel on a 95RON map before the problem begins to show itself? Logic says that if there's less knocking to cause the ECU to peg back the advance/retard, then the problem will stay away for longer, assuming the knocking is happening across all six cylinders as would be the case with a general fuelling issue. Of course this could only be tested fairly after an ECU reset but with a little running on a 95RON map (to ensure it doesn't switch to 98 quite so eagerly), but needs to be done before the retardation starts to take place. However, if 98RON on a 95RON map still deteriorates the same way then it would point to one or two cylinders maybe having supply or combustion issues that are triggering the knock sensor, rather than a blanket problem across all six.

The next question is, if you reset the ECU with 98RON in the tank, so you're running with a 98RON map, do the same set of deteriorating readings show on the scope over time? Again, logic says with 98RON in the tank and a 98RON fuel map, then the readings should be identical (or at least proportionate according to the improvement gained from the greater octane) and deteriorate in the same way (regardless of octane, the engine will run just the safe side of knocking, yes? So the same amount of "problem" will push it into the knocking zone?).

The only reason I'm flagging all this up is that I remember you having a debate with yourself between two ECU's and the obvious performance difference between the two. I wonder whether there could be a single map in the ECU which is screwing things about. There are lots of hardware variables, but it seems odd that the lean running is there with two completely different sets of injectors (albeit slightly different) and the outcome of that lean-running is exactly the same. Something makes me think that if you could shake up the base points (like trying higher octane fuel), then it might help narrow things down a little.

[Mandrake]

Well, changing the pressure regulator requires 80% of the work of changing the injectors - the manifold has to come out, the front fuel rail has to be lifted up, and its very difficult getting the regulator out with the rail only lifted up a bit - especially if its stuck in firmly as it was last time. Then there's the risk of breaking the breather tube again... I'd rather just do the job once and fit the regulator and injectors together. Belt and suspenders.

Bah, and there was me thinking it might be an easy job, seeing as the regulator is sat there staring me in the face when I open up the bonnet

Are you sure you're not confusing the fuel pressure regulator and the pulsation dampener Steve ??!

The fuel pressure regulator definitely does not stare anyone in the face when you open the bonnet. Looking into the engine bay its on the front (closest to the radiator) rail on the left hand (cam belt) side where one of the fuel hoses comes in. You can't even see it without removing the upper manifold much less get to it.

After the manifold is out the front rail has to be unbolted and lifted up, and the regulator comes out downwards towards the bottom manifold. As its usually very tight its very difficult but not impossible to get it out without completely removing the rail, but the risk of damage to breather hoses is high trying to do it in place.

You might be thinking of the pulsation dampener which is near the throttle body on the end of the rear rail - it looks just like a pressure regulator but without a vacuum hose. I could easily see that being removable without taking the manifold off, but that's not the one we're after...

This brings up two questions. How the hell does the engine run right down to 0.2 bar? I wouldn't have thought that (or at least any less than 1.0 bar) was sufficient to push past an injector filter, let alone create a spray that could be ignited - and if as you say, the injectors are opening longer to allow more of this fuel to pass through them, then there's even less chance of a spray forming - it'd be more like a dribble, surely? And a dribble won't combust. My fuel hose had a LOT more than 0.2 bar resting in it when I disconnected my fuel pump - I would guess in old money it would've been around 15-20psi, but with no means of measuring it I can't be certain. Of course the clue here might be the fact that I'm talking about the pump and your readings are at the rail - is it possible you've got a blockage or a kink in the supply hose? I wouldn't imagine the hoses are prone to too much expansion either, as most of the length is stiff-walled plastic, but that makes them a lot more at risk of getting kinked or pinched.

Yes it's surprising how low it will go before it finally stalls, at idle - it runs on for at least 5 seconds after you pull the fuel pump fuse. If you don't have a fuel pressure gauge you can only guess what the pressure might be, or take my word for it.

The second question is a simple one. If the ECU opens the injectors for longer to compensate for perceived lean running (based on O2 readings), then why isn't yours? If all six injectors are seemingly working fine after an ECU reset (and I'm sure you'd have a few more erroneous readings on your scope right across the rpm range if they were inconsistent in any way), then why would they gradually decide they don't want to open that little bit longer to counter lean-running when it happens? I'd say it's more likely the injectors are doing their bit, but something else is amiss that deteriorates with use/heat/whatever and the injectors reach a point where they can do no more.

You need to understand the difference between closed loop operation and open loop, then you will have the answer to your question.

In closed loop mode the ECU is controlling the mixture based on feedback from the O2 sensor - the sensor says the mixture is rich, the ECU progressively keeps leans the mixture until the sensor switches to reading lean. The sensor says it is too lean, the ECU progressively richen's the mixture until the sensor switches to saying rich. Back and forth, back and forth, flipping back and forth across 14.7/1.

At idle you're in closed loop mode provided the engine is at least partially warmed up and has been running for at least 30 seconds. The moment the fuel pressure starts dropping the mixture goes lean and the ECU starts increasing the injector time to compensate. It probably can't keep up as it only changes the mixture at a certain speed, but it will keep increasing it so long as it reads lean, thus it postpones the engine stalling until there really isn't anything left.

In open loop mode the O2 sensor is not monitored. Injector times are calculated based on engine RPM, MAP reading, TPS, (mainly TPS movement not absolute position) coolant temperature, air intake temperature, and even knock sensor activity. The calculations are all made based on assumed engine characteristics of a new engine, that includes induction and exhaust flow rates, (which affect volumetric efficiency) fuel pressure, injector flow rates and so on.

A change to any of the assumed engine parameters will NOT be compensated for by the ECU in open loop mode. So if injector flow is low, it will simply run lean in open loop mode and it won't even know that it is as it doesn't monitor the O2 sensor. In closed loop mode it can make corrections for problems like injector flow.

There is no problem with the mixture in closed loop mode - it compensates for that. However any sudden throttle opening (including snapping the throttle at idle) throws the ECU instantly into open loop mode until conditions stabilise again. Also any engine load higher than a certain percentage (near WOT) puts you in open loop mode because the engine MUST run rich under heavy load/acceleration to avoid knocking, avoid flat spots/hesitation, and give the best power output, yet closed loop mode can only regulate the mixture to be stoichiometric, so can't be used at high load.

See the problem ? In high load conditions we can't use the O2 sensor as we need to run rich, but the ECU can only pre-calculate the mixture in open loop mode and doesn't know whether it has achieved the target mixture or not. It's all done by dead reckoning based on known parameters of a good engine.

Coming back to the knocking issue. Running 98RON on a 95RON fuel map will reduce the likelihood of knocking, yes? But would 98RON on a 95RON map also change the O2 reading at the sensor through the combustion being different?

No. The RON of the fuel won't have any effect on the O2 reading, that purely measures left over oxygen. If its indicating the mixture is lean that's independent of the RON of the fuel. Only the knock sensor provides any feedback that allows the ECU to lean the RON characteristics of the fuel.

Assuming the answer to the latter is either no, or the O2 readings would reduce, then the question is do you get further on 98RON fuel on a 95RON map before the problem begins to show itself? Logic says that if there's less knocking to cause the ECU to peg back the advance/retard, then the problem will stay away for longer, assuming the knocking is happening across all six cylinders as would be the case with a general fuelling issue. Of course this could only be tested fairly after an ECU reset but with a little running on a 95RON map (to ensure it doesn't switch to 98 quite so eagerly), but needs to be done before the retardation starts to take place. However, if 98RON on a 95RON map still deteriorates the same way then it would point to one or two cylinders maybe having supply or combustion issues that are triggering the knock sensor, rather than a blanket problem across all six.

I think we've already covered this before - the car runs a lot better on higher octane petrol at the moment, the retarding of the ignition happens much more readily with 95 RON. Mixture richness and RON both contribute to the maximum timing advance you can use - lower octane fuel causes knocking to occur more easily, but so does a lean mixture. A higher octane fuel can partially compensate for a lean mixture (in so far as knocking is concerned) while a richer mixture can partially compensate for lower octane. Lean mixture and low octane is worse still.

A lean mixture also increases the required spark voltage which increases the chance of an ignition misfire occurring - even if the ignition system is working normally.

The next question is, if you reset the ECU with 98RON in the tank, so you're running with a 98RON map, do the same set of deteriorating readings show on the scope over time? Again, logic says with 98RON in the tank and a 98RON fuel map, then the readings should be identical (or at least proportionate according to the improvement gained from the greater octane) and deteriorate in the same way (regardless of octane, the engine will run just the safe side of knocking, yes? So the same amount of "problem" will push it into the knocking zone?).

See here's the problem - we don't know what octane map the ECU has selected at any given time, nor do we know how many different maps it has. Only Bosch know the answer to that question. If the engine is not running properly due to lean running, it may not even be using the 98RON octane map despite running 98 fuel. Selection of octane map assumes a normally operating engine with a normal fuel mixture ratio.

If the engine is knocking due to a lean mixture its very likely (I'd say almost certain) that after a while it drops back to a lower octane map. And if it keeps knocking on that map it may drop back to an even lower (protective) octane map. I actually believe that the really severe version of the "invisible caravan" symptoms are a protective octane map being selected which has really retarded timing designed to protect the engine against some idiot putting really low octane fuel in it, like 91RON. If your lowest ECU octane map was for 95 RON and you put 91 RON in the car and drove it hard with a 10.5 to 1 compression ratio you'd very likely damage the engine due to overheating and detonation. So there has to be something it can do to protect itself in this situation - a failsafe map, if you will. I know I'd design one in if I was designing the ECU.

The only reason I'm flagging all this up is that I remember you having a debate with yourself between two ECU's and the obvious performance difference between the two. I wonder whether there could be a single map in the ECU which is screwing things about. There are lots of hardware variables, but it seems odd that the lean running is there with two completely different sets of injectors (albeit slightly different) and the outcome of that lean-running is exactly the same. Something makes me think that if you could shake up the base points (like trying higher octane fuel), then it might help narrow things down a little.

Yeah there were some noticeable differences between the ECU's but I think all it was was a slight difference in sensitivity to the knock sensor signal. David has driven the car which my old ECU's went into (which has now had the engine stripped for an Activa V6 conversion) and he reported no problems with performance at all.

When I had my scope connected to the knock sensor I noticed the performance was quite significantly better with less power fluctuation under load - the reason will be that the 1MOhm input impedance of the scope would have loaded the knock sensor signal down slightly effectively reducing the ECU's sensitivity to the knock signal. In reality the engine was still knocking but the ECU wasn't reacting quite so aggressively with the timing, so less power was lost. Not good for the engine though for anything other than a short test! The knock sensor has so far saved the engine from what could have been significant mechanical damage due to lean running...

By the way I noticed an interesting temperature related symptom yesterday - picking up the car on the way home we had our first really hot day (17 degrees and sunny all day, woohoo!) with the car parked in the sun all day. On the short drive home it was ok for the first minute or so then started to struggle and splutter - not only was power really poor it was actually misfiring and hesitating to the point of bucking! On the scope it was running lean, lean, lean! A lot more lean than it even has been recently.

I think I know why too - heat soak of the air inlet temperature sensor with the car sitting in the sun all day would have been giving an artificially elevated intake air temperature reading that would have taken a few minutes to come down to reality as cool air passed through the inlet. So you would have had the sensor temporarily reading maybe 10-15 degrees too high while the plastic housing cooled down, but cooler air immediately coming in.

A hot sensor reading would have caused the open loop mixture to be further leaned out as it assumes the air density is lower. After maybe 5-10 minutes driving the air flow through the inlet would equalise the temperature reading of the sensor and the actual air temperature of the inlet air flow.

The performance and misfiring was so bad that when we got home I went out for a quick drive around the block for 5-10 minutes to try to work out what was going on, and the symptoms gradually abated.

One thing I could try for a quick test is working out what resistor value corresponds to about zero degrees and substituting that for the sensor to see how the car drives - if my theory is right the mixture will be richer and the car will run much better for it, but of course its not a cure, just a way of temporarily tweaking the mixture to prove a point!

Does anyone know if the air intake sensor is a PTC or NTC and know what it's resistance range is ?

[stevieb]

By the way I noticed an interesting temperature related symptom yesterday - picking up the car on the way home we had our first really hot day (17 degrees and sunny all day, woohoo!) with the car parked in the sun all day. On the short drive home it was ok for the first minute or so then started to struggle and splutter - not only was power really poor it was actually misfiring and hesitating to the point of bucking! On the scope it was running lean, lean, lean! A lot more lean than it even has been recently.

I think I know why too - heat soak of the air inlet temperature sensor with the car sitting in the sun all day would have been giving an artificially elevated intake air temperature reading that would have taken a few minutes to come down to reality as cool air passed through the inlet. So you would have had the sensor temporarily reading maybe 10-15 degrees too high while the plastic housing cooled down, but cooler air immediately coming in.

A hot sensor reading would have caused the open loop mixture to be further leaned out as it assumes the air density is lower. After maybe 5-10 minutes driving the air flow through the inlet would equalise the temperature reading of the sensor and the actual air temperature of the inlet air flow.

The performance and misfiring was so bad that when we got home I went out for a quick drive around the block for 5-10 minutes to try to work out what was going on, and the symptoms gradually abated.

One thing I could try for a quick test is working out what resistor value corresponds to about zero degrees and substituting that for the sensor to see how the car drives - if my theory is right the mixture will be richer and the car will run much better for it, but of course its not a cure, just a way of temporarily tweaking the mixture to prove a point!

Does anyone know if the air intake sensor is a PTC or NTC and know what it's resistance range is ?

That is EXACTLY what was happening to mine this time last year. The car would run perfectly fine to work in the morning, then be sat in the sun all day and 100 yards out of the car park on my way home performance would drop like a bomb. I'd actually feel the power drop off, like the handbrake had been pulled on or the invisible caravan had been hitched-up. I'd always claimed the ignition was retarding a huge amount, and you've probably just confirmed it. It sounds like you're getting the point I was making earlier about the IAT sensor (though I made the point very badly).

I know you've said in the past that there was a difference in performance between setting off quickly after a start-up and setting off after a short delay (though you put that down to the oxygen sensor), but I'm sure you've also said that the IAT only influences the ECU for a set period after start-up or within a certain rpm-range. The thing I was getting at before was that if the ECU is fuelling based on an incorrect IAT reading, but just staying the right side of knocking, then there is no need for any change in the knock table used. But as the throttle is blipped and load is increased, it may be that only under that particular set of circumstances the fuelling strays from the safe side of knocking into the dangerous side. If the fuelling drifts over into the dangerous side at the exact moment the ECU stops listening to the O2 sensor (because it's assuming air density is different to what it actually is), then it won't know there's a lean condition and won't attempt to correct it (which is what you see on the O2 readings), and this causes the lean-burn knocking to begin (which IS being detected) so then the timing map changes to compensate.

I do know mine would sometimes audibly knock last year when I had the invisible caravan issue, suggesting it was a long way past anything the ECU could compensate for, but you haven't mentioned that you can hear any knocking. My knocking went completely when the MAP sensor was changed, though I bought a new IAT around the same time and that may have helped.

As usual though, I'm just throwing ideas around to see what sticks (this board needs a shrugging smiley).

If it's any use, a lot of the sensor ranges are listed in the document here - http://www.coupe406.com/documents/Tutor ... nOu6vX4Bwh" onclick="window.open(this.href);return false;