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

[addo]

Fitting kits are readily available, genuine and aftermarket. A hacksaw blade between the coils is how you dismantle the old set.

As a good example of amateur (no facilities) timing, last weekend I removed a Xantia auto box in six and a half hours working solo in the weather. So, with a prior session of "cracking" ball joints, dropping the exhaust and changing the shifter you should come close to a box swap in one weekend. With a semi-competent helper, it'd be a doddle.

[CitroJim]

Simon, delighted to hear the rapid changes of oil seem to have had some good results

You're definitely down on power though... If the cat joint is that bad have you considered to possibility it's leaking there and drawing in air and thus upsetting the oxygen sensor? The boom you speak of could be weak mixture...

I can't recall exactly now but getting at the joint with enough room to hacksaw off the bolts might be a bit tricky. I know on a TCT Activa it would be a right 'mare...

If I had the kit I'd go at them with an oxyacetylene cutter! It's what an exhaust place might do...

The 'dipping revs' after starting a warm engine is absolutely normal. they all do that!

I'd replace the MAP sensor and the Coolant Temperature Sensor now regardless and if that doesn't cure it, go for the cat. You can get a 'kitten' from online automotive for a good price and if the MOT is a PITA next time then just pop the old cat on for such purposes and then afterwards pop the kitten back on. The cat joint bolts will be easy enough to get apart if it's done regularly!

Also, the backbox, if it's ancient, might be bunged up if the internals have collapsed - another possible source of the boom perhaps.

Any resistance in the exhaust is not going to help...

Still, it's coming along and that's good

[Mandrake]

Hi Jim,

I took the car for a drive tonight for the first time since Sunday, it was well and truly iced in with 3 days worth of deposited hoar frost... it started first time in -4 degrees (so glad I did the starter relay mod a while back ) but I could tell the battery voltage was down a wee bit and it wasn't turning over as fast as normal...

No sign of the groaning noise even under heavy low rpm load when fully warmed up (the worst case scenario) and in fact the gearbox was generally well behaved apart from a couple of jolts from 2nd to 1st where I had slowed right down for a roundabout and it must have just been in the middle of changing from 2nd down to 1st as I suddenly put my foot down, resulting in a bit of an abrupt change down into 1st under heavy throttle.

Performance at low rpm was flat again as described - at higher revs above maybe 3000rpm everything seems perfectly normal, it seems to suddenly open up as the revs increase, a bit like a turbo kicking in.

Regarding an air leak on the input to the cat, wouldn't that cause a rich mixture as the oxygen sensor would be reading too lean leading the ECU to richen the mixture ? Either way I don't think there is a leak there, and the oxygen sensor is still disconnected anyway. I tried reconnecting it for a while on Sunday it didn't make much difference to be honest, despite confirming with the Lexia that it was cycling up and down properly.

Regarding dipping revs - when I first got the car it did do that, but for many months it stopped doing it, recently its started it again...I wonder if its temperature related as it didn't seem to do it through the summer.

Possibly related - in the last couple of months I've noticed that as I slow down and stop, instead of the revs coming gently to rest at normal idle the revs always undershoot the idle revs just as I come to a complete stop such that it almost tries to stall, staggers, and then recovers again. It does it more or less every time I come to a stop. Any thoughts on that ?

I'm thinking MAP sensor might be a good next choice to replace, (and dead easy to do, even in the middle of winter) I've got a feeling that it could be related to both the lack of power low down and also the dipping revs when starting the engine or pulling to a stop. The symptoms are very much like those I observed when I once left the MAP sensor accidentally disconnected, except not as severe.

Coolant temperature sensor I'm not so sure about - every time I've checked it it seems to be working, however I will say one thing - it seems to read too high, I see readings as high as 97-99 degrees on the Lexia immediately after coming back from a moderately speedy jaunt on a country road, however I don't believe this is correct because I have the pressure cap loose on the expansion tank (on purpose) and not only is the water not bubbling or agitated, there is barely even any steam coming off it - I would have thought at 99 degrees there would be obvious bubbling in the expansion tank when running unpressurised but there is none at all.

What are the normal operating temperatures for the V6, both dashboard indicated and Lexia indicated ? The dashboard gauge typically says around 80-85, although its hard to be precise. If we assume the ECU temperature sensor is reading too high at the hot end of the range, what impact would this have ? Does the ECU have any "throttle the power back if overheating" safeguard ?

The reason I ask is that the "torque reduction" retard is often active when I think it shouldn't be - I would have thought that it only comes into play when the gearbox commands reduced torque during gear changes etc, but it seems to be frequently active (applying as much as 15-20 degrees of extra retard against the base advance figure) even during normal driving...something causing the ECU to apply this much timing retard would kill the performance quite badly I'd imagine.

Your comments about the back box have made me realise that I haven't even considered that a disintegrating back box or centre silencer could be causing a blockage due to collapsed partitions... in some ways that could be more likely than a fault cat (and a lot cheaper and easier to do something about) as both silencer boxes are very old and rather rusty... food for thought. Pity there isn't a way they could be x-rayed.

[CitroJim]

Simon, my brain is about to go on strike so I'm going to leave a proper reply until tomorrow if that Ok?

[addo]

A blocked exhaust causes incorrect MAP readings for the engine's operating state.

[Mandrake]

Simon, my brain is about to go on strike so I'm going to leave a proper reply until tomorrow if that Ok?

Of course Jim, I'm not expecting you to carry the forum (or this thread for that matter) on your two shoulders.

[CitroJim]

Sorry Simon, I'm late replying. Blame my brain - it's still not firing on all cylinders...

It's why most of my posts have been one-liners recently

Anyway...

Best I can do is this:

The revs dipping below idle as you slow down and stop is definitely wrong. To check this is not the TC staying locked up knock the gearbox into N before coming to a halt and see if it does the same thing...

If this is not conclusive then check the calibration of your throttle pot and check its actually working. Easy to do. Firstly, turn the ignition on and leave it on for 30s without starting up and without touching the throttle. Then, with the Lexia and with engine stopped, ensure that the throttle position is shown as 'IDLE' when the throttle is at rest and goes through (I think - from memory) 'MID-POINT' and 'FULL' respectively.

Cat leak. Yes, you're correct - it would run rich...

Temperature. Remember the coolant temperature sensor (which the Lexia 'sees' is in the top hose effectively so will record the hottest coolant. The gauge sender is effectively in the bottom hose so what the gauge sees will be lower. 97-99 does seem hot even for the top hose as that's a full 10 degrees above stat opening and at that temperature the fans should be cutting in to bring it down. Definitely worth swapping that sensor just to rule it out. Normal operating temperature should be around 85-88 degrees (stat temperature). On whether or not the ECU does anything clever when it detects overheating I don't know but the gearbox does. It invokes a special cool-down map. It has to be jolly hot to do so though.

What is the MAP reading with the engine off but ignition on? It should be close to prevailing atmospheric. On idle it should be around 350 mb or thereabouts. But again, the MK2 MAP sensor is very cheap and worth a swap just to see.

I've no idea when torque reduction is active or if it can be seen in operation on the Lexia. I just can't recall and would have to take mine for a spin to check.

As for the back box. you could always disconnect it for a test and see if performance is (noisily) improved!

Sorry Simon, that for now is the best I can do...

If you want me to do any comparisons against mine with the Lexia then I would be able to and happy to as soon as circumstances permit.

[Mandrake]

The revs dipping below idle as you slow down and stop is definitely wrong. To check this is not the TC staying locked up knock the gearbox into N before coming to a halt and see if it does the same thing...

The torque converter lockup not releasing soon enough while coming to a stop was my first thought as well, for some reason I haven't tried coasting to a stop in neutral, I really ought to try that. The other possibility might be that the change from 2nd to 1st that occurs as you stop is too slow.

If this is not conclusive then check the calibration of your throttle pot and check its actually working. Easy to do. Firstly, turn the ignition on and leave it on for 30s without starting up and without touching the throttle. Then, with the Lexia and with engine stopped, ensure that the throttle position is shown as 'IDLE' when the throttle is at rest and goes through (I think - from memory) 'MID-POINT' and 'FULL' respectively.

I've checked this before and it definately says idle when the throttle is not pressed, as far as I know the butterfly sensor is calibrated.

Temperature. Remember the coolant temperature sensor (which the Lexia 'sees' is in the top hose effectively so will record the hottest coolant. The gauge sender is effectively in the bottom hose so what the gauge sees will be lower. 97-99 does seem hot even for the top hose as that's a full 10 degrees above stat opening and at that temperature the fans should be cutting in to bring it down. Definitely worth swapping that sensor just to rule it out. Normal operating temperature should be around 85-88 degrees (stat temperature).

I'll have to check more carefully, but I believe the fans don't come on until the temperature indicated by the Lexia is 99 degrees, which does seem a bit high even for the top hose. As I say, at an indicated temperature of 99 degrees the water in the expansion bottle is not even agitated, it just has a little bit of vapour rising from it but is otherwise calm, so whilst working it does seem to be indicating a bit on the high side.

A new water temperature sensor looks like around £16, so its a no brainer to replace, and doesn't look too hard to do. If I put the front of the car up on ramps to tilt the engine back I shouldn't loose too much coolant either.

On whether or not the ECU does anything clever when it detects overheating I don't know but the gearbox does. It invokes a special cool-down map. It has to be jolly hot to do so though.

The highest I've seen on the gearbox is 105 after giving it real stick on country roads, which according to the docs is below the cool-down map threshold, so I don't think the gearbox is overheating.

What is the MAP reading with the engine off but ignition on? It should be close to prevailing atmospheric. On idle it should be around 350 mb or thereabouts. But again, the MK2 MAP sensor is very cheap and worth a swap just to see.

The ones I've seen are around £62, so not exactly cheap, with the engine off its around 980mb, when idling around 380mb. I don't think its a question of whether its working, but rather is it accurate, and is it consistent. If its "working" but sometimes its inaccurate, it could easily be the intermittent problem I'm chasing. Without knowing more about how the sensor actually works its hard to say what the likely failure mode is and whether intermittent inaccurate readings or "sticking" are possible.

I've no idea when torque reduction is active or if it can be seen in operation on the Lexia. I just can't recall and would have to take mine for a spin to check.

Yep, you can watch it on the Lexia under Measure parameters in the ignition section. There is a base advance figure which then has a "torque reduction" figure subtracted from it to give the actual timing applied to the engine. The torque reduction figure is 0 degrees when its not being used thus base advance equals actual advance. An example of it reducing torque would be a base advance of 30 degrees with a torque reduction figure of 15 degrees resulting in an actual advance of 15 degrees.

[CitroJim]

Simon, if you know where to look, a MAP sensor is quite cheap.

The part number for cars since RP 8078 is 1920 AN. Before that 1920 9H

here's a reasonable 'AN' on the 'bay...

This is a '9H' just a bit dearer from the 'bay

When should torque reduction by retard be Activa Simon and when have you seen it in operation? If you can tell me then perhaps I can do a road test and see if mine is similar...

[Mandrake]

Second attempt to reply, a browser crash lost my whole post....

The revs dipping below idle as you slow down and stop is definitely wrong. To check this is not the TC staying locked up knock the gearbox into N before coming to a halt and see if it does the same thing...

Torque converter lockup not releasing quick enough was my first thought too, that or changing from 2nd to 1st too slowly, which is normally done just before coming to a stop. For some reason I haven't got around to try coasting to a stop in neutral, I'll give that a try.

If this is not conclusive then check the calibration of your throttle pot and check its actually working. Easy to do. Firstly, turn the ignition on and leave it on for 30s without starting up and without touching the throttle. Then, with the Lexia and with engine stopped, ensure that the throttle position is shown as 'IDLE' when the throttle is at rest and goes through (I think - from memory) 'MID-POINT' and 'FULL' respectively.

I've checked the butterfly calibration before and it seems to be fine, IDLE, MID and FULL are all indicated correctly.

Temperature. Remember the coolant temperature sensor (which the Lexia 'sees' is in the top hose effectively so will record the hottest coolant. The gauge sender is effectively in the bottom hose so what the gauge sees will be lower. 97-99 does seem hot even for the top hose as that's a full 10 degrees above stat opening and at that temperature the fans should be cutting in to bring it down. Definitely worth swapping that sensor just to rule it out. Normal operating temperature should be around 85-88 degrees (stat temperature).

I'll need to check it more carefully but I think the fans are not cutting in until the top hose reading from the Lexia is reaching 99 degrees, which does seem too high. However at the same time the water in the expansion tank only has a bit of vapour coming from it but is otherwise calm and bubble free, so I find it hard to believe that even the top hose is hitting 99 degrees.

There is no sign of the engine actually overheating and a replacement sensor is about £16 so I think its a no brainer to replace it, and it doesn't look too difficult either. If I put the front of the car up on ramps to tilt the engine back and also drain the expansion tank I shouldn't loose too much coolant either.

On whether or not the ECU does anything clever when it detects overheating I don't know but the gearbox does. It invokes a special cool-down map. It has to be jolly hot to do so though.

The highest I've ever seen the gearbox reach with vigorous country road driving is 105 degrees, which is still below the cool-down map threshold so I don't think there are any problems with the gearbox overheating.

What is the MAP reading with the engine off but ignition on? It should be close to prevailing atmospheric. On idle it should be around 350 mb or thereabouts. But again, the MK2 MAP sensor is very cheap and worth a swap just to see.

The MAP sensors I've seen are about £63 so I'm not sure I'd call that cheap. With the engine off the reading is around 980mb, idling around 380mb, so the sensor is "working", at least when I have had the Lexia on the car, however the point is whether its accurate and consistent or not. If the reading is sometimes inaccurate and other times ok it could easily explain the intermittent poor running. Without knowing the operating principle and construction of the sensor its hard to say whether such an intermittent problem is possible or not, but my guess is that it is possible if the sensor is old and flakey.

I've no idea when torque reduction is active or if it can be seen in operation on the Lexia. I just can't recall and would have to take mine for a spin to check.

The torque reduction can be seen in the ignition section of parameter measurement. There is a base advance figure which is calculated based on rpm, load, throttle position and so on as per normal, then there is a torque reduction figure which is subtracted from that to give the actual actual advance figure that is used, all three are displayed.

Sometimes the torque reduction figure is 0 which means base advance = actual advance. Other times you might see something like base advance 30 degrees, torque reduction 20 degrees, actual advance 10 degrees, in which case you can see the advance is only 10 degrees but would have been 30 degrees if the ECU wasn't throttling back the torque.

We know for sure that the gearbox can command torque reduction for example during gear changes (the change in engine note during a gear change with heavy throttle is quite obvious) but what we don't know as its not documented is whether the gearbox might command torque reduction for any other reason than gear-changes, for example gearbox overheating or clutch slip, and whether the engine ECU itself might also apply torque reduction to counter overheating or other perceived running problems...

[Mandrake]

Drat!! Just spent time retyping my post only to find that the first one did go through despite the browser crash...

Simon, if you know where to look, a MAP sensor is quite cheap.

The part number for cars since RP 8078 is 1920 AN. Before that 1920 9H

here's a reasonable 'AN' on the 'bay...

This is a '9H' just a bit dearer from the 'bay

Ah thats more like it, £20 I can stomach My RPO is 07905 so I guess I need the 9H version. I wonder what's different about them...

When should torque reduction by retard be Activa Simon and when have you seen it in operation? If you can tell me then perhaps I can do a road test and see if mine is similar...

I see it in operation even when idling or blipping the throttle, as well as when driving, it seems to vary constantly. I was under the impression from the documentation that it should only be used during gear changes...

I'll see if I can talk my "co-pilot" into watching the laptop during a test drive on the weekend to get a better idea of when the torque reduction is in effect and whether theres any correlation with the poor performance.

[CitroJim]

Ahh, that's better Simon, the last paragraph makes more sense now The first post was there but a bit was missing!

I'll have to see what mine does/says on the torque reduction front and get back to you. Should have a chance over the weekend.

The difference in the MAP sensors is the plug. The early (9H) ones have a square sided plug similar in style to say the one on the throttle pot on the throttle body or the coolant temperature sensor and the later (AN) one has a curvy sided plug with small round pins. In fact the pictures in the links show the plugs correctly for each variant.

[xantia_v6]

re the rpm dropping after coming to a halt, R252DVC would do that occasionally, after a 2 hour non-stop run at 90 MPH it even stalled, so this might be a quirk of the auto-adaptive idle speed control (my current V6 does not have this behaviour).

If you could remove the oxygen sensor from the pipe, I imagine that a blocked exhaust would become evident through the release of pressure?

[addo]

I welded in a 1/8" NPT female bung for pressure gauge readings between fanimold and cat with a cheap boost gauge and capillary tube, however latent heat kept melting the tube connection (soft solder). Brazing would have worked better.

[Mandrake]

re the rpm dropping after coming to a halt, R252DVC would do that occasionally, after a 2 hour non-stop run at 90 MPH it even stalled, so this might be a quirk of the auto-adaptive idle speed control (my current V6 does not have this behaviour).

Back in March when it was cold I did have mine stall a few times soon after starting due to the revs falling too low and not quite recovering, then through the summer it stopped doing it completely for many months.

Can you elaborate a bit on the "auto-adaptive idle speed control", is this a feature of the engine ECU or the gearbox or a bit of both ?

If you could remove the oxygen sensor from the pipe, I imagine that a blocked exhaust would become evident through the release of pressure?

I would imagine so, however without another to compare it against, I wouldn't know how much pressure through the hole would be normal...