Exhaust gas leak. Suspect manifold

[Andy 105 1 Cars]

Morning,

 exhaust leak will not help, but I would be surprised if it caused quite the problems you are facing. Pretty sure that the R32 would need 6 ignition drivers so Emerald would have reconfigured to accommodate . Converting back should not be the issue . Your laptop disconnecting is a headache . Presume you will do as planned and give Emerald a call. They may have met this issue before . You are being ultra thorough. Best of luck. I really hope you manage to find the problem 

[Thijs_Rallye 87 4 Cars]

Allright, so no wiring issues, that's clear.

 

When does the smell become apparant? Immediately after starting or does the engine have to warm up first? A wrongly mapped of placed IAT (intake air temp) could cause strange issues as well. I assume the engine is batch fired because of the lack of a cam sensor, how many times do the injectors fire per engine revolution? Could fuel puddling be an issue?

 

Tbh, I still find it hard to believe exhaust gas coming into the cabin causes a clear petrol smell. But never say never .

[Telf 224 2 Cars]

4 hours ago, Thijs_Rallye said:

 

Tbh, I still find it hard to believe exhaust gas coming into the cabin causes a clear petrol smell

So if the car is over fuelling and not burning the fuel how would the emissions not smell of petrol?

[jackherer 543]

OK some updates;

 

I have not called Emerald yet, I'm trying to make sure I have answers for any questions they will have when they ask them.

 

Interestingly Paul's ECU runs absolutely perfectly in my car. I can connect to it with the laptop with the engine running indefinitely.

 

However it took a while to get it to run...

 

After triple checking the pin outs between the M3D and K3 I plugged the K3 into my loom and installed my map. Then I calibrated the TPS and inverted its operation as it is on the opposite end of the throttle shaft on my Jenveys and turns the other way.

 

Then I tried to fire the engine but it just coughed/spluttered/backfired. And incidentally made the whole area (inside the car, the driveway, in the garage, in the house) absolutely stink of petrol/exhaust fumes. As Paul said, how could it not?

 

I then compared settings between the two and noticed that Paul's ECU was set for sequential injection to four individual driver pins so I changed it to group and it ran. Badly but that was because it had fouled the plugs, new plugs and it was running on Paul's ECU just the same as it runs on mine.

 

This is potentially significant because Paul's car obviously isn't set up to run sequential injection. The cam sensor was disabled in the software so it can't have been trying to run proper sequential injection but the injector driver settings are obviously having a large effect as my car wont run when it is active.

 

Both of our looms are wired to run the injectors in 2 groups of 2. But Paul's ECU thinks it is triggering them individually so I guess it is pulsing two of them when it thinks it is pulsing one and at other times it is pulsing none when it thinks it is pulsing one. I wonder if the timing of the pulses is resulting in the bad mixture at certain RPMs?

 

I still don't get how it works at Emerald and changes afterwards.

 

I also don't know why it is disconnecting the laptop only in Paul's car.

 

BTW @Telf It's hard to concentrate with all these bloody military helicopters flying over but I have put your ECU back and reinstalled your map and undone the config changes I made, it is running your engine again just like it was before. Complete with over fuelling and laptop disconnecting.....

 

FWIW I tried running Paul's car with my wasted spark coil pack and crank sensor in case there was some sort of noise issue with either of them but they made no difference.

[jackherer 543]

5 hours ago, Thijs_Rallye said:

A wrongly mapped of placed IAT (intake air temp) could cause strange issues as well.

The IAT is actually mounted in a Jetronic air flow meter that is only fitted for cosmetic purposes.

 

5 hours ago, Thijs_Rallye said:

I assume the engine is batch fired because of the lack of a cam sensor, how many times do the injectors fire per engine revolution? Could fuel puddling be an issue?

Maybe, see my comments above, it wasn't set to batch fire despite being wired that way. (Actually group fire, two groups of two, but definitely not sequential)

[Telf 224 2 Cars]

Kieran regarding the sequential firing/batch firing of the injectors- surely this is a major issue? 

 

Also I know the car appeared to be ok at emerald but I wonder if they would have noticed the problem. Let's say they set the initial map at idle and it appears ok.. they pretty much went straight into higher rpm type runs, maybe it wasn't noticeable as the car was rolling, the rpm constantly changing and the general fumy environment of the workshop. I can't account for the laptop disconnect though

[Telf 224 2 Cars]

27 minutes ago, jackherer said:

of 2. But Paul's ECU thinks it is triggering them individually so I guess it is pulsing two of them when it thinks it is pulsing one and at other times it is pulsing none when it thinks it is pulsing one. I wonder if the timing of the pulses is resulting in the bad mixture at certain RPMs?

 

Can you change the pulse configuration and see if that changes anything on my car?

[jackherer 543]

6 minutes ago, Telf said:

Also I know the car appeared to be ok at emerald but I wonder if they would have noticed the problem.

I've been thinking that, a rolling road doesn't simulate actual low speed driving very well I don't think.

 

3 minutes ago, Telf said:

Can you change the pulse configuration and see if that changes anything on my car?

I think it would change it drastically if I changed it from sequential to grouped so given that it actually runs very well for most of the map on the current setting I don't think it would even run if I changed it.

 

I'll phone Emerald shortly...

[Telf 224 2 Cars]

16 minutes ago, jackherer said:

 

I'll phone Emerald shortly...

Mate,

 

It seems to me if the injectors have been set to the wrong type of firing sequence ( if I am correctly understanding what you are saying) then Emerald wouldn't have a cat in hell's chance of mapping the car properly? If that's the case they need to redo it right?

[Thijs_Rallye 87 4 Cars]

2 hours ago, Telf said:

So if the car is over fuelling and not burning the fuel how would the emissions not smell of petrol? 

Huge HC emissions at the tail pipe for sure, but an actual petrol smell in the car I find hard to believe. That would mean that exhaust gases are entering the cabin somewhere.

 

It is very hard to describe a smell on a forum , but the difference between partly burnt petrol and pure petrol must be distinguishable.

 

edit 1:

But I still do not know if you smell this immediately after starting or when the engine has been warmed up. (or both)

 

edit: 2

The injector firing doesn't matter that much, it still can be setup properly if they fire the two injector pairs at the same time. I mean the standard managment fires them all four at the same time, four times per 2 crank revolutions.

 

Only thing I can imagine is that since the injection charge is double as big, with less time to evaporate you will get bigger puddles, which could lead to bore wash, but that is all speculation and cannot be determined from here behind my keyboard. I reckon the oil must smell very "petrolly" as well then.

 

edit 3:

@jackherer IAT position is ok then, I was thinking possibly heat sock which leads to (too much) enrichment of the mixture.  Is it possible for you to have a drive around with your wideband in the tailpipe?

Edited September 16, 2019 by Thijs_Rallye
Lots of forgotten stuff

[Telf 224 2 Cars]

11 minutes ago, Thijs_Rallye said:

edit 3:

@jackherer Is it possible for you to have a drive around with your wideband in the tailpipe?

Edited 2 minutes ago by Thijs_Rallye

Thijs, if you read the thread several pages ago you will see this has already been done, also will answer your question about when it appears to overfuel

[jackherer 543]

I've just got off the phone with Dave Walker who agrees something weird is going on...

 

He has offered to look at it again, @Telf phone them now and make an appointment.

 

Basically for every suggestion he asked me to try I have already done it.

 

Regarding the sequential/grouped setting he didn't think it would make any difference because it was mapped with it set (and wired) that way.

 

I'm going there with you again and this time I am taking my laptop and wideband with.

[Thijs_Rallye 87 4 Cars]

17 minutes ago, Telf said:

Thijs, if you read the thread several pages ago you will see this has already been done, also will answer your question about when it appears to overfuel 

I did follow the thread but I am not able to remember it all appearantly , just trying to help. I just have re-read the whole thread, and it still isn't clear to me if it is a pure petrol smell or a "rich" mixture smell.

 

Anyways, did you attend the mapping session? Did they do full rev band pulls or did they do steady state measurements as well? Alpha N isn't ideal for a street car, if setup like a proper Alpha N it is all perfect at WOT, but pig rich part throttle unless there is somekind of compensation mechanism as lambda feedback.

Edited September 16, 2019 by Thijs_Rallye

[jackherer 543]

It's a rich mixture smell, not a raw fuel smell like a leak. The oil doesn't seem to smell of fuel.

 

I was there for both mapping sessions, they spend a long time on steady state, gradually building up through the day.

 

My 205 (1.9 Mi16 and ITBs) was mapped by Emerald as Alpha-N too and is absolutely fine on the road.

 

We did connect a MAP sensor but Emerald didn't use it for reasons I still am unsure of.

 

It does seem to over fuel while driving, between 1500-2500 RPM, just like when it is static. I would like to look at the maps while it is doing it but as stated the laptop disconnects when the engine is running. It seems to do that at any load I can achieve at those RPMS, obviously it's impossible to hold it there for long without a rolling road.

 

I'm sure the laptop disconnecting is a clue, it must be due to electrical noise.

 

Amongst many other things I have probably forgotten to mention in this thread is that running with the alternator belt removed makes no difference to either issue (fuelling or disconnecting) so I'm running out of sources to check for noise.

[Telf 224 2 Cars]

2 hours ago, jackherer said:

Regarding the sequential/grouped setting he didn't think it would make any difference because it was mapped with it set (and wired) that way

Kieran I'm not sure about this. As you have said the injectors are wired as a pair . They aren't sequential in anyway. I mean if the ECU is trying to fire sequentially it seems to me it's going to miss a cylinder every so often . Surely this would account for the fuel smell due to non detonation and therefore the lamda being all over the shop.  If I understand the meaning in terms of the management then sequential the ECU is triggering the correct injector/coil signal so 4-2-3-1 or whatever but if the injectors are wired as a pair which we know they are then the injectors must be triggered as a pair .  Dunno how that effects the coil but as it's a wasted spark then it occurs to me that the coil is also wasting one 'spark' . I don't know how that would interact with a sequential injector sequence when they are paired as per the diagram.

 

That probably makes no sense but I'm trying to puzzle it out

[Telf 224 2 Cars]

So looking at the diagram my loom is built exactly like yours. Do I need 4 separate injector signal lines from the ECU? Does our fuel system differ between the MI16 and the 8v ( as in mine is a common fuel rail ). I'm wondering if I've blindly followed your diagram and missed a vital point.

[jackherer 543]

When your ECU thinks it's triggering the injector for cyl 1 it's actually triggering cyls 1&4, then when it thinks it's triggering cyl 2 it's actually triggering cyls 2&3. And while it thinks it is triggering 3 and 4 there is nothing connected. If you consider that the standard Jetronic injection actually just fires all 4 at once I don't think it actually makes a huge difference but the bottom line is it was mapped that way so it must have fuelled OK with those settings while at Emerald.

 

The coils wasted spark occurs during the cylinders exhaust stroke so there is nothing left to ignite.

 

No, yours is wired correctly, you only wire the injectors individually if you are running sequential injection and that requires a cam position sensor which isn't feasible in your engine or mine.

 

Our fuel systems are basically the same, the only real differences between your engine and mine is 8 extra valves and 3 extra throttles.

[Telf 224 2 Cars]

Ok so post our chat earlier regarding the whole sequential versus group option for the injectors. If the map is set to sequential as you noticed it clearly says in the Emerald manual 'sequential fuel injection operation will only synchronise correctly when a cam phase sensor is fitted' 

 

I've also searched a bit more online and there are a few mentions in various tuning forums that state sequential injection is only really relevant at low rpm (below 2500rpm) and then as rpm increase it becomes less important. 

 

This seems to fit with the issue somewhat. They have definitely mapped it on sequential and yet the car is physically set up as a grouped pair.

[jackherer 543]

The ECU is set to sequential but it is not wired for sequential so it is not actually running sequential. And that is how it was mapped.

 

BTW is that some top secret military stuff in your photo?

Edited September 16, 2019 by jackherer

[Telf 224 2 Cars]

That's not the point mate. If the software is set to sequential it will be trying to operate in that mode. So think about it . The software is trying to operate in that mode and is driving the injectors. At the same time it's looking for a cam sensor signal that it isn't receiving and therefore it won't synchronise, result:Fueling is out  and we get mad AFR figures.

 

Now when you drive it it's fine but it's now in the high 2000 rpm band at least . Sequential now doesn't matter and it appears to drive ok. 

 

The person you spoke to earlier said it didn't matter if the ECU is in grouped or sequential but he probably doesn't realise I don't have a cam sensor etc. Essentially he has never seen the car!

[Telf 224 2 Cars]

Added to that could he ECU be crashing the laptop connection because it doesn't see the Cam sensor input? It ran fine on your car but it had your Map installed that we know is 100% correct

[jackherer 543]

The cam sensor is disabled in your ECU and always has been.

 

I'm not quite sure what the Emerald is doing when the injectors are set to sequential but it is configured to only use a crank sensor as it is an impossible configuration, sequential requires a cam sensor. (technically I think it is possible to run sequential without a cam sensor with some ECUs but starting the engine is literally 50:50 every time, sometimes it will be synced, other times it will be 360 crank degrees out.)

 

Having it set to sequential with no cam sensor obviously does something because my car wouldn't run that way but it isn't actually running sequential.

 

Also don't forget I have seen your injectors firing into a bucket and it was definitely a logical group fire pattern that I was seeing (not that I was thinking of that at the time)

 

This definitely does not explain why my laptop is being disconnected! And anyway the cam sensor is turned off in your ECU.

[jackherer 543]

14 minutes ago, Telf said:

The person you spoke to earlier said it didn't matter if the ECU is in grouped or sequential but he probably doesn't realise I don't have a cam sensor etc. Essentially he has never seen the car!

That was Dave Walker and he knew it was a 205 GTI!

 

He literally wrote the book on engine management - https://www.amazon.co.uk/Engine-Management-Optimising-Carburettors-High-performance/dp/1859608353

[Telf 224 2 Cars]

1 minute ago, jackherer said:

15 minutes ago, Telf said:

The person you spoke to earlier said it didn't matter if the ECU is in grouped or sequential but he probably doesn't realise I don't have a cam sensor etc. Essentially he has never seen the car!

That was Dave Walker and he knew it was a 205 GTI!

I wasn't naming him on purpose!

[jackherer 543]

Well, he says he is personally going to resolve it so let's just take it there.

 

If I take my wideband and laptop we can test it before leaving so we'll know if it's fixed or not.