Exhaust gas leak. Suspect manifold

[Andy 105 1 Cars]

In which case we will wait for Kieran’s diagnosis. For the wiring, screened cable for the crank sensor, and twisted pairs for the temp sensors and throttle , with the grounds for all the sensors connected to the same pin on the ecu, good ecu ground  

and all should be well on the interference front . A poor earth will give an earth offset, which will give some random sensor readings .

 

[MrStock 5]

I have sometimes a very strong fuel smell only when windows are open. Sometimes there is nothing and sometimes i could throw up while driving. I replaced and checked every hose and line, MOT is no problem, car is running fine. I blocked fuel vapur cause it was rotten. It changed nothing.

I read about from many guys and i dont think all have solved that problem. For me, i think its maybe a little hairleak in the fuel tank or something like that, i dont loose petrol or have high fuel/mile.

I have that problem now for 7 years, i have adapted to it.

[Telf 227 2 Cars]

Andy,

 

In terms of the cabling the ECU loom was built by me using the emerald kit. The crank is shielded and as you say all sensors go to a common ECU earth. The main earth is at the gearbox. 

[Telf 227 2 Cars]

7 hours ago, DamirGTI said:

of the books i have is written by Dave Walker , i can make you an copy of pages

Damir, 

 

Once we saw this random changing of the AFR yesterday it's one of the first things Kieran said-maybe there is a cable to near the crank or the coil etc. We removed the alternator belt so the car was running just on battery to see if that might be the issue as the main engine loom is near but alas it wasn't. 

 

Feel free to post I will read anything that is provided to assist trust me!

[Telf 227 2 Cars]

25 minutes ago, MrStock said:

have that problem now for 7 years, i have adapted to it.

Mr Stock. This is not an option in mind . The car is a daily driver and I'm not being booked up to my tail.pipe for 2 hours a day! 

[Andy 105 1 Cars]

1 hour ago, Telf said:

Andy,

 

In terms of the cabling the ECU loom was built by me using the emerald kit. The crank is shielded and as you say all sensors go to a common ECU earth. The main earth is at the gearbox. 

 

Right. So built by the book. Next port of call has to be the injectors . Reconditioned or new ? One injector with a slight or intermittent leak would also cause the afr drift and the fuel smell. If Keiran has a spare set, it might be worth trying a different set, just to rule out another potential source of the problem. As for living with a fuel smell. No thanks . I would rather take the bus 

[Thijs_Rallye 87 4 Cars]

What is the configuration of the Emerald, as in what sensor inputs are you using?

 

As for the shielded crank cable, is the shielding/earth connected at the ECU and the engine ground? Should be earthed on one side only.

 

And in general, how is your ECU grounded? All sensor and ECU grounds should be attached to one point only (preferably at the engine). So in case if you ground the ECU inside the cabin, and the sensors on the engine it may be possible you will get a ground offset and the ECU will start going to crazy stuff, potentially burning itself. I am just thinking out loud here, since I haven't followed the build page for page.

[Telf 227 2 Cars]

The ECU is earthed at the gearbox

[Telf 227 2 Cars]

There's the basic schematic. It also has a Map sensor ( not used) and radiator control 

[Telf 227 2 Cars]

And the pinout at the ECU

[Telf 227 2 Cars]

Fan control

[Telf 227 2 Cars]

And the basic pinouts from the cable supplied from emerald.

[Telf 227 2 Cars]

41 minutes ago, Thijs_Rallye said:

What is the configuration of the Emerald, as in what sensor inputs are you using?

It's using water temp,air temp, crank and throttle position. I don't know it's name but it's throttle v rpm mapped into load sites as posted earlier

[Telf 227 2 Cars]

Multi plug under the dash

[Telf 227 2 Cars]

And the starter relay wiring

[Andy 105 1 Cars]

2 hours ago, Thijs_Rallye said:

What is the configuration of the Emerald, as in what sensor inputs are you using?

 

As for the shielded crank cable, is the shielding/earth connected at the ECU and the engine ground? Should be earthed on one side only.

 

And in general, how is your ECU grounded? All sensor and ECU grounds should be attached to one point only (preferably at the engine). So in case if you ground the ECU inside the cabin, and the sensors on the engine it may be possible you will get a ground offset and the ECU will start going to crazy stuff, potentially burning itself. I am just thinking out loud here, since I haven't followed the build page for page.

Looks like Paul used a flying loom from Emerald and then cut and terminated to suit. In which case all his sensors and shielding will go to the same pin on the ecu. Then he has earthed the ecu on the gearbox which is also the battery ground point, so all looks well there . If he had also connected the crank sensor shield  at the sensor end too, he may have got a spurious crank signal, in which case the ecu would have lost sync , spat its dummy out and cried  for its mummy . In other words, he didn’t and all is well on the wiring front . My next port of callwould be injectors 

[Telf 227 2 Cars]

Yep Andy I did use a flying loom

[SRDT 188]

Check your HT leads, a bad plug can even mess with a stock ECU.

[Telf 227 2 Cars]

Checked. Infact replaced with a different set. No difference.

[DamirGTI 344]

Resistor spark plugs fitted ?

 

D

[Telf 227 2 Cars]

Damir.. yes!

[jackherer 543]

I've spent a bit of time looking at Paul's car and I've narrowed it down but I haven't got an answer yet.

 

First of all it is definitely a fuel mixture problem and not a fuel hose leak or anything else. Paul bridged the fuel pump relay and checked for leaks/fumes for about ten minutes with none apparent.

 

My wideband is reading as rich as 11:1 between 1500-2500 in neutral and while driving. Below that it is lean, 17/18:1! Everywhere else seems fine, it actually drives really well.

 

There is another problem though that is making further diagnosis tricky. The laptop disconnects from the ECU after a short period of time and wont reconnect again until I restart it. The laptop that is, I can leave the ECU powered up and reconnect to it, I only have to restart the laptop.

 

If the engine is not running the laptop will stay connected indefinitely, it only disconnects when the engine is running.

 

I checked the laptop and serial adaptor etc. are working by connecting them to my own 205's Emerald (albeit an M3D, not a K3 like Paul has) and with the engine running I was able to trim fuel/ign, change settings etc. everything I had been trying on Paul's worked fine on my 205.

 

As Paul's car is only running rich (or lean!) at certain parts of the maps I really need to see which parts it is on while it is rich and the same for the supplementary maps etc. but it's impossible with the laptop disconnecting constantly.

 

This reminds me a lot of the behaviour of my own Emerald-equipped 205 when I accidentally fitted non-resistor spark plugs (good call @DamirGTI) but Paul definitely has resistor plugs and also to rule them out I actually fitted the plugs from my 205 which made no difference.

 

Water (and air) sensor values always appear correct when the laptop has been connected so I think we can rule them out. The TPS is calibrated and it's easy to move between load sites with the pedal (I can see that on the laptop with the engine not running when it will stay connected indefinitely)

 

Interference and shielding are also good calls, I have Dave Walker's book so I am aware of all the details in there. I did find some issues with the CPS wiring and its shielding and I briefly thought I had found the issue but I corrected those and it made no difference whatsoever. All other grounds appear correct, I added a toothed washer to the gearbox stud as it is good practice but it made no difference either.

 

I took the injectors out and span the engine over with them pointing into a bucket, the spray patterns were identical and they all fired consistently for quite a while. They also had brand new seals, caps etc fitted which makes sense as Paul had them cleaned/tested before he put his car back together.

 

 

There are still some things for me to check but I think I will contact Emerald about the laptop connection problem tomorrow to see what they think.

 

I may try putting Paul's ECU in my 205 to see if I can connect to it from the laptop with my engine running... Does anybody know if you can import a map from an Emerald M3D into a K3? I had already discounted doing it the other way round because there are more cells in the K3's tables so Paul's map definitely wont go into my M3D but the other way should be possible even if it is a bit crude, Emerald upgrade M3Ds to K3 spec and I've never heard of that involving a remap.

[Andy 105 1 Cars]

Wow. You have done lots of thorough checks. Impressed. At 17 -1 I am surprised that the engine will run.  Atthe other extreme, 11-1 at 2000 rpm under no load is really rich. Looking at the map , the injector pulse times would suggest a reduction in pulse width at speeds abouve 1000rpm with the throttle shut, making it weaker .  However, the ignition timing is flat at 0 beyond 1000rpm. That would give very poor combustion . Try the following,

 connect up the ecu to your laptop, power it up and launch the software, so that you have access to the map but without the engine running. Go to the ‘live adjustments ‘ screen and alter the ignition values in the table on the top line from 1000 to 4000. Start at 15 and increase steadily as you go to 4000rpm up to about 32degrees . As these are live adjustments , they should be saved to the ecu. 

Then, start the engine and see if that has made any difference to the afr at no load and rpm between 1000 and 3000 

[jackherer 543]

I forgot to say, I think there might be a small exhaust leak so that could explain the lean AFR at idle up to nearly 1500.

 

Then it goes rich (11:1) from 1500 up and gradually leans slightly to about 12:1-12.5:1 by 2500 after which it suddenly jumps to about stoich.

 

Incidentally I was a bit suspicious of the readings from my wideband especially as I was using a brand new sensor but I tested it in the tailpipe of a modern car with a lambda/cat etc and it behaved exactly as it should.

 

I'll try that tomorrow but I think I'd have to save the changes to the ECU first, the live adjustments only take effect while the laptop is connected, as soon as you unplug it the ECU reverts to the saved map.

 

The thing I keep going back to is that it has spent several hours at Emerald on their rollers with their wideband in the tailpipe achieving sensible AFRs and with their PC connected to it without disconnecting. And that's happened twice now!

 

Something else worth mentioning is that Paul bought his ECU second hand. It turned out to be a custom configuration, something to do with VW R32 ignition, it had extra drivers, or fewer, I can't remember. The bottom line is it had to go back to Emerald where they put it back to standard and thoroughly tested it but presumably that involved some hand soldering so I wonder if there is a dry joint or a fleck of solder somewhere it shouldn't be.

[jackherer 543]

It does look like the K3 will accept an M3D map, I just opened my map in the K3 software and it seems fine so I'll try isolating Paul's ECU in my car.

 

2 hours ago, jackherer said:

I forgot to say, I think there might be a small exhaust leak so that could explain the lean AFR at idle up to nearly 1500.

 

I also forgot to say, I'm using an Innovate exhaust clamp so there is definitely the possibility for an exhaust leak upstream of it causing a false lean reading but Innovate are adamant the sensor will never read falsely rich so if anything it is running richer than the 11:1 I am seeing!

 

 

I assume the effect of a leak will diminish as the exhaust flow increases with RPM hence it reading approximately stoich once it's passed the 2500RPM trouble spot.