Ulysses

Good job matey!

Cinderella will go to the Ball

Been a bit of a while since I updated the thread. 

The trip out with Brett and Dave was great and a much needed day out. Then this week the MOT was due...

 

Fails on the emissions spectacularly . It was reading so high on their tester that it was almost off the scale. They think the cat in the exhaust has collapsed so I've brought it back home to get it off for a look. Its also overfuelling at idle (or there is too much unburnt fuel in exhaust) so I need to look at the ecu mapping a bit more carefully. Booked back in saturday for a retest so hopefully I can do something about it before then. 

Have you had any head gasket issues? Coolant can take out lambda sensors and make it run very rich.

wrote:

Have you had any head gasket issues? Coolant can take out lambda sensors and make it run very rich.

Hi Les,

Nah, the head and cooling is fine. Its just not that simple with Ulysses. The factory ecu is gone and replaced with a DTAfast S80 ecu to run the individual bike throttle body setup on the M18vvt engine. Most of the sensors are there apart from anything in the exhaust now, so the ecu doesnt do any fuel trimming based on lambda.

The map it runs was setup on a rolling road by a very clever fella, but needs a bit of looking at 2 years later really. I will ask about putting a o2 sensor in the exhaust as the ecu should be able to support it. I think there was a reason why not to use one, though I cant remember what it was...

The cat in my home made exhaust is steel cased and looks pretty sorry for itself on the outside, so I imagine the internals are past it by now too. The MOT guys were suggesting it might have collapsed and be doing nothing now.

O hang on... I think I've just figured it out after re-reading above...

A while ago I changed the sensor scaling for the coolant temp as the one it was on was calibrated much higher than it should have been (after comparing with a thermometer).

I wonder if thats meaning its always running on the cold start side of the ecu mapping, which is set to be rich to keep it from stalling. That would explain why its vastly different to last years MOT at the same garage with the same car.

wrote:

The factory ecu is gone and replaced with a DTAfast S80 ecu to run the individual bike throttle body setup on the M18vvt engine. Most of the sensors are there apart from anything in the exhaust now, so the ecu doesnt do any fuel trimming based on lambda.
 

Ah, I forgot you are using ITBs. I run my trials car on a Megasquirt which also does not use any lambda sensors. Luckily in my case the car is old enough to only need a visible smoke test. I tend to run on the rich side to try to control pre ignition.

Running rich can overheat cats, which may explain why yours doesn't work very well now.

Lambert, here's a bit of an explanation about what is going on.
Modern cars have sensors in the exhaust (lambda sensors) to measure how well the fuel is burning. They can tell if too much or too little fuel is being used for the best emissions and economy. Riccy is using a custom fuel injection system that does not use these sensors. The system was calibrated on a rolling road when it was first installed. This method is not as accurate but can give more power at the expense of fuel economy and emissions.
When you start a cold engine the fuel does not vapourise very well so the fuel injection adds more fuel to compensate. It's the equivalent of the choke on an old carburettor. When Riccy's injection system was set up the sensor that measures engine temperature was set up wrong. He later changed the calibration of this sensor which means the system now continually thinks the engine is cold and dumps in too much fuel, killing the emissions. As it is not using a lambda sensor it does not know it is running too rich.

Hopefully setting the temperature sensor back to it's old calibration and maybe replacing the cat should solve the problem.

 

Thanks Les, That is a masterful summary of what I think is happening. Essentially the mapping was set when the coolant sensor scaling was wrong, so the map temperature was calibrated based on the wrong measurement. I changed the sensor scaling without altering the map, resulting in the ecu thinking is -10 most of the time and going into cold start mode unless I get it really hot. This may also explain the 140ish miles-to-a-tank comedy fuel consumption I get.

Lambert, you are showing off, but Stoichiometry (yes i had to look it up...)  is working, but not as efficiently as it should currently, therefore the emissions are worse than a 40yr old tractor 

Yay! a pass  

Set the sensor scaling back and there was an immediate improvement with the idle at running temperature as well so hopefully that was all