Freelander running rich

[Mark Adams]

Hi,

I have just had a friend bring his S reg 1.8 Freelander round to me with the problem of it running rich.

I initially checked the O2 sensor and it is fixed at 900mv initially thinking that it was faulty I disconnected a vacuum pipe and the o2 sensor responded and gave a reading of 200mv. On WOT it also responds so I feel the sensor is fine. No DTC's

Car generally runs fine, he tells me it also has a slightly rough idle when warm so I checked the CTS but it is fine (voltage decreases as temperature increases) but this may be a separate problem.

Air filter is new, recent oil change. Connected a vacuum gauge and got a steady reading of 18 " mercury at idle (got the same reading from live data of the MAP sensor), so does not seem to have any vacuum leaks, sprayed carb cleaner in the normal places to confirm and no increase in RPM.

Got new platinum plugs about 1000 miles ago when I removed them they are all clean and look normal (no deposits).

Fuel pressure is 3.5 bar while engine is running.

Very similar problem to what I had on my own freelander, but can't figure out the cause of this one.

[Ally V8]

What diagnostic tool are you using ? Testbook will show if the ecu is running in closed or open loop,if the oxy sensor is giving 900mV then the ecu ought to be trimming back the injector pulse width - but only if its closed loop.In open loop mode it will just run default values of injector opening and usually the map errs on the rich side.

When you start the engine you should see it go from open to closed loop fairly quickly,within 30 secs even from cold.But, it will soon wimp out and go open loop if there is something it does not like - usually a lack of response from varying oxygen sensor outputs.

Mems ecu's are not the most reliable,if nothing else can be found I would code up another ecu and try that before tearing too much apart or using the parts shotgun !

[Mark Adams]

Hi AllyV8,

I used an SP ACR4, I took it for a test drive with the unit connected and displaying live data, I noticed that the Lambda reading mainly stayed at 902mv. But also noticed that the car was in Open loop on the whole journey (as you mentioned).

The car starts as normal in open loop, then went into Closed loop but soon went back to open loop and stayed there. Exactly as you mentioned.

I have always taken the approach of only replacing parts that I have tested to be faulty so I know what has caused a problem before replacing any parts.

It is helps, I own a 4 gas analyser, the SP ACR4 and a 2 channel scope, fuel pressure gauge and other test equipment.

[Mark Adams]

Went out early this morning to get a new NTK O2 sensor as I was informed that the one on the car is the original sensor so it will be a bit slow to give good readings even when working properly.

I removed the old one with no problem, but cant get the new one to fit, the fins on the sensor seem to catch on the inside of the exhaust manifold, if it is forced in it closes all the openings. I noticed on the one that come out that the slits were hardly open. Unfortunatly due to the location of the sensor I cant see what is causing the problem.

Has anyone else come across this same issue? Is it just a build up of rust on the inside of the manifold? I would rather not remove the manifold as I know what the nuts/studs will be like and it always turns out to be a pig of a job as everything rusts solid.

The sensor would be far better with holes to sample the gas rather than these open slits.

Any idea?

[Ally V8]

This is a common problem with the K series downpipes,there really isnt much room in there.You may need to resort to removing the downpipe assembly and "modifying" it. i have to say though,that if the oringinal sensor is still capable of switching up to a rich signal the ecu ought to be trying to narrow down the injector pulse width to get the sensor to switch back down.

If the cambelt is timed and tensioned correctly,compressions are all good and the igntion side is all good - I'd go for an ecu swop. Other things to do are to scope the HT's - I'm sure if you Google HT waveforms there will be plenty of help there.You need more than 3/4 of a millisecond of spark burn time reliably at each plug for proper combustion.

Checking the individual downpipe temps with a lazer thermometer will help pinpoint and large differences between cylinders too.

[Mark Adams]

Not sure if this is going to help but I captured some live data (taken just after starting but o2 sensor had warmed up).

You can see it goes into closed loop them back to Open, I have shown Injector Advance, the Injector Dead time is a constant -0.6 which does not seem right.

[Ally V8]

I still think the ecu is failing to change the injector pulse width,one of the ways these Mems ecus seem to fail.Do you have access to another unit ?

[Mark Adams]

unfortunately I dont have one I can just swap to try it. I would need to buy one to try it. will it just work if I swap it out or will I need to take it to a main dealer to get it re-coded?

Mark.

[Ally V8]

Allcar electronics,(Alfreton,Derbyshire) used to do a loan service where you borrow an ecu that has had its coding function removed - will plug straight into any car.You try it,if it works you keep it,if not send it back.Cant remember quite how they do the money bit,but its a good service.

If you are in the westcountry I could code up my spare on your car for bottle of plonk or similar...

[Mark Adams]

Ally V8,

Sorry for the delayed reply.

Thank you for the kind offer, unfortunatly I am based in Wales just outside Swansea and I am not planning a trip in your direction for a while. Be interested to know if Allcar still have the service you mention?

What would happen if I just swapped the ECU? Would nothing work I guess due to the Immobiliser?

Mark.

[Ally V8]

An uncoded ecu will allow the engine to fire a few times then just die out.Looks like Allcar are still doing the hire service;

http://www.allcarelectronics.com/

[Mark Adams]

Hi,

The service was very quick, I received the test ECU today, I plugged it in and drove the car to warm it then retested it, unfortunatly it has not changed. Lambda sensor still fixed high.

Thank you

Mark.

[Ally V8]

Thats a pain,before you send the ecu back you could try scoping the injector pulse width and seeing if the ecu is trying to change it while the oxygen sensor signal stays high.Also stick a pin through the black wire from the oxy sensor and scope that,it should start to give an output after about 15 secs,engine hot or cold.When you have finished the scoping seal the pin hole with nail varnish,or the conductors will corrode leading to more problems.

If you can datalog the scope signals you should be able to see the pulse width narrow down and then the oxy sensor signal drop down below 420mv.This is where the pulse width should then widen and the cycle starts again.Pulling off the crankcase breather a little ought to let the map sensor see less depression and cause a weaker mixture,again the oxy sensor ought to respond.Seeing its output on the scope is far more accurate than viewing from a scan tool whose refresh rate can be quite slow.

Also just a thought,does the new ecu give very similar readings from the map sensor under the same conditions as the old one did ?

[Mark Adams]

Thank you for the reply, I took the reading using my Fluke scopemeter as I was hoping to be able to measure the duration of the wave form but unfortunatly I just have a fixed voltage of approx 845mv. I scoped the injectors and the pulse width did change when opening the throttle, if you WOT then the lambda signal also drops suddenly (as it should) and goes straight back to 845mv. I had the lambda sensor on one channel and injectors 1/3 on the other channel. I can see the pintle hump close on the voltage trace and open on the current trace.

When the pipe is removed slightly the Lambda sigal drops to approx 200mv, the MAP signal changes and the pulse width widens until the car eventually cuts out.

I rechecked the MAP signal on a drive up the hill and when idle and it gave the same results.

I perform all of the above tests with both the old and new ECU and the results were the same.

[Ally V8]

Have you made any progress on this ? Its good that you have the kit and knowledge to measure and test,most dont. What troubles me is that the ecu is clearly able to alter the injector pulse width - but not enough to get the oxy sensor cycling under idle or cruise conditions.One last thing you could play around with is to modify the coolant temp resistor value to fool the ecu into thinking its hotter than it is.I had to do this once on a Rover V8 in a Ginetta to get it to pass its Mot emissions test.This was with a modified Gems ecu set up for high performance.It shouldnt be necessary on a standard 1.8 FL,but it may go some way towards finding whats wrong with your car.

[Mark Adams]

Hi,

Tried what you suggested by wiring up a spare temp sensor and heating it slowly with a blow lamp. Unfortunatly no change.

I took a video of injector 1 (as this is the only one you can get to unless you use the connector) which is here:

The pulse width at idle is 3.945 ms. little bit high I thought.

[Mark Adams]

Am I right in thinking the pulse width should be approx 3.5 ms when warm?

Mark.

[Ally V8]

Sorry to take so long to reply,I think it should be more like 2.5ms than 3.9ms. Trouble is I dont often see them now,my wife still has one,but it rarely visits my workshop,(Thankfully) and hasnt been on Testbook or scoped for a long time.If I get a chance I'll borrow it and have a look.

[Mark Adams]

It would be interesting to see how another car compares.

Look forward to hear the results.

Mark.

[Ally V8]

I ended up with Testbook and my Snap on Modis at home this evening... Unfortunately neither will show injector pulse width in live data.So I 'll try to remember to bring home my Picoscope and do it from that.Please be patient - I have rather alot going on at the moment.

[Ally V8]

Ok, I finally managed to borrow her car,took two lots of readings,scoping injector 1.

1st, at idle;

RPM 780

Coolant temp 95c

Map reading 310 mB

Oxy sensor switching nicely - average injector duration 3.15ms

2nd reading;

RPM 2750

Coolant temp 95c

Map reading 230 mB

Oxy sensor switching nicely - average injector duration 2.67ms

The car has its original engine,running very well at 107,000m,no faults logged in the Mems ecu and passes emissions well. The scope is a 4 channel Pico and I used my Modis to read the other live data.(Quicker to set up than Testbook)

So my wifes car clearly has a shorter pulse width than your problem car,but I guess the problem now is where to go in finding why Mems is putting out such a large pulse width, when it clearly is seeing a high reading from the Oxy sensor.

[Mark Adams]

Thank you for the reply,

The other interesting thing I found which may or may not be related is:

When trying to get a steady fast idle speed it is impossible, you can hold the TPS steedy but the revs keep increasing, I have check the TPS on the scope and it is fine and the reading for TPS do not change in live data.

Also when you WOT and let it idle again, instead of going back to idle speed straight away it sits at about 1500 for about 30 seconds then drops back down.

Ever come across these?

Mark.

[Ally V8]

They are very difficult to hold at a steady rev,esp for Mot purposes,I once clamped off the idle bypass hose during an Mot which helped.Today when I was using the scope on my wifes car I wanted to hold it at 2000rpm but it wouldnt stay there,ended up at 2750.So I dont think yours is any different in that respect,but taking 30 secs to come down to idle is not right.It may be worth getting it on Testbook to see if the idle settings are all correct,Testbook will automatically control the engine to check idle response etc.I cant remember its exact proceedure as its so long ago since I last did it,but it does make a big difference to how they run.