Emissions test/closed loop/ RoverMEMS diagnostic software

[Superdunx]

So... My MGF mk2 VVC Emissions saga continues. Basically it failed an MOT in Feb for emissions and since then we have been trying to get it to pass. I think I now know what the 'F' in MGF stands for, its certainly a word I have muttered many times over the last 4 frustrating months!

To cut a (very) long story short we replaced a physically damaged lambda sensor, identified that the oil temp low fault was because someone had crossed the VVC decrease and oil sensor cables, and then identified that the incorrect oil temp sender had been fitted by a previous owner (which may explain why the cables crossed as it was the later MEMS3 connector/sensor colours). On the way I have continuity checked all wiring between sensors and the ECU, cleaned up all body earths, oscilloscoped almost all sensors, reset the ECU adaptive settings, reset the air idle valve, fitted a new genuine MGR catalyst, and experimented with the garages emissions tester and plugging the LHS tail pipe of my TT mk7 exhaust. (which I can confirm DOES lower the lambda, but I now need to persuade the MOT tester that his measurement technique is incorrect!)

I am now clear of any ECU faults, the live data and oscilloscope traces seems to make sense and the car seems to reliably go into closed loop, but what is concerning me is that some of the data I have captured using RoverMEMS diagnostic software does not make sense. Also occasionally I am getting either a 'feedback failure on clamp' error or 'fault system high voltage' error. The fault seems intermittent and when it is not playing up it can now pass the emissions limits, but I am not wanting to take it for an MOT as occasionally it seems to 'freak out' and the lambda and Co fail, usually when the engine is hot and has been left to idle. (i.e. the usual conditions that they use to heat the car up before an MOT!)

I attach a spreadsheet of an example run where I started the car and ran it randomly between idle and around 3000 rpm for a while to heat the engine up to normal operating temperature until the rad cooling fan cut in. (note, no MOT so cant drive it!) The scans are 2 a second, so each two rows are a second of captured data. I have highlighted the figures that look odd in light red.

So what concerns me are the following - if anyone can shed any light on these I would be most appreciative:

1.) what is the 'system fault high voltage' that appears on row 4033 of the spreadsheet? Why has this appeared? could it be that the O2mV have a limit that has been exceeded?

2.)The ECU dropping out of closed loop occasionally seems to be associated with 'RPM error'. Occasionally the ECU seems to register a significant 'RPM error' It seems to appear from nowhere, persist for a while and then settle back down. I have scoped both the crank position sensor and cam position sensor and signals look good and clean, also there is no issue with ignition timing, which I would expect if these means of measuring RPM were in fault. So what could be causing this? why does the ECU suddenly decide that there is a large RPM error when the RPM measurements looks consistent for the true engine speed and throttle position? any ideas? how does the ECU determine what the correct engine RPM should be?

3.) Despite the fueling percentage increasing/decreasing there seems to be little change in injector pulse width and only 2 of the four injectors are reporting their data. Is this a quirk of the RoverMEMS software or indicative of them never going into full sequential injection mode but remaining in grouped mode of operation?

Other than that I am really struggling to know what to do next, I am really at the end of my tether on this. Whilst I have good electrical knowledge and tools at my disposal I seem to be constantly defeated by understanding what goes on in the 'black box' of the ECU, especially deciphering fault codes and working out what the ECU's given response is to certain faults. For example I spent ages chasing my tail on the oil sender issue only to conclude that the ECU synthesises its own oil temp values if the sensor input is erroneous. How it does this I do not know, I assume using some calculation based on the coolant temp?? nobody seems to know.

Are there any Rover Mems J2 experts out there or even ex MGR or Powertrain Ltd people on the forum who may be able to help? I'm sure that somewhere in the documentation or service information there must be a fault tree analysis, failure mode effect analysis or specialist diagnostic guide that explains all.

Many thanks in advance and I am most greatful for any technical knowledge that anyone may be able to share.

Best regards,
Duncan.

[Roverlike]

Have you tried to disconnect plugs from two (2) Solenoid Control Valves and make measures without them connected?
In case you have measurements orrect now, then you just need to change both solenoid control valves.
Disconnecting plugs from them will limit rpms to 5000, but that is just to test how engine behaves in such situation.

[Deepfat]

All this is beyond my scope apart from one tiny bit.
Tell the MOT person to get stuffed, they must test the car "as presented" and are not allowed to remove the bung, if it passes the emissions that's it, they can if they so wish make a note under advisories but they can't fail it because of the bung.
This information came via another forum and the person who gave it works for Driver and Vehicle Standards Agency and part of his job is the ongoing testing of MOT testers.

[Superdunx]

So an update on my emissions problem.... It passed! But only just

Its still high on Co and obviously running rich so I still think I have an issue, but the car runs, drives and sounds great.

Roverlike, I have now fixed the VVC issue and I can see on the diagnostic that the cam period increases and decreases as it should. Previously with the wires switched it would not decrease the cam period and popped and banged on the overrun, which it does not do now so that part is definitely fixed.

Deepfat, I intend to make my own bung to plug one side for next years MOT. With one of the twin tailpipes reverse flowing at idle you will never get an accurate lambda reading. My other pet hate is idling the car to heat it up for the test (hot engine cool cat) MGF's hate being left on idle and the cat won't get properly hot and working unless you have some good exhaust gas flow!

Regarding the remaining issue responses seem a bit thin on the ground.... I thought they'd be some techies who would have been well into the ECU code and my spreadsheet data but seems not I even emailed the RoveMEMS guy a while back but no reply. ONly explanation I could come up with is maybe the RPM error is caused by magnetic detritus on the crank sensor (from the started gear), but surely if that was the case then it would occur more regularly and it would also upset the ignition timing too? any ideas?

[Deepfat]

I fell lucky with my local (1/2m away) MOT station they are into old cars, I give it the Italian tune up and its straight onto the emissions, might be worth asking around to find something similar. Either that or move to NZ where I think if the car is over 20 years old they don't test emissions which seems logical given its a very small % and your asking quite a lot out of what is very old kit.
Re your mems info have you tried the T-Bar forums they have a dedicated pscan forum which is chocka of stuff about mems that I don't understand, I go on there when I can't sleep, works a treat.
Congrats on the pass

[Roverlike]

Maybe I was not clear enough. I suggested you to disconnect solenoids in relation to your high volatage and RPMs errors. Faulty solenoids can cause this kind of errors on MEMS2J VVC engines.
Disconnect solenoids and try the test you are doing and see if high voltage and RPM errors will be present. If not, then you might have faulty solenoids and you would need to replace them.