[SteveB]

At the risk of sounding a complete idiot this is my (very basic) layman’s understanding of the system. When you press the accelerator the throttle opens and the TPS sends a signal to the ECU. The ECU then sends a signal to the injector to provide an appropriate amount of fuel.

Your first problem describes major loss of power but the engine doesn’t actually stall, which suggests the ignition system is probably ok – you’ve still got a spark. It’s also getting enough fuel to keep running. If there was a blockage in the fuel supply I presume it would cut out completely. Could it be that, at that precise moment, the ECU doesn’t ‘know’ you’ve got the throttle open and sets the fuelling to tickover? Having the throttle disc open would weaken the mixture even further?

The second problem describes massive over-fuelling. The engine’s ticking over, but it’s as if the ECU suddenly ‘thinks’ you’ve opened the throttle and provides too much fuel.

I think both scenarios could suggest a problem between the TPS and ECU. Since you’ve checked the sensors and their outputs I would suspect a fault with the ECU. I had to replace mine last year on the MPI. There was an intermittent problem – it would run fine then start cutting out for a fraction of a second. Then it would run fine again, then be a non starter. After a lot of trial and error, fitting a reconditioned (and ‘decoded’) ECU sorted the problem.

My first Rover, a ‘91 SPI had no remote central locking and the alarm was simply a switch under the dashboard, so no 5AS. I once had an intermittent problem with it and as part of the troubleshooting process changed the ECU. It didn’t need any coding so was a straightforward swap – I presume yours will be the same, so I think it would be well worth trying your spare.

[g259fsg]

Yes, that's exactly how I understand the system works, Steve. The other parameter is the MAP sensor, which in combination with the TPS, determines the fuel injection level based on power demand. The MAP did seem to be working perfectly from the RoverMEMS App. Other parameters like coolant temperature and air temperature also influence the data, but I think they have less importance than the main parameters and if the sensors are faulty, the ECU uses default values which allows the engine to work.

When I first had the problem in November (as per the first message on this thread), like you, I thought the TPS might be faulty - the throttle was open but the ECU wasn't getting the message, thinking the accelerator was still in the idle position. The RoverMEMS App didn't show any problems with the TPS, however I suppose that doesn't prove anything since the fault has occurred extremely rarely and just for a short time. I did try putting the car into first gear in idle to see if it would move, but it didn't, so the low-power fault condition is more complicated than just the ECU thinking the engine is in idle, I do get very weak forward motion. I agree with you, that all the symptoms I've had so far indicate the ignition side is working correctly, including the crankshaft position sensor.

From what you and others have said, my current thought is the main potential culprits are the ECU malfunctioning in an intermittent way, or a fault/contamination in the injector/fuel-pressure regulator.

My next step is to change the ECU to the spare, as you and others have suggested. The ECU is MNE10013 which is a basic version with no immobiliser so not complicated to swap. This will be useful anyway as I haven't actually tested it and the car is currently working 99% of the time, so a useful check. After that, I'll need to take off the the throttle body and examine the injection head.

It's a bit too cold to work on cars at the moment, so I'm planning to continue to use the Rover for shorter local journeys, where if the worst happens it's not too inconvenient to get a taxi and recovery. I'll keep logging incidents. For longer trips, we'll use the three Rootes cars which are all runners and very reliable.

Thanks again for your insights, much appreciated.

I'll report back when I have more news.

Hugh

[g259fsg]

Here's the latest instalment in this saga. I ran the engine up to temperature and connected the RoverMEMS Diagnostic App. Interestingly it showed logged faults from when I lasted tested a while ago. There was a Coolant Temperature Fault and a MAP Fault. Once up to temperature, I got the following readings:

Coolant temperature 86
RPM 874
Air intake temperature 43
MAP 33 kPa
Throttle 8%
Idle control 15
Ignition advance 10.5
coil time 775 microsecs

I fitted the spare ECU. The engine stalled a couple of times, but then ran smoothly up to temperature. I got the following readings:

Coolant temperature 83
RPM 859
Air intake temperature 40
MAP 32 kPa
Throttle 8.4%
Idle control 18
Ignition advance 10.5
Coil time 806 microsecs

It also seemed to respond when I switched on lights etc., holding the idle at roughly the same speed.

After a bit of further running, I tried driving it out of the garage. Every time I tried to get some power, the engine died. Got it back in the garage and reconnected the Diagnostics. This showed that the MAP was now permanently at 10 kPa (virtually a vacuum), way too low. It didn't respond to opening the throttle. The MAP looked to have gone faulty. I tried refitting the original ECU and the engine ran normally with the correct MAP reading and the car was fine on a road test. Looks like I need to source another spare ECU.

The one on the car is a MG Rover replacement from 2002 (it has the E suffix indicating it's a refurbed unit). Interestingly, the original one was replaced because of a fault with the integral MAP sensor. I think all the R8 ECU use a similar system including the MPi versions. It means if the MAP is faulty, the whole ECU needs to be replaced.

A couple of things have come to light during these tests. I think there may genuinely be a fault with the coolant sensor. I noticed the diagnostic sometimes showed a fixed 60 degC for a while then reverted to 85 degC. 60 is the default value used by the ECU. Looks like this is one of the faults and was picked up in the fault log. I also now realise that the MAP sensor is critical to the operation, and if it doesn't register the pressure correctly there could be no power. Maybe the short loss of power I experienced at the beginning of this thread was an intermittent MAP reading. Maybe it's significant that there was a MAP fault logged in the ECU memory.

I wonder if anyone has dismantled these R8 ECUs? It would be interesting to see if the MAP sensor can be replaced, even maybe with an external sensor.

Finally, I have a second spare ECU which I got quite a while ago. Again this idled OK, but the MAP showed about 9 kPa (not 32 kPa). Like the other spare, there was no power when I tried to drive the car. Looks like the MAP is the weak link.

Hugh

[GTiJohn]

Richard Jessett, the RCOC parts and electronics expert, has suggest that this is a replacement for the MAF sensor inside these Rover ECUs -
https://www.ebay.co.uk/itm/254557199751

Might need somebody with a steady hand to solder in

[g259fsg]

Thanks, John, that's very helpful. I'll have a go at opening one of the faulty ECUs to see what's involved. As my work was in electronics, fitting a new MAP is not a problem.

I wonder if anyone has worked on one of these. There seem to be a lot of crosshead screws on the back of the board. I assume only some of them hold the board to the metal case.

While searching around, I found this MG Rover video about the MEMS system on YouTube. Gives quite a bit of useful information. Although it applies to the Montego in 1989 and there are some differences, much of it is the same. https://www.youtube.com/watch?v=vzcviSCytfI This seems to say that the air/fuel mixture is largely controlled by the MAP not the TPS. The throttle position sensor seems to be used to give the ECU early warning that an increase in power is required before the MAP responds. A bit like the accelerator pump on a carburettor, it avoids hesitation when the throttle is opened quickly.

Hugh

[g259fsg]

Opened up one of the non-working ECUs. To take out the board, remove the crosshead screws marked S and SL, the SL screws are long ones going through the heatsinks. You don't need to remove the screws in between as these hold the heat sink to the PCB.



The board should then ease up, but there will be some resistance due to the thermal grease on the heatsinks. Remove the rubber tube from the MAP inlet spigot.



This is the other side of the board with the Motorola MAP sensor. The one in my ECU is 5141550T02 (the board is dated 9025). The one recommended above is 5143770U01. I assume this is backward compatible.

[g259fsg]

Did a bit more work on this today. As mentioned, the spare ECU I bought last week initially appeared to work with good data readings on RoverMEMS App. However, when I tried driving the car the engine died. The Diagnostic now showed the MAP sitting permanently at 10 kPa, so it had clearly failed after previously working. The company I bought it from, sent another one which arrived on Friday. These are all now very old parts, MNE10013 was only used in the early 1990s, so up to 35 years old.

I was not that optimistic that the replacement would work but it did. Sometimes takes a bit of time for the ECU to settle down on idle with a new engine. I took it out for a drive and it all seemed fine, plenty of power. So, I've left this one on the car, and put the (99%?) working one in the boot in case of an emergency. The readings I'm getting off the replacement unit are:

RPM 907
Coolant 80 deg
Intake 34 deg
MAP 33 kPa
Throttle 8.8%
Idle Control 33
Advance 8.0
Coil time 788 microsecs

The one that was on the car was a remanufactured MG Rover part. I found this was fitted in November 2003. The garage also fitted an additional vapour trap in the vacuum line. One thing I've noticed with these old units when you open them up is that there's a film of oil or hydrocarbon on the connecting hose and the MAP sensor. Maybe stuff that came from the inlet manifold over the years.

Here's the receipt for that work:



Finally, while searching on the Internet, I found this conference paper written by Rover engineers Crabb, Duncan, Hiljemark and Kershaw about the design of the Rover MEMS. Seems to have been designed in-house including the software, though it looks like the PCB and surface mount hardware was made by Motorola. It was first used in the 2 litre Montego in April 1989 and then in the Rover 214 in October 1989 followed by the 414 in March 1990. Here's the link https://ssims.co.uk/Downloads/MEMS%20Paper.pdf

I'll report back when I get any further developments.

Hugh

[rjessett]

This is the other side of the board with the Motorola MAP sensor. The one in my ECU is 5141550T02 (the board is dated 9025). The one recommended above is 5143770U01. I assume this is backward compatible.

Interesting thread - I pointed John at that sensor as I know it was used in the later mems, no idea if its backward compatible or not im afraid

[GTiJohn]

Well done Hugh !

Your persistence has paid off

[g259fsg]

The replacement ECU has so far performed well. Apart from local journeys, we took it on a 50 mile round trip yesterday without incident. However, it's still early days. Given that the previous faults occurred very infrequently, I may have to wait a while to know if the old ECU was the problem.

I have been doing some further reading on the Rover MEMS, and found a couple of useful documents. There's been a lot of interest in this thread, so I'll give the links below. I'm beginning to understand a bit more about the operation. The Rover MEMS uses speed-density to measure the volume (and therefore mass) of air entering the engine. The other method used by injection engines is the Mass Air Flow (MAF) sensor, usually a hot-wire anemometer. In the speed-density system, the Manifold Air Pressure (MAP) is measured along with the Intake Air Temperature (IAT) to calculate the density of the air and therefore the mass of air taken into the cylinder on each stroke. This algorithm then sets how long the injector is pulsed-on to inject just the right amount of fuel to give an approximate air/fuel ratio of 14.5:1. This increases with RPM, the speed side of the algorithm. The throttle Position Sensor (TPS) has a secondary role, mainly telling the ECU to inject a bit more fuel for acceleration or full throttle performance, and also the conditions for coasting or over-run when the the throttle is partially or fully closed. On later models (1993 onwards I think), it initiates the idle mode instead of the accelerator switch on earlier models. There are other parameters such as coolant temperature and corrections to IAT to make the engine run better. I'm full of admiration for the guys that designed this system, there's so much to it, and it was designed to accommodate future developments. Also, it's one thing to do the science, but these cars drive so well too.

I found a very good description of all this in what appears to be a Haynes Manual about fuel injection. Does anyone know what the book is? The pages on MEMS (Chapter 14) have been scanned by somebody in the Mini Club in France. It's here as a PDF: http://britishminicorp.chez-alice.fr/RoverMEMS.pdf

One thing which was highlighted in this, is the difference between SPi and MPi engines. In SPi, the injector is in the throttle body (an electronic carburettor) so everything downstream of the butterfly valve has got petrol vapour, described as a wet manifold. This means there is potential for petrol vapour to contaminate the MAP sensor. In MPi, the injectors spray petrol on the inlet valves so the rest of the inlet manifold is dry, hence much less risk to the MAP.

There's also a very interesting analysis on temperature corrections by I assume James Portman (creator of the RoverMEMS Diagnostic App). It's here https://rovermems.com/discussion/mpi-ecu-fuel-calc.html

Hope you find this information interesting.

Hugh