[g259fsg]

Just a brief update. I've now done 170 miles without any funny business from the engine. So far so good, but I'll need to probably do 1000 miles to be sure the problem was in the old ECU. Think I'll take it along for an emissions test, well ahead of the MOT. I believe it is possible to trim the fuel level through the diagnostic App. The blurb says that the ECU adapts to a new engine and it takes a while, presumably that is the idling settings? I've not seen the optimisation algorithm described anywhere. Does anyone know how it works?

I found quite a good summary of the Rover MEMS on Wikipedia which describes some of the evolution: https://en.wikipedia.org/wiki/Modular_E ... ent_System

I also found out what the Haynes Manual is on the Fuel Injection systems. It's called "Automotive Engine Management and Fuel Injection Systems" by Charles White, published in 1997. Haynes Manual 3344, ISBN 185960 344 0 . It's very comprehensive covering several different systems. There's a chapter on the Rover MEMS and also a chapter on the Honda/Rover PGM-Fi used in the 216/416 models. A very useful book with some introductory chapters on the history of fuel injection, operation and testing. I got a nice copy on eBay, seem to be several available. A useful addition to your library of Rover manuals.



Finally, I've got two ECUs with faulty MAP sensors, so I'm going to try fitting an external 1 bar sensor to one of those and see how it goes. I think they are fairly standard, operating from the 5V supply in the ECU, so worth having a go. Putting the MAP sensor inside the ECU was maybe not the best choice since it does seem to fail.

I'll report back when I've done more miles and also done some tests with the modified ECU.

Hugh

[g259fsg]

Done about 370 miles now without incident. So far so good.

I took it round for an emissions test this afternoon, ahead of the MOT next month. Here are the results:

CO 0.45%
HC 216 ppm
Propane 463 ppm
CO2 14.3%
Lambda 1.032
O2 1.17%

The CO reading is quite low for a non-catalyst car. I did suspect this as the tick-over exhaust is very slightly "poppy" which is what you get on carburettors when the slow mixture is a bit lean (quite a good way to adjust carburettors). The readings with the previous ECU at the last MOT were CO 1.83% and HC 244 ppm.

I think I'll leave it alone until after the MOT as the guy said it would pass no problem. It drives well and doesn't stall, so maybe it's OK to leave the idle mixture where it is. Has anyone tried trimming the idle fuel level through the RoverMEMS Diagnostic App?

Are there any views on having a lower CO level? The Haynes Manual (above) lists the CO for my non-catalyst car as 0.5 to 2.0%, so I'm just on the bottom of the range.

Hugh

[g259fsg]

Here’s an update on this topic. I’ve now done about 850 miles on the replacement ECU and had no further problems, so the original fault does seem to have been due to the old ECU, most likely the MAP sensor.

The 214 had the MOT last week, and the emissions test results were pretty much the same as the previous test, possibly a bit on the low side.

CO 0.37%
HC 160 ppm

I’ve also followed up on the other idea of fitting an external MAP sensor to replace the faulty component on the ECU board, the Motorola 5141550T02. I’ve got a couple of ECUs with faulty MAP sensors – on the RoverMEMS diagnostic App, both showed a fixed 10 kPa, way too low. Although they idle OK, as soon as you tried to get some power from the engine to move off, it stalled.

I found a universal 1 bar MAP sensor from a company called EFI Parts which looked like it might be suitable https://efi-parts.co.uk/product/1-bar-map-sensor/. I assumed the Motorola device also used a 5V supply and would have a similar pressure to voltage ratio.

I deduced the likely three pin connections to the existing Motorola part. I opted to just cut the connecting wires to the faulty device on the board and remove it, rather than attempting to de-solder it. I think the ECU PCB is four layer, so there are two internal printed circuits that are only interconnected by the through-hole plating on the component holes. There's a slight risk of damaging the through-hole plating when de-soldering and removing the component legs, so I just left them in situ. I then soldered on three wires, red for +5V, black for GND (0V) and blue for the MAP signal. Cut a small hole in the edge of the ECU cover for the cable to come out.



I used some six-core cable intended for wiring alarm system which is flexible and small diameter. the PCB tracks are quite small so you need a small soldering iron.

The MAP sensor connects with a three pin plug/socket, so I joined these two parts with connector blocks. Before connecting the MAP, I installed the ECU and checked the voltages with the ECU switched on, confirming I had got the 5V supply. Here are a couple of pics of the complete installation.





Next the moment of truth. With ignition on but the engine not started, the MAP reading on the Diagnostic should be around 100 kPa. I got 102 kPa. With the engine started and idling, I got around 34 kPa, pretty much the same as I got with the working Motorola MAP sensor. So it looked like the EFI part was pretty well compatible without modifying the output.

I took the car for a run, and it behaved normally with plenty of power and acceleration. The idling tends to be a bit on the high side, but I've never run this particular ECU on the car before. It might settle down. I don't think the MAP has much input on the ECU idling parameters.

After the run, I took a set of readings using the RoverMEMS diagnostic:

RPM 924
Coolant 92 degC
IAT 46 degC
MAP 32kPa
Throttle 9.2%
Idle Control Valve Position 45
Ignition Advance 7.5
Coil Time 847 microsecs

So, initially this looks like a satisfactory way to repair the ECU. I'll need to run it for a while to see how it goes.

The question is, will I be brave enough to use it to drive down to POL later this month?? But I will have the other working ECU in the boot just in case.

[g259fsg]

Have just returned from a long trip starting with POL, so here's a final report on the performance of the modified ECU with external MAP sensor. I had the courage to use it to travel to POL and beyond. The first leg was from Dunfermline down to Longbridge, mostly on motorways at around 70 mph, though with some hold-ups. Following POL, I drove down to Sidmouth in Devon on the Saturday afternoon to visit my sister. A very smooth trip down to Exeter on the M5. My return trip was a more scenic route through the Mendip Hills to Bath and then on to Gloucester, mostly on country roads, and then up the Fosse Way to Coventry for an overnight stay. Finally up the M6/M74 home. A total of 1163 miles on motorways, country roads and some urban driving.

The performance seemed well up to expectations. In fifth gear on motorways it could quickly accelerate above 70 mph when required and maintained speed up Shap and Beattock, so the fuelling algorithm seems to be working OK. I measured the fuel economy by finding out how much petrol was required to completely refill the tank. On the first 819 miles, I used 79.4 litres which is around 17.46 gallons. This gives an average of 46.9 mpg.

Think I'll leave the experiment on the car a bit longer and see if there are any issues. So far, everything looks good. Maybe write some of this up in an article for Viking as there's been a lot of interest in the topic.

Hugh

[GTiJohn]

You really know your stuff

This may become a necessary mod for owners, so a write up would be very helpful.

Might you even consider a postal update service for members ?