The MGF Register Forums

Hi, Hope someone can help me with this one.

I have a random EPAS (power steering) failure light and no power steering. It has happened about a dozen times in the last year.

Some times when driving. Sometimes does not cancel after self-check on start up.

2010 MG TF 135 (SAIC) . Fuses and Links removed and replaced. Does not have 40amp Fuse separate on wing which is a known issue on earlier cars. All fuses inside fuse boxes.

After every failure if you drive it for a short distance (50yds) and switch off and restart all will be ok again.

I have Pscan OBD2 diagnostics. No error codes stored. However here is the bit that has be baffled.

On Pscan live data the EPAS ECU is showing engine speed about 1000 rpm above true engine speed at 1915 rpm. Live data from MEMS3 (engine) ECU is showing true engine speed at 905 rpm.

I know that if EPAS ECU sees too many revs with too low speed it can trip out. So what do we think? Has anyone else seen live data from EPAS ECU and is it just a quirk that it shows RPM a 1000 higher that true or a fault and I ned a new EPAS ECU.

All thoughts gratefully received

Cheers, Mark

Wish I could help but I have no idea about the SAIC cars.

I don't think the SAIC cars were any different than Rover/MG cars when it comes to EPAS. There seem to be only 2 types depending on how the EPAS ECU received its speed data. The early F's had cable speedometers and later ones had electrical speed sensors driven by the ABS. As far as I know..........

SAIC just used up or remade RoverMG parts in the main and bolted them together at Longbridge. The only main differences were the N series engine and instruments.

Cheers Mark

I think Philip is away on hols at the moment, but it would be good to ask him his thoughts on what the pscan is seeing, but if the EPAS ECU is seeing twice the actual engine speed, then I guess that could throw some faults (the EPAS will go into temporary fail-safe if it detects high engine speed for a period of time without a corresponding road speed signal, as it will assume, correctly, that either the engine speed or road speed sensor is incorrect, and won't know what would be the most appropriate level of assistance to provide.

Rather surprised that this doesn't flag up as a retained fault code though?

Until you mentioned that this was a SAIC car without the underbonnet 40A fuse holder, that was going to be what I was going to point my finger at, as I had a similar intermittent fault caused by corroded connectors.

In this case, I'm pretty sure the problem is the engine speed signal value that you've identified. But why this is so high I am not sure? Is it always too high?

Hi Rob,

Yes, ran Pscan 3 times now and engine speed live data from EPAS ECU is about 1000rpm higher than from MEMS3 ECU. No faults logged. All other data looked right.

All elec plugs cleaned with contact cleaner and fuses replaced.

Will keep an eye out for a SAIC model EPAS ECU to swap out . They have a different part number from MG-Rover items but I do not know if this is just for a different parts system or if there is anything different inside the box. The late MG-Rover and SAIC EPAS ECU are both made by Mitsubishi and are Type G, just different part numbers. Just a bit wary of sticking a MG-Rover box in in case there is a difference. I found that the MEMS3 ECU is different on the SAIC to MG-Rover. Made by Continental and can not be programmed via OBD2 port. Mark at KMaps removed ECU and bolted it to his magic box to map it.

Good idea to ask Philip at Pscan, went to his site he only has live data view for early MGF, but that does show normal rpm. Tends to point to my ECU being dodgy.

Cheers Mark

I don't think Philip has tested the SAIC EPAS - there is a chance that it is a little different, but I think this unlikely. He's in Spain at the moment, but he will have a look for you.

Anyone else here with a SAIC MGTF and a Pscan who can have a look for Mark? Both my MGFs are early ones. And I've junked the EPAS completely from one of them!

markvrs wrote:I don't think the SAIC cars were any different than Rover/MG cars when it comes to EPAS. There seem to be only 2 types depending on how the EPAS ECU received its speed data. The early F's had cable speedometers and later ones had electrical speed sensors driven by the ABS...........

I recall later TFs having an electronic speed sensor on the gearbox which supplied the EPAS ECU via the speedo. May be mistaken though!

Yes - that is the road speed signal. However, it is the engine speed signal that appears to be wrong on Mark's car (or the pscan is incorrectly calibrated for SAIC cars - although why this would be the case is far from clear!)

The later MG Rover and SAIC TF seem to have the same inputs into the EPAS ECU for Road Speed and Engine Speed.

Road Speed direct from Speed Transducer T142 above L/H Drive Shaft on diff/gearbox

Engine Speed direct from Engine Control Module D131 (MEMS)

The confusing part is that both these signals are correct when going to either the speedo or the rev counter. Also live data from MEMS3 is correct for both. It is just live data from EPAS ECU that is showing incorrect engine RPM. Rob is spot on, is it pscan reading incorrectly or a faulty EPAS module?

Thanks for the views, dialogue has helped me learn what my car is doing, and how it does it. Much appreciated.

Cheers Mark

Just to finish off the post, in case someone searches in the future. Nothing worse than finding someone else had the same problem but did not post how they fixed it

Replaced my 2010 EPAS ECU with one from 2004 MG TF. (Chinese v British) Different part numbers but both type G. No compatibility issues.
Both showed engine speed at double actual engine speed so looks like it is a bug with Pscan software.

After 6 trips no EPAS failure so it seems ( I hope ) it was a faulty EPAS ECU, and the engine speed indication was a diagnostic bug not a real problem.

Thanks for posting a final up-date Mark, it helps to see how you resolved your problem. Hope all goes well in the future

Thanks Mark - great update. We should alert Philip of the problem...

I no longer have access to the local MGF that I used when I developed the EPAS code. It belonged to a fellow just down the road but last year he sold it.

Can anyone else confirm the same problem with live data on EPAS ?

Is pscan showing double the revs? or the actual revs plus 1000?

Fixing this in the software is relatively easy but it would be nice to be absolutely sure of getting it right.

If anyone with an MGF/TF happens to be in Ruislip and can pay me a visit then I could fix this quickly.

Hi Philip,
pscan shows double revs from Type G (late) EPAS ECU on TF's. The "1000 rpm" came from me doing live data at idle on cold engine. RPM gauge showing just under 1000 and pscan showing just under 2000. When I changed the EPAS ECU it showed the same issue at idle and when I rev'd it I then noticed it was actually double engine speed.

I noticed on your website that screen shot of MG F EPAS ECU live data shows correct engine speed (863rpm) Maybe a MG TF with MEMS3 issue?

Cheers, Mark

My MGF has the earlier EPAS system. I can check this. Has anyone got a Pscan and the later EPAS as found on post MY2000 MGFs & TFs?

If anyone with a suitable car is at MGF25 on Saturday maybe I can try it out? I will be on the SAWS stand.

Yes I've just replicated Mark's EPAS observation using vers.0432 on my SAIC TF.
I did a cold start and Pscan indicated an idle of ~2100 when rev counter showed a bit above 1100rpm.
Dunno if it's +1000 or doubled - but certainly reading high on mine too.
Hope this helps.

I need access to vehicle that has this to get to the bottom of it. Hopefully at MGF25.

I connected to an LE500 yesterday. The engine revs signal from the EPAS is reading double what the engine ECU says.
My current working theory is that this is nothing to do with the SAIC MGF/TF, but more to do with pre MEMS3 and post MEMS3.
My understanding is that pre MEMS3 the EPAS ECU got an engine speed signal from the instrument cluster, and post MEMS3 it got it from the engine ECU.
The fact that markvrs put a 2004 EPAS ECU in a SAIC car and it still works re-enforces this conclusion in my mind.
The EPAS software was developed on a pre MEMS3 car.
The good news is that the T4 worked correctly (well badly, but the engine RPM numbers were correct).
I just need to figure out how the T4 knows whether it's a pre-MEMS3 or post-MEMS3 and then it should be possible to correct this in pscan.

There are several different EPAS columns and EPAS ECUs.

Model-MGF to VIN XD511058 - Not Tilt Adjustable column- QMB101570 and Control Unit - YWC103773
(I believe this ties in with the move to MEMS3 and the digital speedo)

MGF from VIN XD511059 - Tilt Adjustable column - QMB102270PMP and Control Unit - YWC106740
MGTF - Tilt Adjustable column - QMB102270PMP and control unit - YWC000720
The later F and the TF share a column but have different control units. The only reason I can see for a different control unit is that the TF has a quicker steering rack than the F.

MG Motor LE 500 has yet a different column but I have no details of the control unit.

I may be able to put my hands on a couple of different ECUs, I will let you know when I have had a look.