MG-Cars.net

Welcome to our resource for MG Car Information.

Recommendations

Parts

MG parts spares and accessories are available for MG T Series (TA, MG TB, MG TC, MG TD, MG TF), Magnette, MGA, Twin cam, MGB, MGBGT, MGC, MGC GT, MG Midget, Sprite and other MG models from British car spares company LBCarCo.

MG MGF Technical - Engine warning light and smoke

Hi, Me again!

Just changed the radio on my new 2002 TF and it works well.

Later today I started the car, after a few minutes the engine started to misfire and the engine warning light started to flash.
I turned off the engine and noticed a strong burning smell. Upon opening the boot, clouds of smoke billowed out from somewhere in the engine compartment. After a couple of minutes the smoke cleared. The engine had not reached running temp. I daren't start her up again.
Any suggestions what the cause could be?

If the new radio had shorted somewhere it would have fused or not worked.

Mike (now beginning to regret the change from my trusted 13 year old VVC).

Mike
M King

I expect your alternator has seized and the smoke comes from the overheating drivebelt.
Charless

what warning light was it ? if it was the orange eml light it would idicate a engine misfire smoke could be from oil / water burning ? if it was the battery light yes poss alternator or electrical fault cant see how you caused either fault fitting a stereo ?
m hambidge

It was the flashing amber light with an icon of an engine block.

The engine was misfiring too, sounded as if running on 2 cylinders. The strong smoke could have been cables burning? I only ran the engine for less than a minute, so I doubt overheating.
Mike
M King

Flashing amber light likely to be triggered by unburnt fuel being detected at the cat, and the fact that it sounded like it was running on two cylinders would suggest a coil pack has failed spectactularly - each coil pack runs two cylinders, and the coil is one of the few things that could get hot enough to cause smoke that quickly. If you remove the cover strip in the middle of the rocker cover, I suspect you may find evidence of the source of the smoke.
bandit

Bandit to the rescue....(Again)

Thanks Bandit, I do hope your diagnosis is correct. My mind was racing last night thinking all sorts of terrible things!
Can I access the coil pack through the grill opening, or does it require removal of the parcel shelf?
Don't suppose you know of a web page with instructions?

Thanks again, Mike
M King

The engine access panel under the parcel shelf needs to come out for proper inspection. No special instructions particularly - across the middle of the valve cover is a plastic finishing strip, unscrew that and you'll see the coil packs. I'd expect there to be fairly obvious evidence if one has got hot enough to produce smoke but you may have to remove them to inspect the underside.


bandit

Spot on, bandit!

Quite a simple job to remove the obviously burnt out coil (See pic). Most challenging task was to unzip the siezed up rear screen!

My only concern is what may have caused it to burn out. As I said earlier, I had been fitting a new radio and had the the ignition switched on for several minutes without the engine running. Could that have been the cause?

Thanks again for you correct diagnosis, will have saved me quite a lot of cash.

Mike


M King

Mike

Cause? Difficult but suggest change the spark plug as well as the coil in case there was a short there
J Foulger

Glad to hear your car obliged by being broken in the correct way ;-)

Ignition on for a few minutes without cranking the engine shouldn't be taking the coil packs outside of 'normal' parameters, so I'd suspect it had already begun to fail before you swapped the radio. And better it fails on your drive than many miles from home, eh? Again, obliging car you have there!

The warning light may persist for the next few starts of the engine after the coil pack is replaced, or it may not - I'm not sure whether a cat distress signal triggers the memory function on the ECU. Either way, once you have 4 cylinders firing you can declare it fully fit again.
bandit

I am surprised to see the rusty bolt heads and general surface deterioration in this area. It looks as if it is old which may be indicating an underlying problem.
Geoff F.
Geoff Farthing

I fitted the new coil and started the engine. But still running on just 2 cylinders!

After a few seconds, I stopped the engine and felt the new coil, it was very hot. I removed it and it was too hot to hold and started to smell of burning.
I'm at a loss now as to what is causing the coil to short out / overheat.
Any suggestions would be welcome.

Mike
M King

Julian's suggestion of changing those two spark plugs is a good one, or at least taking them out to see if there's any defect visible, ideally swapping them to the other two cylinders to see if that causes the other coil pack to get hot, and while you're at it check the condition of the HT leads - any cracking of the insulation could be allowing the HT to short and thus put an unusually high load on the coil.

bandit

Thanks,

If there is just a 12v supply to the coils, and the problem coil is overheating, the supply must be OK.
I'm thinking the fault must lie from the coil to (and including)the plugs. There is no visible cracking or damage to the HT leads, although the pencil type plug connecters to cylinders 2 & 3 are cracked.

Of course, I could have damaged the new coil when it over heated.

So, another coil, 4 plugs, and an extended plug socket, and keep fingers crossed!

Mike.
M King

I have just had this response from another BB:

"Leaves the question whether the short is in the wiring (bad) or the ECU (a lot worse).
You can check this out if you disconnect both the coil pack in question and the main connector on the ECU and measure the switching wire (either white/black or white/red iirc) if this shows little or no resistance when checking against the body".

Is this a credible answer? If so, could you clarify the procedure in a more 'Basic idiot' language.

Also the LT wires in the pic are brown/yellow. Are these the wires in question?

Mike
M King

What Mykel is suggesting is disconnecting the plug connector from the suspect coil, and with a multimeter set to read Ohms check the resistance of each wire, with the other tester probe earthed on the body or engine block. With the ignition switched off, you should get infinite resistance i.e. no continuity from either wire. If all is well, current should not have any path beyond just the ignition circuit, if there's a short then it'll have a path into another circuit and that will show up. Do the same test with and without the ECU multi-plug connected and see if that produces different results.
bandit

Hi,
I connected the multimeter, with the ignition OFF.
I got a reading of 12v across the two LT wires, No resistance between the yellow and earth, and 12v between the orange and earth. Any ideas what that could indicate?

Mike
M King

What do you get from the wires to the other coil?

When you say no resistance, do you mean no continuity i.e. didn't show a voltage?

Can't find my workshop manual to check, but I'd be surprised if there should be a 12v supply to the coil when the ignition is off, as far as I know the ECU will switch on the constant feed to each coil when the ignition is switched on, and then starts firing the trigger charge when it gets an input from the crank sensor. If there's a circuit between the two wires with the ignition off I'd suspect that a switching component inside the ECU has failed. Not sure if that can be investigated by TestBook off the car or whether you'd need to get the car to a TestBook-equipped workshop.
bandit

Dieter's excellent site makes finding the workshop manual irrelevant - http://www.mgfcar.de/library/ENGINE_MANAGEMENT_SYSTEM_%20MPi_VVC_MEMS_3.htm

That confirms that the coil should only be getting a feed when the ignition is switched on, via the main relay. One way to confirm the 12v you're seeing at the coil is coming from the ECU would be to perform the same test with the multiplug disconnected from the ECU. If that produces no continuity i.e. voltage from either wire, the ECU is the culprit. If you still get voltage, the wires are shorting against something that gets battery voltage while the ignition is off, which narrows it down.
bandit

Hi,
Thanks, the 12v across the 2 LT wires was found after the ignition had been turned on, then turned off. This voltage dissipated to zero after a few minutes with the ignition off.

I'll check to see if there's any current with the multiplug disconnected in the morning. If the problem is within the ECU, can that be fixed? or should I think about getting my old VVC through the MOT test!

Mike
M King

I believe some faults inside the ECU can be tackled, but it's not designed to be serviceable so it'll be full of challenges. And really not worth trying in my opinion, there are enough TFs & late MGFs being written off that the supply of secondhand ECUs far exceeds demand so prices are low, and if you get both the engine ECU and the alarm ECU (and keyfobs) from the same car it's effectively plug'n'play, avoiding the need to have the 2 ECUs 'introduced' by TestBook. The EKA code will be the one for the donor car but if you can get the registration or VIN number for the donor you'll still be able to get the code should you need it.

Or you could just get an Emerald ECU, thus opening up a performance tuning path that will empty your wallet nicely ;o)
bandit

I think you are right.

If I can confirm it's a fault within the ECU, I'll have to get a replacement :-(

I take it the ECU is the box fixed to the rear bulkhead?

Is the replacement some thing I could do with, as you have realised, my limited knowledge?

Mike
M King

Just had a quick look on ebay for ECUs (there are quite a few listed).
But seem to be for steering, srs airbags, alarms etc..
Which is the one for my problem??

Mike
M King

The alloy enclosure on the rear bulkhead contains the engine ECU, and the alarm ECU is concealed behind the centre console, with the microwave unit hidden between the fold-down compartment and the T-bar cover. I'm fairly sure you don't need to swap the microwave unit, but maybe someone else will be able to confirm that..?

They're all very DIYable to remove & re-fit, the engine ECU is just a bit fiddly - elusive mounting bolts & an array of plug connectors to be disconnected and there's not much slack in the cabling. If both replacement ECUs came from the same car they'll already have the 'trust' relationship so once fitted & connected they'll work, and the associated keyfobs will work too. The replacement ECU may have been on the shelf a while exposing the contacts to atmospheric moisture so treat all the connection points (and the plug connectors on your car) with electrical contact spray (Maplins, Toolstation, some motor spares stockist have it too) to ensure they make a good connection when you plug it all together.

Check which part number you need for your car - http://www.rimmerbros.co.uk/Item--i-GRID000045
and then look for a set that includes the right engine ECU - http://www.ebay.co.uk/itm/mgtf-mgf-full-ecu-kit-include-ecu-1-x-fob-and-2-x-modules-mg136-/220932327759?pt=UK_CarsParts_Vehicles_CarParts_SM&hash=item337097c94f perhaps?
bandit

OR... http://www.ebay.co.uk/itm/160718447807?ssPageName=STRK:MEBIDX:IT&_trksid=p3984.m1426.l2649 but that'll probably cost you ten times as much, plus it'll start you on the path to fast road cams & verniers, DTH throttle bodies, and all sorts of fun on a rolling road getting it all set up, so all in probably 30 times as much...

bandit

Thanks bandit,

Am I right in thinking that if I have to replace the ECU, I will need both the Engine and Alarm units?

In the meantime, could you take a look at Mykel's listing at: http://www.mgfregister.org/forum/viewtopic.php?f=8&t=9354&p=76510#p76510. I would appreciate your comments.

Mike
M King

Mykel agrees there shouldn't be a circuit between the two LT wires when the engine isn't running, as the only time there should be voltage there is when the ECU has received an input from the crank sensor that tells it the engine is turning & where it is on the stroke cycle, and then only when the stroke is at the point where a spark is required. The fact that you were able to see a circuit between the two wires (even if it faded away fairly quickly) indicates all is not as it should be inside the ECU. If you were to do the same test on the wires of the working coil, I'd be fairly certain you'd not see a circuit/voltage at all.

Although there's a bit of work involved in swapping the engine ECU & the alarm ECU, bearing in mind they can be sourced for a pretty modest sum, that would be the route I'd take. Testbook diagnosis sessions won't be cheap and have logistical obstacles, whereas swapping the ECUs you can do yourself without the car moving. In many cases with electrical problems swapping bits isn't the smart route, but in this case I think it is - all the evidence is pointing at the engine ECU being the culprit.

If you can find an ECU set from a breaker that has both keyfobs it would be worth paying a bit extra, but one keyfob is enough for now, and your current keyfob can be re-programmed to the replacement alarm ECU as & when you like for not a huge cost.
bandit

mike, did you actually do the resistance check, or did you just check for volts? I'd go back and check for resistance between the switching wire and the earth. With the ignition off!!

And have you tried changing plugs and HT leads?
Leigh Reid

I disconnected the multiplug from the ECU and found infinite resistance between both LT wires and ground.

I guess it looks like the ECU is at fault. I managed to source one complete with fob and black security box.
Before I buy it, is there any thing else you can think of that may be causing the problem. i.e. cam sensor? main relay?
I have also noticed that the battery discharges very quickly once it is connected.

Mike
M King

The only point where the ignition cycle splits into two is at the ECU, and one coil pack was/is running fine, so the problem isn't with the inputs to the ECU, only with what the ECU is sending to the coil that melted.

The battery discharging quickly isn't surprising, it takes an awful lot of current to get a coil pack hot enough to fail, and when the alternator isn't spinning that'll all be coming from the battery.
bandit

Thanks bandit.

The battery discharges even with the ignition off. The only time the coil gets hot is when the engine is running (albeit on 2 cylinders).

Looks almost certainly the fault lies within the ECU.
Unless I hear otherwise, I'll send off for the ECU, security box and fob tomorrow.

Mike
M King

Oh dear...
Changed the ECU and black security box and they linked together ok.

Started the engine, it just ran on 2 cylinders again and the same coil became very hot. This car has been off the road for 2 months now and I'm just beginning to grow a teeny weeny bit despondent!

Any ideas what could be causing this problem would be welcome.


Mike
M King

Mike

I am surprised this thread is still going.

Did you replace the spark plugs as I suggested early on?
Julian S Foulger

Err no, do you think it could be that simple? I'll renew the plugs and kick myself if that's the solution!

Mike
M King

It seems to me that you have done everything else!

I assume you have checked the cabling carefully for shorts?
Julian S Foulger

Hi Julian,
Yep, checked the cabling with a tester. Had a look at the plugs, they seem OK, but I will fit 4 new ones when they arrive.

Mike
M King

No harm in fitting new plugs.

Next option is a good Auto Electrian.

Best of luck!
Julian S Foulger

This thread was discussed between 05/01/2012 and 26/02/2012

MG MGF Technical index

This thread is from the archive. The Live MG MGF Technical BBS is active now.