16 replies to this topic

[northy]

I'm having a problem with my ECU.
It's a 95 spi with a stage one exhaust and a worked head from a mightymini, and sometimes is has a bit off oomf.
Sometimes it feels like a badly tuned 998.
It feels like the ECU is auto-calibrating the fuel map every time i start up+drive, the car varies in power quite a bit, even with the same throttle position and load, over about a couple seconds each time. It actually feels like someone's messing with the fuel mix screw on a carb as i drive...

Just wondering if this was normal, it seems logical to me that the ECU should have a map already built in and that it'd know how much fuel to put in at all load levels/revs/throttle position/temp/air temp etc. Also there's a blue/green wire under the steering column which is frayed/slightly burnt, don't know if that's anything to do with it.

If I were to guess what it was, i'd say that the permanent live to the ECU which presumably keeps the flash memory running was faulty. But that's me using computer knowledge rather than actual ECU knowledge.
Also the exhaust is always really sooty and my fuel economy is shocking, around 20mpg.

Any ideas?

EDIT: Spelling, sorry :/

Edited by northy, 17 July 2011 - 04:44 PM.

[Ethel]

The mems system rover used is actually quite sophisticated and is able to adjust itself to compensate for varying conditions, like a blocked air filter for example. It won't adjust that quickly though so suspect a faulty sensor or wiring feeding the ecu duff info. Most will be measure voltage drop across a variable resistor sensor, so poor wiring can also alter the resistance.

Search techie posts by Sprocket - he's very good on SPI's

[northy]

Cheers i'll do a quick visual on the sensors tomorrow, unless sprocket gives me some more definitive info.
I've looked through a few threads and found little in the way of people with the same problem. Is it definitely not normal though?

[icklemini]

sounds more like a sensor issue to me.... probably the MAP pipes messing around... or if really unlucky the MAP sensor in the ECU itself...

[Sprocket]

If its ony happening with the engine cold, it might be the manifold heater. if the heater is not working you get slight hesitation on acceleration.

Check the wiring is connected properly, also check the relay pack connector and the PCB tracks inside the pack for scorching.

[northy]

Just to add a spanner into the works, I had it plugged into a code reader, but it read no fault.

:S

[Sprocket]

Just to add a spanner into the works, I had it plugged into a code reader, but it read no fault.

:S

No, that's correct. The code reader cannot read any faults with the human using it

[northy]

But will it not indicate if the manifold heater or any of the sensors are nacked?
Don't want to replace that relay pack, having looked at the price...

[Sprocket]

If only life was that simple.

Unfortunately diagnosing the 'fault' is all down to the person using the 'tool' to read the available information to compare against measured readings and with a bit of experience on what to look for, deduce where the problem is likely to lie.

The ECU will not loose its memory.

To check if the manifold heater is working easily, you will need a Hall effect DC current meter and start the engine cold, then measure the current draw on the brown/ blue wire at the relay pack. If it draws a fair chuck in amps and reduces as the engine warms up, the heater is probibly ok, if it draws little amps and changes little as the engine warms up, there is something wrong with either, the relay pack (noted by scorching on the large connector pins, or scorched tracks on the PCB inside the relay pack (common), the wiring (either brocken or not connected onto the heater itself (common), or the heater itself has burnt out.

The other way to check the heater is less accurate and difficult to get access to but perhaps a quick way to determine whether the heater might be broken or not. Using an Ohm Meter, measure the resistance between the spade terminal in the center of the heater (on the underside of the manifold), and the body of the manifold itslef. with the heater cold, the resistance will be low (cannot for the life of me remember what the actual values are. 10's to 100's Ohms) and with the heater hot, the resistance will be much higher (1000's Ohms). if the resistance is infinate cold, the heater is dead.

Obviously check the simple things first like the wiring and the relay pack (remove the cover with lt disconnected of course, and have a look)

You could be clever and disconnect the large relay pack connector, find the pin on the loom end connector with the brown/ blue wire and measure the resistance between there and the manifold, thus measuring the whole circuit. if the resistance is infinate, there is a problem, but do check the wiring is connected!!

[icklemini]

But will it not indicate if the manifold heater or any of the sensors are nacked?

Nope sadly not, you could have the map pipes disconnected, blocked water pipes, low fuel pressure, wonky temp sensors and a few other things and no 'Fault' will be logged...

As sprocket said ^^^ you need to look at the data coming back to the reader.. not just the codes...

[northy]

It's not a problem that only occurs when the engine is cold though, it's almost random, but sometimes i drive and it never happns and the rest of the time it's fairly common.
It's like when, with a carbed car with an incorrect needle, when you put your foot down it burbles a bit and it seems to have MORE power as you return the throttle to closed, so strange.

That makes me think either MAP pipes, Injector, temp sensors or lambda sensor, because it seems to be over-fuelling. Any idea on how to check the lamba sensor or how to check for map sensor problems?

[northy]

I have located a single plug ecu from a cooper, if i fit this, will i have to have my immobiliser re-coded, or is that only for the later, twin plug varieties?

[icklemini]

Any idea on how to check the lamba sensor or how to check for map sensor problems

I use a code reader to be honest....

Lambda sensor can be checked with a voltmeter... MAP sensor readings cannot though and a code reader is needed...

[northy]

Right, I've tracked down a guy who has kindly agreed to let me use his fault finder, meanwhile i took apart the map pipes and elbows and discovered lots of oily, nasty liquid?
Very strange, also the elbow for the ecu MAP input thingy was very loose around the pipe, good ol' zip-tie sorted that out.
So i took the drastic action of fitting a home-made filter box (made of a plastic box with foam in it) between the crankcase breather system and the throttle body, as I figured oily badness might have been interfering with the map reading?
Seems to be a bit better now, but seen as my initial problem was intermittent, I have literally no way of telling...

[northy]

The crypton reported this:
Oxygen sensor circuit failure
Air inlet circuit failure

So I guess that means change the lambda sensor and the air temp sensor?
any clues of anything else air inlet circuit could refer to?