AFM reading widely fluctuating in Rovergauge

[taylormj4]

On way home Friday, I opened the throttle widely quite quickly on a hill. I usually open it more progressively.
This was greeted by a big stutter and rough running.
At light throttle openings, normal service resumed.
I pulled over and the engine starts running rough again - needed extra throttle to stop it stalling.

Drove quietly home OK but then on a sharp bend, engine cuts out. Engine restarts and gets me home running badly.

Next morning, put Rovergauge on and the reading from the AFM is fluctuating wildly from normal to maximum and back again. This seems to be causing the fuel map to flip from its usual row to the bottom row and back again, which is causing very rough running and from the smell of it, overfuelling.

Oddly, no AFM related fault code was shown on RG but the MIL light on RG did light. What causes the MIL light to come on in RG ?

The AFM fluctuations presumably suggests knackered AFM or connector to it ?
Pretty annoying as I had this happen around 2 years ago and replaced the AFM with a brand new genuine LR unit at significant cost.

I'm going to also look at the wiring/connector to the ECU. Could it be the ECU itself?

Thanks for any suggestions as this is a daily drive car.
Cheers.

[blaze_away]

Easy stuff first. Pull the plug off the AFM and give it c9ntacts a clean and WD40 spray.
Then reset ECU by either removing the ECU plug oŕ disconnecting the battery.
Re connect and see how things are.
RG has a "show fault codes" under "options"

[taylormj4]

blaze_away said:

Easy stuff first. Pull the plug off the AFM and give it c9ntacts a clean and WD40 spray.
Then reset ECU by either removing the ECU plug o? disconnecting the battery.
Re connect and see how things are.
RG has a "show fault codes" under "options"

Yes, it was the RG fault codes that I was talking about and the RG MIL light indication.

Tried cleaning connectors - used proper contact cleaner. Pulled AFM plug and teh ECU plug. AFM plug is actually relatively new as I replaced it at the same time as the AFM around 2 years ago.

The ECU connection clean seemed to help a bit - not sure whether that's becassuse of the clean or the ECU reset via disconnection - not running quite as rough now maybe but RG still showing fluctuations in the AFM output. The engine runs properly when the AFM reading drops to zero. Can it be zero ?

Edited by taylormj4 on Sunday 26th March 19:06

[blitzracing]

Id double check what the AFM is actually doing with a test meter - any wild swing in voltages will show up very easily. Dont forget RoverGauge is reporting just what the ECU sees- so you could have something like a bad earth on the ECU causing the voltage on the ECU ground to move. Are all the other readings stable?

AFM output:

Red/black- Ground

Blue/Green -Output voltage

[PhilH42]

Its a real pain...I'm on number 5 in the space of 18 months. First one failed and I put it down to being old and replaced for aftermarket which was subsequently replaced three times for free, I put the first two of these down to poor quality and then realised there was a definite problem when another failed. Seems to happen after a lay up and then fails quickly after driving again. I now disconnect if it sits for any time but I'm chasing the problem.

Symptoms of mine seem to be misfire at around 3500 revs but have differed slightly each time. I've had cutting out at idle and when giving it full throttle only but the worst one was almost undrivable. Last one showed fault code 19 on rovergauge.

Can you think of any situation that may have been similar between your first failure and this one (potentially) ?

Agree with Blitz...multimeter is the best way to trace whats going on with yours and my next job.

Edited by PhilH42 on Sunday 26th March 20:53

Edited by PhilH42 on Sunday 26th March 20:57

[taylormj4]

PhilH42 said:

Its a real pain...I'm on number 5 in the space of 18 months. First one failed and I put it down to being old and replaced for aftermarket which was subsequently replaced three times for free, I put the first two of these down to poor quality and then realised there was a definite problem when another failed. Seems to happen after a lay up and then fails quickly after driving again. I now disconnect if it sits for any time but I'm chasing the problem.

Symptoms of mine seem to be misfire at around 3500 revs but have differed slightly each time. I've had cutting out at idle and when giving it full throttle only but the worst one was almost undrivable. Last one showed fault code 19 on rovergauge.

Can you think of any situation that may have been similar between your first failure and this one (potentially) ?

Agree with Blitz...multimeter is the best way to trace whats going on with yours and my next job.

Edited by PhilH42 on Sunday 26th March 20:53

Edited by PhilH42 on Sunday 26th March 20:57

Yes odd. My first one lasted nearly 18 years. I replaced it with a Genuine LR AFM at considerable cost and expected it to last similarly. In fact I made up a small heat deflector for it to lengthen its lifespan - made from sheets of foil, looked rubbish but did the job. However, helpfully the garage that did the last service binned it and I haven't replaced so now kicking myself.

Car has been running slightly rough and smelling rich on start up for a while but I thought it was just new ignition components needed; I put a post up a week or so ago asking for recommendations on ignition leads and new dizzy parts.

Now, I suspect the AFM was going but the lambdas were giving enough correction for correct fueling. RG shows they are both at -100% constantly except for when you snap open the thottle. So they are now not able to correct it seems.

Only similarity to yours seems that a sudden opening of the throttle appears to be what finally finished the AFM off.

What actually breaks on these? Is it the hot wire itself that breaks or something else. Are they user-repariable ?

[taylormj4]

blitzracing said:

Id double check what the AFM is actually doing with a test meter - any wild swing in voltages will show up very easily. Dont forget RoverGauge is reporting just what the ECU sees- so you could have something like a bad earth on the ECU causing the voltage on the ECU ground to move. Are all the other readings stable?

AFM output:

Red/black- Ground

Blue/Green -Output voltage

Cheers Mark,
Didn't think a digital meter would be fast enough to pick up the fluctuations but I'll give it a go.

All other readings look stable:
Throttle position, short-term lambdas as both -100%, long terms are in the centre as I've just reset them after unplugging the ECU.

Where is the earth for the ECU ?

[taylormj4]

PhilH42 said:

Its a real pain...I'm on number 5 in the space of 18 months. First one failed and I put it down to being old and replaced for aftermarket which was subsequently replaced three times for free, I put the first two of these down to poor quality and then realised there was a definite problem when another failed. Seems to happen after a lay up and then fails quickly after driving again. I now disconnect if it sits for any time but I'm chasing the problem.

Symptoms of mine seem to be misfire at around 3500 revs but have differed slightly each time. I've had cutting out at idle and when giving it full throttle only but the worst one was almost undrivable. Last one showed fault code 19 on rovergauge.

Can you think of any situation that may have been similar between your first failure and this one (potentially) ?

Agree with Blitz...multimeter is the best way to trace whats going on with yours and my next job.

Edited by PhilH42 on Sunday 26th March 20:53

Edited by PhilH42 on Sunday 26th March 20:57

Phil,
Which replacement units have you tried please? I was going to avoid those that you found to fail quickly but there doesn't seem to be much choice.

Has anyone tried using a company that can do refurbishments, such as ATP ?

Anyone tried refurbishing their own ?

[blitzracing]

taylormj4 said:

Cheers Mark,
Didn't think a digital meter would be fast enough to pick up the fluctuations but I'll give it a go.

All other readings look stable:
Throttle position, short-term lambdas as both -100%, long terms are in the centre as I've just reset them after unplugging the ECU.

Where is the earth for the ECU ?

I dont know where its bolted on the TVR from the 14CUX Loom, but if everything else is stable it wont be the main earth anyway. Id think a DVM will catch the voltage unless its spiking really sharply.

[PhilH42]

taylormj4 said:

Phil,
Which replacement units have you tried please? I was going to avoid those that you found to fail quickly but there doesn't seem to be much choice.

Has anyone tried using a company that can do refurbishments, such as ATP ?

Anyone tried refurbishing their own ?

I don't know how much I can honestly help you out with that. I've tried both Second hand OE and new from the main tvr parts suppliers...IME neither choice has been better or worse than the next in terms of how long it lasts. I wouldn't buy cheap from ebay personally but either of the other choices are probably ok in normal circumstances.

It ran best on both OE's though...noticeably.

[PhilH42]

Oh and I took one apart...delicate is the word I'd use to describe the hotwire.

[taylormj4]

PhilH42 said:

Oh and I took one apart...delicate is the word I'd use to describe the hotwire.

Did you manage to repair it ?

Found a bit on line about a guy refurbishing a flapper type AFM, cleaning up the resistance tracks but the hotwires won't have that I suspect.

Is it worth opening one up ? Can someone competent with a multimeter and a soldering iron refurbish one ?
What actually fails due to the heat from the manifolds, just dry joints ?

[blitzracing]

You have a circuit board covered in op-amps from the dark ages, so I even when you do pull them to bits I dont know if you could replace the semiconductors anyway. The hot wire itself is not really the issue here, but its connections are spot welded to the board connections and wont re solder if you remove the PCB that makes life difficult. I've never managed to get a schematic for these units or Id be quite happy to re-solder a few chips once I could tell what was failing, but its difficult fault finding without a diagram.

You can get refurbs- these should work fine on catalyst cars:

http://www.ebay.co.uk/itm/LUCAS-HOT-WIRE-AFM-AIR-F...

If you are precat- you would need to add a 5k trim pot to the loom for the CO trim.

[PhilH42]

taylormj4 said:

Did you manage to repair it ?

Found a bit on line about a guy refurbishing a flapper type AFM, cleaning up the resistance tracks but the hotwires won't have that I suspect.

Is it worth opening one up ? Can someone competent with a multimeter and a soldering iron refurbish one ?
What actually fails due to the heat from the manifolds, just dry joints ?

It was sacrificial to get a better understanding of how its put together.

I've got a heat shield on mine and have wrapped it in a jacket before and it still doesn't last that long, my kittens are out and so it was no hotter under the bonnet before the turbo than anyone elses, so I wouldn't say its just down to heat killing it or everyone would be changing them very regularly.

Electric fault finding is not my forte..I get by. I was wondering if there's a spike happening or whether the ignition relay has a short and is permanently feeding...battery will deplete in 4 or 5 days if not on trickle. I'm changing it anyway.

So you can see what its like inside the afm below. You will note that the hotwire is broken...I poked it gently at first to see if it had separated...it hadn't.

[davep]

PhilH42 said:

It was sacrificial to get a better understanding of how its put together...

... So you can see what its like inside the afm below.

FWIW The MAF clones seem to be based on a schematic similar to this:



The sophisticated part is the 8-pin device, which is probably a PIC:

http://ww1.microchip.com/downloads/en/devicedoc/41...

A reasonable design but very poor build quality.

[blitzracing]

Nothing that advanced in the 5 AM im sure- especially as the output is non linear. Id think it may be worth my while reverse engineering one if its £300 to get a referb'.

[PhilH42]

Not wanting to take away from the OP's remaining (potential) AFM failure but it's important that I update the thread after my previous postings.

I tried another OE AFM today that was proven sound and fault code 19 still comes back after resetting and unplugging the ECU!!

This now may point to a faulty throttle pot or wiring not AFM , but not yet proven!

So it would also seem that although I've had many aftermarket afm's fail, only the original OE one has failed so far!!

[taylormj4]

Thanks folks,
All interesting stuff - good to see inside one.

The fact that the first one lasted 16 years and the second one (new although marked as a 1997 unit) only 3 years suggests to me that the meters are degrading with time rather than from years of use. Either that or something degrades on the car that is causing conditions which make the AFM fail.
Interestingly, both AFM's have failed in the same way - rough running for a period and then a proper boot of the throttle caused very poor running and cut-out.

Oddly, Rovergauge is not showing any fault codes even thought the MAF reading is swinging all over the place with a hardly open throttle.

The MIL light on Rovergauge was up before I reset it. What conditions cause that to come on ?

[taylormj4]

Someone suggested I tried running with the AFM disconnected. Tried this. Very rough running.

Does this put it into limp mode ?

[taylormj4]

blitzracing said:

Id double check what the AFM is actually doing with a test meter - any wild swing in voltages will show up very easily. Dont forget RoverGauge is reporting just what the ECU sees- so you could have something like a bad earth on the ECU causing the voltage on the ECU ground to move. Are all the other readings stable?

AFM output:

Red/black- Ground

Blue/Green -Output voltage

Measuring the dc voltage across these two with engine running gave:

0.9-1.1V at around 900rpm
up to
1.3-1.5V at around 3000rpm

The reading seems relatively stable but the meter show regular fluctuations of around 0.1V, like it is pulsing up and down around 0.1V at a frequency of around 1-2Hz.
Is that correct ?