Rich running problem

[adam_rbts]

Hi all, my griff seems to be running rich but I'm starting to run out of ideas what could be causing it. Hopefully somebody smart can see what I'm missing..

I know it's running rich from the soot spraying from the exhausts, the soot in the manifolds and the sooty plugs. I've also got a wideband O2 sensor installed (ready for MS install) which confirms that it's running approx 9:1 AFR at idle, dropping further under high load.
Idle is a little rough but it still pulls pretty strong through the revs. If I advance the timing at idle it becomes a little more stable and the AFR does increase very slightly.

Here's what I've checked so far (sensors measured at ECU plug):

Coolant sensor: all within spec in bible/haynes
AFM: 1.8v at idle
Fuel pressure: approx 38 psi @ idle (vacuum off)

O2 sensors: somehow I'm seeing a peak of around 2.0v on both sensors with digital multimeter. I've tried a scope on them but I get a quite noisy sine wave but still showing a peak of around 2.0v.

I've played around with the scope but I can't seem to get a better signal form the O2 sensors. What I don't know is are the sensors simply reacting to the rich conditions or are they part of the problem?

Would the lucas ecu revert to a safe map if it saw these kind of figures from the O2 sensors? If so would it really allow it to run so rich?

I've tried a spare coil and ignition amp (not new but spares that Ive used previously) with no change. Are ignition problems generally more noticable at idle or higher up the rev range?

Can anyone see anything I could be missing that would cause this? I'm getting to a point where I can't see the woods for the trees...

cheers

Adam

[adam_rbts]

Yup, forgot to mention, I tried your ecu but no difference.

If you want it back asap just click the 'Buy-it-now' option.

I'll even ship it for free..

Yeah earth problems sound plausible but being a clanky (not a sparky) like yourself I'm not sure of the best way to track this down..

[adam_rbts]

..another thing I forgot to mention in my haste:

I've tested both injector banks and they both measure 4.5ohms.

I've also tried a brand new set of injectors (bosch pintle type but same flow rate) with no change..

cheers

[pwd95]

I know you probably know full well but arent there 3 coolant temp sensors. One for the rad fans, one for temp gauge, & one for ECU. Did you get the right one?

Please correct me if I'm out.

[K4000JB]

hi just a thought, you said you have tried the fuel pressure with the vac disconnected but have you tried it with it connected, it the reading is the same then thats your problem (it should be about 7psi lower)and the reason why it seems to get better as you rev her
hope this helps

[adam_rbts]

K4000JB said:

hi just a thought, you said you have tried the fuel pressure with the vac disconnected but have you tried it with it connected, it the reading is the same then thats your problem (it should be about 7psi lower)and the reason why it seems to get better as you rev her
hope this helps

yeah that's right, it drops to about 32 with the vac connected as expected. It does run smoother through the revs but it's still really rich..

cheers

[cinquecento]

Would a shorted/open tune resistor do this?? Its located in the loom near the ECU.

[rev-erend]

Just junk it all and get on with the M/S install !

[adam_rbts]

rev-erend said:

Just junk it all and get on with the M/S install !

Yeah that was my thought too! I've actually had it running on MS but I don't want to be tuning out any non-ecu related fueling issue..

Actually I'm only able to get it to idle at about 12.5 AFR maximum with the MS anyhow..

[adam_rbts]

cinquecento said:

Would a shorted/open tune resistor do this?? Its located in the loom near the ECU.

Worth a look! cheers

[blitzracing]

The AFM voltage of 1.8 volts is a bit high I think, it should be nearer 1.5, although its not so high as to cause it to run so rich. I expect the ECU is in the get you home mode, as one of the sensor inputs is wrong (outside spec). Problem is you frequently cant tell due to either a lack of ECU warning light, or its not enabled to come on in the ECU firmware so try borrowing a fault code reader. The fuel pressure can be a long way out at idle, and the ECU can still pull the mixture back in range, as long as it gets the correct readings from the Lambda probes. You could try replacing the catalyst tune resistor with a 400 ohm one, so it runs the not catalyst fuel map as a test. You can then adjust the mixture at idle with the setting on the side of the AFM (This is turned off on the catalyst setting). This removes the lambda mixture control, so if it still wont drop into range, its likely to be AFM or temperature sensors. The ECU can also mistake a misfire for a lean mixture (to much exhaust oxygen) and will try and richen the mixture accordingly and incorrectly. If you have a test meter, check across the black (signal) and white wires on the standard lambda probes on tickover. If your wideband readings are correct, these will both be stuck over .7 volts. They should be switching at about half second intervals between .3 and .7 volts.

Mark

[adam_rbts]

blitzracing said:

The AFM voltage of 1.8 volts is a bit high I think, it should be nearer 1.5, although its not so high as to cause it to run so rich. I expect the ECU is in the get you home mode, as one of the sensor inputs is wrong (outside spec). Problem is you frequently cant tell due to either a lack of ECU warning light, or its not enabled to come on in the ECU firmware so try borrowing a fault code reader. The fuel pressure can be a long way out at idle, and the ECU can still pull the mixture back in range, as long as it gets the correct readings from the Lambda probes. You could try replacing the catalyst tune resistor with a 400 ohm one, so it runs the not catalyst fuel map as a test. You can then adjust the mixture at idle with the setting on the side of the AFM (This is turned off on the catalyst setting). This removes the lambda mixture control, so if it still wont drop into range, its likely to be AFM or temperature sensors. The ECU can also mistake a misfire for a lean mixture (to much exhaust oxygen) and will try and richen the mixture accordingly and incorrectly. If you have a test meter, check across the black (signal) and white wires on the standard lambda probes on tickover. If your wideband readings are correct, these will both be stuck over .7 volts. They should be switching at about half second intervals between .3 and .7 volts.

Mark

Thanks Mark. I've checked the lambda probes again and they're showing fixed at around 1.1v which confirms the wideband readings. Not sure how I was reading 2v the first time around but hey ho..

I'll try the non-cat tune resistor tomorrow to see if anything changes.

I'm using the same coolant sensor through the same loom when hooked up to the Megasquirt and I'm seeing pretty sensible temp readings so I'm pretty happy this is OK.

cheers

[blitzracing]

A simple test for the AFM output would be to remove the air pipe from and mesh off the input side, and observe the lambda outputs if you pass a finger infront the hot wire port with the car at tickover. The loss of airflow will drop the AFM output, and alter the fueling. The ECU should then try and re adjust the base fuel fueling, but we know it cant at the moment as the mixture is wrong for some reason.

Mark

[adam_rbts]

slight progress..

I've been looking into issues with the Innovate LC-1 controller on various forums. I suggest if anyone has one of these that they go take a look at some of the issues on Innovative's own forum.

Mine has been intermittently recalibrating itself (within the exhaust gases rather than free air) during cranking which has been causing spurious readings. I'm waiting to find out from Innovative whether my unit is one of a known faulty batch...

After recalibrating correctly, I'm now seeing more sensible idle AFR figures (12.5-13) but I still have a rich mixture issue, albeit less drastic than I'd first thought.

I've tried the method suggested above by Mark of swapping the tune resistor and playing with the AFM settings but with no joy.

Dumb question, but turning the trim screw on the AFM affects it's output voltage right? Just I'm not seeing any change when I turn it, either with the engine off (ignition on) or at idle. I'm still seeing about 1.8v to 1.9v at idle.

I guess the next step is to fit some shiny new plugs (coming tomorrow) and find a friendly local Griff owner who can lend me his AFM for an evening..?

Hi Quinny!

[catretriever]

I had a similar problem with my old Griff which turned out to be a bad connection for the lamda sensors inside the ECU which kept putiing it into limp mode.
I know you've swapped ECU's, but could it be a dodgy connector or something?

[blitzracing]

adam_rbts said:

slight progress..

I've tried the method suggested above by Mark of swapping the tune resistor and playing with the AFM settings but with no joy.

Dumb question, but turning the trim screw on the AFM affects it's output voltage right? Just I'm not seeing any change when I turn it, either with the engine off (ignition on) or at idle. I'm still seeing about 1.8v to 1.9v at idle.

Hi Quinny!

The setting on the side of the AFM is only a variable resistance thats basically across 2 pins on the ECU. You can do a simple restistance check with a DVM between AFM ground and CO trim output (outer 2 pins and AFM unplugged). Wind the pot and you should see the resistance change and it goes from around 1 - 1000 ohms. It has no bearing whatsoever on the AFM's output or scaling. The DC voltage change alters the base fueling the ECU starts with, so it can richen the weaken the mixture. Im not 100 % if this is only up to 2500 rpm as some have suggested or across the whole range. This is used on the non catalyst fuel map only, as the lambda feed back does the same job at tickover. If altering the voltage on the AFM on the non cat map does not alter the mixture, then at least you know its not lambda feedback problems making it run rich. That also discounts mis- fires. Its possible if the ECU is in fault mode the DC setting is ignored, but even when it is, it never runs as rich as 9:1. Ive only seen it around 13-14:1 in this state. As you get the same readings on both banks you can discount individual leaky injectors. If you part block the AFM hot wire port, does the engine idle / mixture change ?

Mark

[adam_rbts]

Right, had chance to borrow Quinny's AFM tonight (cheers Quinny!) and found no difference.

After reseting the ECU and firing it up with Quinny's AFM, the newly calibrated wideband sensor showed a good AFR figure of around 14 but after a minute or so it returned to 12-13 AFR as with my own AFR. I'm guessing it's dropped into limp-home mode at this point.

Measuring the narrowband O2 sensors I'm seeing around 1v at the sensor (confirming it's rich) but oddly I'm seeing 2.0v at the ECU plug (pins 23 to 4 and 24 to 4). I'd seen the 2v at the plug a few days ago but put it down to my dodgy measurement.

catretriever said:

I had a similar problem with my old Griff which turned out to be a bad connection for the lamda sensors inside the ECU which kept putiing it into limp mode.
I know you've swapped ECU's, but could it be a dodgy connector or something?

catretriever, can you recall what issues you had with you lambda connections? I've a feeling I could have a similar issue.

I'm thinking of bypassing the loom and wiring the O2 sensorss directly to the plug to see if that helps..

cheers all

[blitzracing]

Id suspect an earthing issue for this sort of voltage. The lambda cable is screened (like audio) as the signal levels are relatively low. To get an accurate signal back to the ECU, the braided part of the cable would need to be connected at each end, and not rely on the various earth paths through the engine and chassis. Worth doing a resistance check from the ecu earth pin, to the lambda earth, to make sure its no more than a fraction of an ohm. A lambda voltage stuck high will make the ECU drop into fall back. Also the ECU does not start making fueling corrections untill it sees a rising voltage from the probe after its warmed up (more noticable after a reset as the ecu has to reset the base fueling again), so that could account for the mixture changes. You need to borrow a fault code reader really.

Mark

[GreenV8S]

blitzracing said:

The lambda cable is screened (like audio) as the signal levels are relatively low. To get an accurate signal back to the ECU, the braided part of the cable would need to be connected at each end, and not rely on the various earth paths through the engine and chassis.

Don't entirely agree with that part. Screened cables are usually used for time-varying signals such as wheel/cam/crank position sensors where external noise makes it harder to lock on to the signal. The lambda signal only varies slowly and wouldn't normally justify screened cables. The standard lambda sensor cables aren't screened. If you were using a screened cable it would only be earthed at one end (usually the receiving end) to avoid creating an earth loop.

blitzracing said:

Worth doing a resistance check from the ecu earth pin, to the lambda earth, to make sure its no more than a fraction of an ohm.

That is a good idea.

[adam_rbts]

All sorted!!

Turned out the lambda sensor earth cable at the ECU plug wasn't connected to anything. The sensors themselves were earthed, that's why I was able to read the correct voltage at the sensor, but this earth didn't connect to pin 4 in the ECU plug so the ECU couldn't see what was going on..

Idle AFR is now where it should be and she now pulls like the proverbial!

Many thanks to all who chipped in to help diagnose