XR4x4 bad misfire and uncommunicative ECU - help!
Discussion
After 70,000 faultless miles my beloved XR4x4 2.9 EFI has suddenly begun misfiring badly. It starts up and ticks over perfectly, and it drives fine at speed, but if I touch the throttle at around 2,000 revs to accelerate, the engine dies and the car judders like a kangaroo.
I've changed all the obvious culprits - new spark plugs, air filter and fuel filter, and I've cleaned the distributor, rotor arm and HT leads and coil connections. I've adjusted the valve clearances, and the compression is around 11 bar in all cylinders. I've also checked the alternator output, but it all makes no difference. It's not the petrol (BP Ultimate) because I've used half the tank without incident.
I assume the fault is somewhere in the injection or engine management system or its sensors or the ignition module. I've tried to interrogate the ECU with a Gunson 5-pin fault finder, but I don't get any response at all (I've checked out the fault finder - it actually contains just two resistors, a diode, a lamp and a switch, all of which work OK). Can anyone point me in the right direction, as I don't want to dismantle and replace everything.
I've changed all the obvious culprits - new spark plugs, air filter and fuel filter, and I've cleaned the distributor, rotor arm and HT leads and coil connections. I've adjusted the valve clearances, and the compression is around 11 bar in all cylinders. I've also checked the alternator output, but it all makes no difference. It's not the petrol (BP Ultimate) because I've used half the tank without incident.
I assume the fault is somewhere in the injection or engine management system or its sensors or the ignition module. I've tried to interrogate the ECU with a Gunson 5-pin fault finder, but I don't get any response at all (I've checked out the fault finder - it actually contains just two resistors, a diode, a lamp and a switch, all of which work OK). Can anyone point me in the right direction, as I don't want to dismantle and replace everything.
Several things spring to mind. First of all the ECU plug. There are 2 sockets which look identical! 1 is the ECU diagnosis socket, and the other controls the standard ignition retard and idle speed. I think the ECU one is next to the battery, and the idle/retard one is next to the HT coil (although the position may well change). I used to diagnose mine with a bit of wire and an LED with a resistor soldered to it (so much for being a professional electronics engineer!)
What year is it? If it is one of the few K reg 2.9s it may have a catalyst which can block, and an oxygen sensor which may have failed.
Some other things to try: The vane airflow meters may be going faulty. If you are handy with a multimeter you may be able to compare the output of one versus the other. Also check the connections haven't oxidised over the years. Another thing to try is the throttle position sensor. Either measure the resistance with the engine and ignition off, or the voltage output with the ignition on and engine off (not running). It should be a smooth progressive output. As with the airflow meters, check the connections for oxidation. It is probably worth checking the crank position sensor and connections, although I'm not sure of it's location on the 2.9.
Also check the vacuum pipes around the engine, particularly the one to the manifold air pressure (MAP) sensor. They can easily get blocked in a car this age.
What year is it? If it is one of the few K reg 2.9s it may have a catalyst which can block, and an oxygen sensor which may have failed.
Some other things to try: The vane airflow meters may be going faulty. If you are handy with a multimeter you may be able to compare the output of one versus the other. Also check the connections haven't oxidised over the years. Another thing to try is the throttle position sensor. Either measure the resistance with the engine and ignition off, or the voltage output with the ignition on and engine off (not running). It should be a smooth progressive output. As with the airflow meters, check the connections for oxidation. It is probably worth checking the crank position sensor and connections, although I'm not sure of it's location on the 2.9.
Also check the vacuum pipes around the engine, particularly the one to the manifold air pressure (MAP) sensor. They can easily get blocked in a car this age.
Thanks Zad, you've given me a lot of useful information to experiment with. However, since posting, I've reseated one of the HT leads that wasn't firmly home in the distributor cap, and I've sprayed the HT leads and injection wiring with WD40, which seems to have solved the problem (fingers crossed). I'm now going to treat Black Beauty to a new set of leads - Whoooa!
I'm still looking to solve the mystery of the dead diagnostic socket. I have been plugging in to a 5-pin socket dangling by the coil (only 3 pins are actually used) so maybe I've been using the wrong one - I'll search by the battery for an alternative, and report back.
PS The car is early 1989 - one of the first 2.9 models, and it still has a couple of legacies from the 2.8, eg the positioning of the fuel filter.
I'm still looking to solve the mystery of the dead diagnostic socket. I have been plugging in to a 5-pin socket dangling by the coil (only 3 pins are actually used) so maybe I've been using the wrong one - I'll search by the battery for an alternative, and report back.
PS The car is early 1989 - one of the first 2.9 models, and it still has a couple of legacies from the 2.8, eg the positioning of the fuel filter.
Zad, if you're still with this one you're a master mechanic. I found the correct diagnostic socket bundled with the leads on the positive battery terminal. After 18 years Black Beauty has spoken to me and says "11 11 1", which means all-clear, so I don't have to fit a new ECU. However . . .
For the static ignition test, the Gunson fault finder manual says to start with the Gunson switch open ('0'), turn the ignition on, then close the switch to '1' (which grounds the control pin) and wait 10 seconds for the light to flash. What actually happened was that I turned the ignition on, which immediately illuminated the Gunson light, and then the light flashed off and on to signal "all-clear". This is totally the wrong way round, because I had not closed the switch, and the light should start off then flash on, not start on and flash off.
I tried turning the ignition on with the Gunson switch closed ('1'). The light flashed on and off once. I opened the switch ('0'), and the light immediately came on and then flashed off and on to signal "all-clear".
I'm mystified at this, and I'm reluctant to carry out a live engine test until I understand it. The fault finder was bought cheap in a clearance from Halfords, and was a return. I wonder if it has been repaired incorrecly. The internal wiring is as follows: the live pin is wired to an LED which is wired through a 1.6K resistor and a diode to the earth pin, which is also wired through a 3.3 resistor and switch to the control pin. Needless to say, I have checked the pin connections have been inserted in the correct socket holes as per the Gunson manual. Any thoughts?
For the static ignition test, the Gunson fault finder manual says to start with the Gunson switch open ('0'), turn the ignition on, then close the switch to '1' (which grounds the control pin) and wait 10 seconds for the light to flash. What actually happened was that I turned the ignition on, which immediately illuminated the Gunson light, and then the light flashed off and on to signal "all-clear". This is totally the wrong way round, because I had not closed the switch, and the light should start off then flash on, not start on and flash off.
I tried turning the ignition on with the Gunson switch closed ('1'). The light flashed on and off once. I opened the switch ('0'), and the light immediately came on and then flashed off and on to signal "all-clear".
I'm mystified at this, and I'm reluctant to carry out a live engine test until I understand it. The fault finder was bought cheap in a clearance from Halfords, and was a return. I wonder if it has been repaired incorrecly. The internal wiring is as follows: the live pin is wired to an LED which is wired through a 1.6K resistor and a diode to the earth pin, which is also wired through a 3.3 resistor and switch to the control pin. Needless to say, I have checked the pin connections have been inserted in the correct socket holes as per the Gunson manual. Any thoughts?
Zad, I've had another look inside the Gunson fault finder.
I said previously that the switch is open in the '0' position and grounds the control pin in the '1' position. It's actually wired the opposite way round ie '1' is open. This doesn't sound right to me, but there is no evidence that the switch soldering has been tampered with. However, the incoming live and control leads do look as if they have been re-soldered, and I'm wondering if they have been reversed.
I've been assuming that the apex hole of the diagnostic socket is the live feed to the LED, and the bottom right hole is the control to be grounded. It would be enormously helpful if you could recount which holes you attached your lamp and other wires to in order to provide the control signal, and I can then tell if my fault finder is wired similarly.
I said previously that the switch is open in the '0' position and grounds the control pin in the '1' position. It's actually wired the opposite way round ie '1' is open. This doesn't sound right to me, but there is no evidence that the switch soldering has been tampered with. However, the incoming live and control leads do look as if they have been re-soldered, and I'm wondering if they have been reversed.
I've been assuming that the apex hole of the diagnostic socket is the live feed to the LED, and the bottom right hole is the control to be grounded. It would be enormously helpful if you could recount which holes you attached your lamp and other wires to in order to provide the control signal, and I can then tell if my fault finder is wired similarly.
Further to my previous posts, I've not yet been able to source a new set of HT leads as Halfords don't have them in stock, and my misfire has been returning intermittently. So .... I've taken the old leads off to give them a good clean and resistance check, and lo and behold I've found that no 5 cylinder lead is dead, so I've been running on five cylinders, which probably explains all the symptoms.
PS I still haven't bottomed how the Gunson fault finder is supposed to work, so if anyone has had any success with it please contribute.
PS I still haven't bottomed how the Gunson fault finder is supposed to work, so if anyone has had any success with it please contribute.
Hmm. This may help: http://www.corral.net/tech/maintenance/eecivtest.h...
When I had to source a front transfer/propshaft for mine, I had a rummage around the small ads in the Ford magazines. That might be a good place to source new plug leads. It might *just* be worth ringing your local Ford spares department. It is amazing how much old stuff they keep or can source.
When I had to source a front transfer/propshaft for mine, I had a rummage around the small ads in the Ford magazines. That might be a good place to source new plug leads. It might *just* be worth ringing your local Ford spares department. It is amazing how much old stuff they keep or can source.
Thanks for the info Zad.
As regards the HT lead, I've ordered one from my usual supplier whom I heartily recommend. He is a specialist Sierra breaker called Rob Smith who operates through eBay under the moniker 'sierrashack1', and can supply anything from gearboxes to obscure pieces of trim within a couple of days. He can be contacted through eBay, and likes payment via PayPal, so if you haven't signed up to eBay yet you're missing out.
In relation to the Gunson fault finder, I've read all of the articles accessible via Google, and they and the Gunson instructions all say slightly different things, and all differ from my own experience. I'm convinced that my switch is mounted the wrong way round, so I'm going to thoroughly explore the operation of the ECU's self-diagnostic routine by trial and error, and lay it out once and for all.
As regards the HT lead, I've ordered one from my usual supplier whom I heartily recommend. He is a specialist Sierra breaker called Rob Smith who operates through eBay under the moniker 'sierrashack1', and can supply anything from gearboxes to obscure pieces of trim within a couple of days. He can be contacted through eBay, and likes payment via PayPal, so if you haven't signed up to eBay yet you're missing out.
In relation to the Gunson fault finder, I've read all of the articles accessible via Google, and they and the Gunson instructions all say slightly different things, and all differ from my own experience. I'm convinced that my switch is mounted the wrong way round, so I'm going to thoroughly explore the operation of the ECU's self-diagnostic routine by trial and error, and lay it out once and for all.
Right, I've now worked out by trial and error how to access the EEC IV ECU self-diagnostics for a 1989 XR4x4, so here's how to do it:
The 5-hole diagnostic socket dangles from the leads to the positive battery terminal, and is protected by a rubber cover. Only three holes are wired, and the other two are blanks. The apex socket is the diagnostics output, and pulses from 0v to 5v relative to the bottom left hole, which is common earth. The right hole is the control terminal, which must be grounded to earth to initiate self-testing.
The best way to interrogate the socket is using a Gunson Ford 5-pin fault finder. Paradoxically, this comes with just three loose pins - brown for output, green/yellow for earth and blue for control. It has an LED readout and a switch for initiating tests. Be aware that Halfords is currently selling off these finders for only £5, and mine, if not all of them, was faulty in that the switch is mounted upside down in the casing with the result that '0' is incorrectly on and '1' is off. You need to check with a voltmeter that '1' connects the control pin with the earth pin, and, if it does not, you need to unscrew the casing, unsolder the switch, push it out, turn it round, push it back in and then resolder the leads. Now plug the Gunson pins into the correct diagnostic socket holes and set the switch to '0'. The accompanying Gunson instructions are wildly inaccurate, so use these instead:
STATIC IGNITION TEST
Turn on the ignition (not the engine) and listen for the fuel pump. The LED will pulse on and off once. Flip the switch to '1', and the LED will light up. It will now start pulsing off and on to communicate codes that are defined in the Gunson Ford fault finder instructions ('11' is the code for all-clear). These are then repeated, and the test will end with a single isolated pulse. There is no fault history memory and no wiggle test facility. Now turn the ignition off and return the switch to '0'.
CONTINUOUS LIVE ENGINE TEST
With the switch in the '0' position, turn the ignition on and start the engine. The LED will pulse rapidly - these are computer readable diagnostics that should be ignored. Turn the switch to '1'. The LED will light and then continously pulse any error codes or '11' for all-clear. Whilst this is happening, suspect sensor connectors can be wiggled to see if they produce an error. Once you've seen enough, turn off the ignition.
LIVE ENGINE LOAD TEST
With the switch in the '1' position, turn the ignition on and start the engine. The LED will light - ignore any rapid pulses whilst the engine speed rises to 2500 rpm under the control of the ECU and then drops back down. The LED will now signal with a single pulse off. Immediately floor the accelerator and release. The LED will now pulse any resultant error codes or '11' for all-clear. Turn off the ignition once the LED has stopped signalling and remains continuously on. Job done.
Y'all be pleased to know that Black Beauty has had a replacement HT lead fitted, and is running sweetly again. The above diagnostics have all resulted in 'all-clear', so over and out.
The 5-hole diagnostic socket dangles from the leads to the positive battery terminal, and is protected by a rubber cover. Only three holes are wired, and the other two are blanks. The apex socket is the diagnostics output, and pulses from 0v to 5v relative to the bottom left hole, which is common earth. The right hole is the control terminal, which must be grounded to earth to initiate self-testing.
The best way to interrogate the socket is using a Gunson Ford 5-pin fault finder. Paradoxically, this comes with just three loose pins - brown for output, green/yellow for earth and blue for control. It has an LED readout and a switch for initiating tests. Be aware that Halfords is currently selling off these finders for only £5, and mine, if not all of them, was faulty in that the switch is mounted upside down in the casing with the result that '0' is incorrectly on and '1' is off. You need to check with a voltmeter that '1' connects the control pin with the earth pin, and, if it does not, you need to unscrew the casing, unsolder the switch, push it out, turn it round, push it back in and then resolder the leads. Now plug the Gunson pins into the correct diagnostic socket holes and set the switch to '0'. The accompanying Gunson instructions are wildly inaccurate, so use these instead:
STATIC IGNITION TEST
Turn on the ignition (not the engine) and listen for the fuel pump. The LED will pulse on and off once. Flip the switch to '1', and the LED will light up. It will now start pulsing off and on to communicate codes that are defined in the Gunson Ford fault finder instructions ('11' is the code for all-clear). These are then repeated, and the test will end with a single isolated pulse. There is no fault history memory and no wiggle test facility. Now turn the ignition off and return the switch to '0'.
CONTINUOUS LIVE ENGINE TEST
With the switch in the '0' position, turn the ignition on and start the engine. The LED will pulse rapidly - these are computer readable diagnostics that should be ignored. Turn the switch to '1'. The LED will light and then continously pulse any error codes or '11' for all-clear. Whilst this is happening, suspect sensor connectors can be wiggled to see if they produce an error. Once you've seen enough, turn off the ignition.
LIVE ENGINE LOAD TEST
With the switch in the '1' position, turn the ignition on and start the engine. The LED will light - ignore any rapid pulses whilst the engine speed rises to 2500 rpm under the control of the ECU and then drops back down. The LED will now signal with a single pulse off. Immediately floor the accelerator and release. The LED will now pulse any resultant error codes or '11' for all-clear. Turn off the ignition once the LED has stopped signalling and remains continuously on. Job done.
Y'all be pleased to know that Black Beauty has had a replacement HT lead fitted, and is running sweetly again. The above diagnostics have all resulted in 'all-clear', so over and out.
Gassing Station | Ford | Top of Page | What's New | My Stuff