14 CUX V8 won't run when hot.
Discussion
Apologies for polluting the TVR pages with Land Rovers, but it is the same engine and ignition system. I am at my wits end with this thing, and the next step is to throw the entire CUX system in the bin and Megasquirt it, unless someone has a bright idea.
I have a Defender 110 with a 4.6 non-serp V8 in it. It has always run the 14CUX hotwire system, and has always been "passable" rather than "great". The ECU has a Tornado 4.6 chip in it and sparks are provided by a 123 ignition distributor. It is running a catalyst map with lambdas. Other than that, it is a standard 3.9 system.
The symptoms of the problem are:
1) It starts perfectly, on the button, from cold. At this point, it idles beautifully and is perfectly drivable.
2) With Rovergauge plugged in, all of the readings are sensible - lambdas are saying "about right", engine temperature is sane, fuel temperature is sane. There are no error codes. The air mass and throttle data seem good and react in a sensible manner when the engine is revved.
3) As it warms up, it continues to behave normally.
4) When it gets hot (about 90c), it suddenly flips into a mode where the idle is flaky, and there is zero power - basically the engine dies, it has the ability to idle, and maybe move the truck at walking pace, but it is knocking and banging in the exhaust. You can spin it up to 3000 rpm and it sounds fine, but there is no power at all.
Let it cool down again, and it is perfect.
I've checked the sparks with an oscilloscope, as I assumed it was a "warming up" issue with something electronic, but the sparks are fine - nice profile, no change between poor running and good running.
I've changed the AFM with a spare, no difference
I've disconnected the Lambdas to provoke it into the limp map - no difference.
Anyone hand anything like this - the only thing similar seems to be the "swapping of fuel and water sensors" - but mine seem right. They both start at ambient, the engine temp goes up to ~90 and the fans come on, and the fuel temp rises to about 25 with the bonnet open.
I have a Defender 110 with a 4.6 non-serp V8 in it. It has always run the 14CUX hotwire system, and has always been "passable" rather than "great". The ECU has a Tornado 4.6 chip in it and sparks are provided by a 123 ignition distributor. It is running a catalyst map with lambdas. Other than that, it is a standard 3.9 system.
The symptoms of the problem are:
1) It starts perfectly, on the button, from cold. At this point, it idles beautifully and is perfectly drivable.
2) With Rovergauge plugged in, all of the readings are sensible - lambdas are saying "about right", engine temperature is sane, fuel temperature is sane. There are no error codes. The air mass and throttle data seem good and react in a sensible manner when the engine is revved.
3) As it warms up, it continues to behave normally.
4) When it gets hot (about 90c), it suddenly flips into a mode where the idle is flaky, and there is zero power - basically the engine dies, it has the ability to idle, and maybe move the truck at walking pace, but it is knocking and banging in the exhaust. You can spin it up to 3000 rpm and it sounds fine, but there is no power at all.
Let it cool down again, and it is perfect.
I've checked the sparks with an oscilloscope, as I assumed it was a "warming up" issue with something electronic, but the sparks are fine - nice profile, no change between poor running and good running.
I've changed the AFM with a spare, no difference
I've disconnected the Lambdas to provoke it into the limp map - no difference.
Anyone hand anything like this - the only thing similar seems to be the "swapping of fuel and water sensors" - but mine seem right. They both start at ambient, the engine temp goes up to ~90 and the fans come on, and the fuel temp rises to about 25 with the bonnet open.
I don’t think I was clear on the sensors. There is a well known failure mode where someone mixes up the coolant and fuel temperature sensors - this makes the car think it is cold and runs incredibly rich. Mine seem to be fine: my coolant sensor (as per Rover gauge) starts at ambient and climbs gradually to about 90c when the fans come on. My fuel sensor starts at ambient, and gets to 25 - 35 depending on whether the bonnet is open.
I had very similar symptoms on a 1992 Griffith 4.0. I replaced the ECU on a very Rover knowledgeable persons advice and that cured it. Something internal to the ECU must have failed. I would have thought you could borrow a known good one from someone to test the theory. Make sure its one in the same date range as yours and has the removable chip.
geeman237 said:
I had very similar symptoms on a 1992 Griffith 4.0. I replaced the ECU on a very Rover knowledgeable persons advice and that cured it. Something internal to the ECU must have failed. I would have thought you could borrow a known good one from someone to test the theory. Make sure its one in the same date range as yours and has the removable chip.
Hmmm, that's not a good thought. If the ECU is dead, then it will be a megasquirt job. I'll have to find someone locally with a 3.9 I can pillage for a morning. My current plan is:
- Spoof the temperature sensor with a suitable resistor that says "85C". Just in case the sensor is occasionally breaking the connection.
- Check the fuel pressure. No reason to think it is wrong, but worth checking
- Test with someone else's ECU if I can find one
Failing that, strip the whole lot off and put something modern on.
Steve_D said:
Where in UK are you?
I have a spare ECU we use in our workshop for this type of test plus we have seen a failed ECU on a customers car.
We are Fareham south coast.
steve
Thanks for the offer - I’m in Berkshire, which would be quite a schlep! There is a Land Rover place down the road that might have one on the shelf. I have a spare ECU we use in our workshop for this type of test plus we have seen a failed ECU on a customers car.
We are Fareham south coast.
steve
The coil - yes, it was my first thought, mainly because Rover Gauge shows absolutely nothing changing between good and bad behaviour. But Rover Gauge doesn’t know about sparks. I’ve got a spare coil, so I’ll put that on the “try it out list” - but I’m not hopeful as the coil is out on the inner wing, so not really exposed to heat. The ignition amp (as such) is in the 123 ignition distributor, I’ll put the scope on it and see if there is anything odd, but the spark profile looked fine. The only bit that points to the distributor is the fact that this only seems to happen when the reported distributor temperature hits 60C. 55C - its all fine.
Voltage - everything is reporting fine, both in the distributor and ECU, and the engine is crap whether the fans are off or on.
rxe said:
Hmmm, that's not a good thought. If the ECU is dead, then it will be a megasquirt job. I'll have to find someone locally with a 3.9 I can pillage for a morning.
My current plan is:
- Spoof the temperature sensor with a suitable resistor that says "85C". Just in case the sensor is occasionally breaking the connection.
- Check the fuel pressure. No reason to think it is wrong, but worth checking
- Test with someone else's ECU if I can find one
Failing that, strip the whole lot off and put something modern on.
I went back to my RoverGauge data logs when I was experiencing similar symptoms. Have you run RoverGauge to log the data of the car driving to this cut out point and then converted the 'raw' data into a MS Excel spreadsheet? When I did this on my Griffith I looked through all the data and found the point of failure ie the cut out point. The data showed the Mains Voltage spike to 4,000 volts from the normal 13-ish volt running. It appeared something was causing an ECU mis-read. With some help from someone with more 14CUX knowledge than me, they suggested a new EPROM chip first. I did that and no change, still got the mis-fire. Next was trying a replacement ECU and the problem went away.My current plan is:
- Spoof the temperature sensor with a suitable resistor that says "85C". Just in case the sensor is occasionally breaking the connection.
- Check the fuel pressure. No reason to think it is wrong, but worth checking
- Test with someone else's ECU if I can find one
Failing that, strip the whole lot off and put something modern on.
My recommendation is get the vehicle warmed up then take it for a drive with the data logger running and drive until it exhibits the symptoms. Download the data and convert it into Excel and clean it up and try and identify the time of the failure and check the data around that point. Maybe the chip has gone bad, or maybe its the ECU like mine. Keep us posted.
Running it with logging on is a good shout, I'll try that, I'm pretty handy with Excel so I should be OK! It might show something up.
Timing - I have run various ignition profiles in the distributor. My first change was to go back to a very conservative map from my normal over advance at idle. Made no difference at all.
I can put a trace on the coil switching voltage, and the coil output, and log all of that too.
Timing - I have run various ignition profiles in the distributor. My first change was to go back to a very conservative map from my normal over advance at idle. Made no difference at all.
I can put a trace on the coil switching voltage, and the coil output, and log all of that too.
rxe said:
Running it with logging on is a good shout, I'll try that, I'm pretty handy with Excel so I should be OK! It might show something up.
Timing - I have run various ignition profiles in the distributor. My first change was to go back to a very conservative map from my normal over advance at idle. Made no difference at all.
I can put a trace on the coil switching voltage, and the coil output, and log all of that too.
FWIW I too am 4.6 with 123 DistributorTiming - I have run various ignition profiles in the distributor. My first change was to go back to a very conservative map from my normal over advance at idle. Made no difference at all.
I can put a trace on the coil switching voltage, and the coil output, and log all of that too.
My Ignition map is as below. (Note it was dyno mapped fuel and ignition at this on similar 4.6 with Turbo so has curve extension into +ve pressure part of the map, anyway up to STP is what works on mine which is NA.)
Taking 2 steps backwards-
You cant have the same readings between failed and working- if the HT is failing you will get large amount "adding fuel" trim as the lambda probe mis reads unburnt fuel as a lean mixture so the ECU adds as much fuel as it can to try and get the probe to switch. The devil will be in the detail so to speak- what row is the fuel map in between working and failing at the same RPM with no load? As the airflow is the major control of the injector timing we need to start here. External temperature fuelling is added after the fuel amount is set by the map against airflow. Also check the RPM reading is stable in RoverGauge as a failing amp with give you erratic or no RPM reading
HT wise , remove the king lead and tape the end about 1cm away from the chassis and crank the engine and observe the spark. It should be able to jump this gap easily. I would not exceed this gap as the bigger the gap, the bigger the spikes you get in the coil primary that glitches the amp so can stress the internal semiconductors. 1 cm relates to about 10kv arc.
Where are you in Berkshire?
You cant have the same readings between failed and working- if the HT is failing you will get large amount "adding fuel" trim as the lambda probe mis reads unburnt fuel as a lean mixture so the ECU adds as much fuel as it can to try and get the probe to switch. The devil will be in the detail so to speak- what row is the fuel map in between working and failing at the same RPM with no load? As the airflow is the major control of the injector timing we need to start here. External temperature fuelling is added after the fuel amount is set by the map against airflow. Also check the RPM reading is stable in RoverGauge as a failing amp with give you erratic or no RPM reading
HT wise , remove the king lead and tape the end about 1cm away from the chassis and crank the engine and observe the spark. It should be able to jump this gap easily. I would not exceed this gap as the bigger the gap, the bigger the spikes you get in the coil primary that glitches the amp so can stress the internal semiconductors. 1 cm relates to about 10kv arc.
Where are you in Berkshire?
Somewhere between Ascot & Maidenhead…
I had time for a quick play today, and I thought I had it, but no. Fuel pressure was high - 3 bar. So I turned down the regulator and got 2.5 bar with vac disconnected, and 2 bar with vacuum connected, per the spec. Pressure was rock steady, so I awaited warm up again … and it was still crap. So, not the fuel pressure.
Mark - interesting thoughts on the map. From memory (and this is a bit anecdotal because I wasn’t really looking for it) when it is happy, it is in the second row of the map. The “dark box” is hunting around left & right depending on the revs, but stays around row 2. When it is unhappy, it is down much further, but it is hard really to see as the revs are generally on the floor by then.
It’s got to be the damn sparks hasn’t it? I will try your king lead experiment when I get a chance. Off oh holiday next week, but I have the whole of Christmas off for fiddling!
I had time for a quick play today, and I thought I had it, but no. Fuel pressure was high - 3 bar. So I turned down the regulator and got 2.5 bar with vac disconnected, and 2 bar with vacuum connected, per the spec. Pressure was rock steady, so I awaited warm up again … and it was still crap. So, not the fuel pressure.
Mark - interesting thoughts on the map. From memory (and this is a bit anecdotal because I wasn’t really looking for it) when it is happy, it is in the second row of the map. The “dark box” is hunting around left & right depending on the revs, but stays around row 2. When it is unhappy, it is down much further, but it is hard really to see as the revs are generally on the floor by then.
It’s got to be the damn sparks hasn’t it? I will try your king lead experiment when I get a chance. Off oh holiday next week, but I have the whole of Christmas off for fiddling!
Gassing Station | Chimaera | Top of Page | What's New | My Stuff