Lambda fueling problem. FAO shpub or anyone please
Discussion
Hi Didn't know whether to start a thread or ask Steve Heath and Mark Adams directly as i am still have an issue that is puzzling me, having recently bought steves ecumate and thought i'd fixed my problems.
On cruise and on tickover both the lambdas read correctly and afr is reading stoik, then for no apparent reason and this could be after a 5min period of running correctly it would all of a sudden run constantly super lean on the afr and the ecumate box says BOTH lambda reading 255X (which means out of range rich and ecu locked out) its like a typical lazy lambda failure yet its on both lambdas and symaltaniously and its not all the time. ??
Mark mapped the car a fortnight ago and one of the lambdas was replaced then as it was faulty so i cant believe its the lamdbas at fault as it is also affecting both prety much simaltanious, i see A lock out at 255X almost instantly followed by B.
All other readings seem to be ok, and giving a little acceleration sometimes clears it and they read correctly for half a minute then go off the scale again.
The afr wide band is clearly reading very lean and the car is definately running poorly when this happens, so i dont understand why the narrow band lambdas arent also reporting back its lean and not rich i could understand it if i was getting different reading from lambda A & B.
Any clues or suggestions on what to check would be very much appreciated
Cheers
H
On cruise and on tickover both the lambdas read correctly and afr is reading stoik, then for no apparent reason and this could be after a 5min period of running correctly it would all of a sudden run constantly super lean on the afr and the ecumate box says BOTH lambda reading 255X (which means out of range rich and ecu locked out) its like a typical lazy lambda failure yet its on both lambdas and symaltaniously and its not all the time. ??
Mark mapped the car a fortnight ago and one of the lambdas was replaced then as it was faulty so i cant believe its the lamdbas at fault as it is also affecting both prety much simaltanious, i see A lock out at 255X almost instantly followed by B.
All other readings seem to be ok, and giving a little acceleration sometimes clears it and they read correctly for half a minute then go off the scale again.
The afr wide band is clearly reading very lean and the car is definately running poorly when this happens, so i dont understand why the narrow band lambdas arent also reporting back its lean and not rich i could understand it if i was getting different reading from lambda A & B.
Any clues or suggestions on what to check would be very much appreciated
Cheers
H
something to check out.What revs is this happening?
You say cruise but what is the throttle pot doing? What value is it?
The narrow band sensors do not react in the same time as the wide band so unless everything is absolutely steady, they will often report differing values. Their only use is to tell the ECU if its either too rich or too lean. Roughly. Sort off. They don't give a quick response.
So comparing the narrow band to an AFR is generally not good as it can lead to a lot of misconceptions. Particularly as the lambda control is frequently switched off by the 14CUX in response to engine revs, sudden changes in engine load (supercharger?), changes in the throttle position and so on. Blipping the throttle is enough to disable it.
Where is the AFR actually located in the exhaust? This can cause problems/differences as well.
The ECUmate reports back the lambda value but that on its own is not enough information. It also looks at the changes and indicates whether the unit is oscillating or not . If it is oscillating/switching then it is operational. If it doesn't then that indicates either lambda control is switched off, or that the mixture is really out of control or there is a fault. The lambda sensors are fed by a common heater supply. If that fails then that can cause problems.
Given that the AFR (and assuming its right) says it is running lead I would go for lambda control suddenly being disabled or some other fault. This could be down to a noisy pot sending a signal which the ECU interprets incorrectly. This might explain why moving the throttle restores normality.
If the throttle/butterfly is worn, that can cause problems as it can move and effectively lean out the mixture and cause the throttle pot value to suddenly jump. Made worse with a blower as the pressure on the plate can be quite high.
I suspect that the lambda is actually being disabled rather than hitting rich. I know from the testing I did that the values go to defaults when the lambda control is switched off and those defaults were very much dependent on which eprom revision and so on.
The other point is that I don't know what Mark A has done to mapping. It is made more difficult because it is a supercharged system and to be honest I am not sure if Mark has had to some tricks to get it to work at all.
I would look at the AFM values and the mapping sites and see if there is any change here before or after the lean running. If these are changing then that would indicate that something is going wrong and the ECU is trying its best. Are the water and fuel temps correct?
You say cruise but what is the throttle pot doing? What value is it?
The narrow band sensors do not react in the same time as the wide band so unless everything is absolutely steady, they will often report differing values. Their only use is to tell the ECU if its either too rich or too lean. Roughly. Sort off. They don't give a quick response.
So comparing the narrow band to an AFR is generally not good as it can lead to a lot of misconceptions. Particularly as the lambda control is frequently switched off by the 14CUX in response to engine revs, sudden changes in engine load (supercharger?), changes in the throttle position and so on. Blipping the throttle is enough to disable it.
Where is the AFR actually located in the exhaust? This can cause problems/differences as well.
The ECUmate reports back the lambda value but that on its own is not enough information. It also looks at the changes and indicates whether the unit is oscillating or not . If it is oscillating/switching then it is operational. If it doesn't then that indicates either lambda control is switched off, or that the mixture is really out of control or there is a fault. The lambda sensors are fed by a common heater supply. If that fails then that can cause problems.
Given that the AFR (and assuming its right) says it is running lead I would go for lambda control suddenly being disabled or some other fault. This could be down to a noisy pot sending a signal which the ECU interprets incorrectly. This might explain why moving the throttle restores normality.
If the throttle/butterfly is worn, that can cause problems as it can move and effectively lean out the mixture and cause the throttle pot value to suddenly jump. Made worse with a blower as the pressure on the plate can be quite high.
I suspect that the lambda is actually being disabled rather than hitting rich. I know from the testing I did that the values go to defaults when the lambda control is switched off and those defaults were very much dependent on which eprom revision and so on.
The other point is that I don't know what Mark A has done to mapping. It is made more difficult because it is a supercharged system and to be honest I am not sure if Mark has had to some tricks to get it to work at all.
I would look at the AFM values and the mapping sites and see if there is any change here before or after the lean running. If these are changing then that would indicate that something is going wrong and the ECU is trying its best. Are the water and fuel temps correct?
Edited by shpub on Wednesday 14th December 13:17
Hi Simon,
The Bosch afm appears to be giving correct readings on the ecumate and the issue is happening engine hot or cold (worse when its cold and it cant hold tickover when its that lean). Sometimes stop starting the ignitions seems to clear it. It did it last weekend and after a half hour drive with it working sometimes but mostly not, stopped for 15mins and it worked perfectly for the return journey and the following 2 days, then it started again yesterday.
The Bosch afm appears to be giving correct readings on the ecumate and the issue is happening engine hot or cold (worse when its cold and it cant hold tickover when its that lean). Sometimes stop starting the ignitions seems to clear it. It did it last weekend and after a half hour drive with it working sometimes but mostly not, stopped for 15mins and it worked perfectly for the return journey and the following 2 days, then it started again yesterday.
Thanks Steve, I'll write down the other readings for you when i get the car out later.
As for the wideband/narrow band speeds it is irrelevant as the narrow bands stop cycling and are locking out for prolonged periods at all points throughout the lambda controlled rev range and only changes with sufficient throttle pot angle to take the ecu off of lambda control otherwise you can speed up slow down or just leave it on tickover and no change in the A & B reading(and no cycling),and then suddenly they start cycling again and the mixture comes back to normal readings sometimes for 30 secs or so and sometimes for the rest of the journey.
MA went through all the tp and afm readings last week, but i will check again whilst it is in fault mode again.
As for the wideband/narrow band speeds it is irrelevant as the narrow bands stop cycling and are locking out for prolonged periods at all points throughout the lambda controlled rev range and only changes with sufficient throttle pot angle to take the ecu off of lambda control otherwise you can speed up slow down or just leave it on tickover and no change in the A & B reading(and no cycling),and then suddenly they start cycling again and the mixture comes back to normal readings sometimes for 30 secs or so and sometimes for the rest of the journey.
MA went through all the tp and afm readings last week, but i will check again whilst it is in fault mode again.
Just been thinking...
It could be possible that the ECU is seeing the Lambdas go rich (in error) and then will try and trim this out. This will lean out the mixture which is what the AFR reports.
If this is the case then the ECU is using an incorrect value which is then upsetting the fueling as you see. As the fault is intermittent, it could be something like an electrical problem or something similar.
Might be worth measuring the actual lambda voltage. If this indicates lean while ECUmate says rich, I would suspect that the problem is ECU related - maybe something like a dry joint that is causing the ECU to misread the lambda sensor and hence the problem.
It could be possible that the ECU is seeing the Lambdas go rich (in error) and then will try and trim this out. This will lean out the mixture which is what the AFR reports.
If this is the case then the ECU is using an incorrect value which is then upsetting the fueling as you see. As the fault is intermittent, it could be something like an electrical problem or something similar.
Might be worth measuring the actual lambda voltage. If this indicates lean while ECUmate says rich, I would suspect that the problem is ECU related - maybe something like a dry joint that is causing the ECU to misread the lambda sensor and hence the problem.
Lean mixture is a low or no lambda voltage, and these are resistive probes that takes the voltage from the heater supply, so if the heater supply is intermittent this will produce a false reading. it would need to be a break in the loom somewhere as its a common 12 volt supply with the fuel pump and AFM (I think?) so if you lost power from these as well the fault would be very apparent.
blitzracing said:
Lean mixture is a low or no lambda voltage, and these are resistive probes that takes the voltage from the heater supply, so if the heater supply is intermittent this will produce a false reading. it would need to be a break in the loom somewhere as its a common 12 volt supply with the fuel pump and AFM (I think?) so if you lost power from these as well the fault would be very apparent.
That is certainly not what I understood? Heater element simply shares ground with the sensor, ECU is sampling against the ground side of the heater, not the positive feed side AFIK.spend said:
blitzracing said:
Lean mixture is a low or no lambda voltage, and these are resistive probes that takes the voltage from the heater supply, so if the heater supply is intermittent this will produce a false reading. it would need to be a break in the loom somewhere as its a common 12 volt supply with the fuel pump and AFM (I think?) so if you lost power from these as well the fault would be very apparent.
That is certainly not what I understood? Heater element simply shares ground with the sensor, ECU is sampling against the ground side of the heater, not the positive feed side AFIK.Steve/Spend, I had suspected a second ecu failure and have a spare comming from MA along with a fresh chip incase the original got spiked when the first ecu went. so will wait till that gets here before i freaze in the garage checking readings
Another possibility suggested to me is electrical noise, if any sensor wires are near ignition circuit or ht leads..I've already had this problem before when the TP lead were touching a spark plug lead, so thats another thing to check.
spend said:
blitzracing said:
Lean mixture is a low or no lambda voltage, and these are resistive probes that takes the voltage from the heater supply, so if the heater supply is intermittent this will produce a false reading. it would need to be a break in the loom somewhere as its a common 12 volt supply with the fuel pump and AFM (I think?) so if you lost power from these as well the fault would be very apparent.
That is certainly not what I understood? Heater element simply shares ground with the sensor, ECU is sampling against the ground side of the heater, not the positive feed side AFIK.The lambda voltages go up for an over rich mixture, so lean is 0 volts, and rich will be around 1.3 volts, but this is right on the limit of where the probe works, and changes as the probe ages so you cant say and AF ratio of say 12:1 will give 1.31 volts without calibrating individual probes. As for electrical noise, the cable feeding the lambda probes is shielded like audio cable within the 14CUX loom, but I dont think the TVR extenders are shielded so noise could be an issue if the extenders run near the HT or like. Under normal opertaion the probes switch slowly enough to give a reliable reading with just a test meter. I think monitoring the outputs would be a very good idea, and check the ground wiring as well, and see how many millivolts you have between the engine ground and the lambda ground. The whole ECU loom has earth wires running through it for all the sensors to prevent small voltage shifts in the engine ground plane from messing the sensor voltages.
The 12 supply for the lambda heaters is shared with the fuel pump, from the fuel pump relay, so the lambda heaters are only on as long as the engine is running, or during the first pulse sent to the pump on power on. The AFM and Injector 12 volt supply is switched from the main relay, so this is on with ignition on.
blitzracing said:
The ECU dose not provide any voltage across the resitive lambda probe, its taken from the heater + ve supply and I presume dropped down through an internal resistor within the lambda probe to allow you to get about 0- 1.3 volts as the sensor "output". The heater and sensor have a common ground.
The lambda voltages go up for an over rich mixture, so lean is 0 volts, and rich will be around 1.3 volts, but this is right on the limit of where the probe works, and changes as the probe ages so you cant say and AF ratio of say 12:1 will give 1.31 volts without calibrating individual probes. As for electrical noise, the cable feeding the lambda probes is shielded like audio cable within the 14CUX loom, but I dont think the TVR extenders are shielded so noise could be an issue if the extenders run near the HT or like. Under normal opertaion the probes switch slowly enough to give a reliable reading with just a test meter. I think monitoring the outputs would be a very good idea, and check the ground wiring as well, and see how many millivolts you have between the engine ground and the lambda ground. The whole ECU loom has earth wires running through it for all the sensors to prevent small voltage shifts in the engine ground plane from messing the sensor voltages.
The 12 supply for the lambda heaters is shared with the fuel pump, from the fuel pump relay, so the lambda heaters are only on as long as the engine is running, or during the first pulse sent to the pump on power on. The AFM and Injector 12 volt supply is switched from the main relay, so this is on with ignition on.
I don't really see why you worry about the voltage reading when they are measuring change in resistance?The lambda voltages go up for an over rich mixture, so lean is 0 volts, and rich will be around 1.3 volts, but this is right on the limit of where the probe works, and changes as the probe ages so you cant say and AF ratio of say 12:1 will give 1.31 volts without calibrating individual probes. As for electrical noise, the cable feeding the lambda probes is shielded like audio cable within the 14CUX loom, but I dont think the TVR extenders are shielded so noise could be an issue if the extenders run near the HT or like. Under normal opertaion the probes switch slowly enough to give a reliable reading with just a test meter. I think monitoring the outputs would be a very good idea, and check the ground wiring as well, and see how many millivolts you have between the engine ground and the lambda ground. The whole ECU loom has earth wires running through it for all the sensors to prevent small voltage shifts in the engine ground plane from messing the sensor voltages.
The 12 supply for the lambda heaters is shared with the fuel pump, from the fuel pump relay, so the lambda heaters are only on as long as the engine is running, or during the first pulse sent to the pump on power on. The AFM and Injector 12 volt supply is switched from the main relay, so this is on with ignition on.
I think you'll find the 'TVR' extenders are std LR parts since the sensors sit way back in the Y piece under the car on Discos, my 3.9 certainly has them and the Rangie exhaust is the same.
spend said:
I don't really see why you worry about the voltage reading when they are measuring change in resistance?
I think you'll find the 'TVR' extenders are std LR parts since the sensors sit way back in the Y piece under the car on Discos, my 3.9 certainly has them and the Rangie exhaust is the same.
As part of doing resistance checks you would also loose the feedback voltage to the ECU, so the whole lambda cycling would stop anyway so you would be none the wiser, hence sticking to voltage readings. Im sure you are correct about the extenders making little odds, its a red herring thinking about it. What we have is an ECU saying its seeing a lambda voltage of over say 1 volt (I don't what voltage when the A-D convertor hits 255) on both sides at the same time, so the ECU backs of the fuel as far as it can to compensate, assuming as Steve has pointed out its not in open loop for some reason. That either means the both lambda voltages have stuck high, (the ECU can take 12v BTW without damage as I found out when a heater wire shorted to the lambda output!) or as already suggested something is wrong with the ECU itself. A test meter on the lambda outputs should prove that one way or the other. The fact that the lambda cycling starts again with a blip of the throttle would imply its not a genuine lambda voltage fault. Example like a broken ground wire that could be common to both probes where the cable screens are joined to the ECU- I dont know if its wired this way however.I think you'll find the 'TVR' extenders are std LR parts since the sensors sit way back in the Y piece under the car on Discos, my 3.9 certainly has them and the Rangie exhaust is the same.
Update;
Decided to scive off work today and try to fix, with success
Turns out that the first issue was crank sensor and tp lead getting spiked by the proximity to ht lead...I re-sheathed all loose wires in conduit and tie wrapped away from ht leads which cured some issues, but the main one was a leak in the intercooler pipework which formed a pressure seal that only opened up on boost.
Interesting note for future ref; cux will believe the afm reading over both lambdas.
Decided to scive off work today and try to fix, with success
Turns out that the first issue was crank sensor and tp lead getting spiked by the proximity to ht lead...I re-sheathed all loose wires in conduit and tie wrapped away from ht leads which cured some issues, but the main one was a leak in the intercooler pipework which formed a pressure seal that only opened up on boost.
Interesting note for future ref; cux will believe the afm reading over both lambdas.
Bluebottle said:
Update;
Decided to scive off work today and try to fix, with success
Turns out that the first issue was crank sensor and tp lead getting spiked by the proximity to ht lead...I re-sheathed all loose wires in conduit and tie wrapped away from ht leads which cured some issues, but the main one was a leak in the intercooler pipework which formed a pressure seal that only opened up on boost.
Interesting note for future ref; cux will believe the afm reading over both lambdas.
I think you will find that the AFM/AMM just as a larger range to influence fuelling over the narrow range of the Lambda probes Bluebottle, glad you fixed it Decided to scive off work today and try to fix, with success
Turns out that the first issue was crank sensor and tp lead getting spiked by the proximity to ht lead...I re-sheathed all loose wires in conduit and tie wrapped away from ht leads which cured some issues, but the main one was a leak in the intercooler pipework which formed a pressure seal that only opened up on boost.
Interesting note for future ref; cux will believe the afm reading over both lambdas.
Gassing Station | Griffith | Top of Page | What's New | My Stuff