Kangaroo Griff
Discussion
I first started this thread after a very annoying kangaroo drive into work. On the journey home however, the little devil exhibited completely different symptoms - It simply wouldn't tick over; it wasn't erratic, it just kept trying to tick over at about 3-400 rpm. The other bizarre circumstance was that the volt meter registered a constant 0.75 volt higher reading than normal (this was independant of load eg when the fans were on or off). This suggests that something wasn't drawing the usual current. That same evening I took her out again and she behaved herself impeccably. I think my missus may have diagnosed the problem when she pointed out that the car was left out in heavy rain the day before and damp got into something electronic.
Stepper moter has been suggested above but I fitted a new one a couple of months ago.
I suspect that damp gets into the Lambda connectors underneath the chassis which in turn confuses the ECU which maybe needs resetting by switching the engine off? I dunno.
Stepper moter has been suggested above but I fitted a new one a couple of months ago.
I suspect that damp gets into the Lambda connectors underneath the chassis which in turn confuses the ECU which maybe needs resetting by switching the engine off? I dunno.
I'm under the impression that there are only two states for the 14CUX - pre-cat or cat and each is selected by a different resistor value. If this is the case, then why does fitting a new Mod-wise resistor make an improvement, unless the original resistor is faulty?
Is this resistor a black plastic square thing with spade terminals at each ends, located in the harness next to the AFM? I have measured a volt drop across this, but never realised that it was the resistor that selected the cat/non-cat fuel map.
Is this resistor a black plastic square thing with spade terminals at each ends, located in the harness next to the AFM? I have measured a volt drop across this, but never realised that it was the resistor that selected the cat/non-cat fuel map.
Thanks for the info Mark, very interesting and helpful. But back to my original thread question; why does my Griff kangaroo, or misbehave in other ways when it gets wet?
Also, just for interest, what is the black square thing which has a volt drop across it, with spade terminals at each end situated close to the AFM? It has "Lucas" moulded into the case.
Steve
Also, just for interest, what is the black square thing which has a volt drop across it, with spade terminals at each end situated close to the AFM? It has "Lucas" moulded into the case.
Steve
I originally started this thread because my car was behaving diabolically in a kangaroo sort of way on one journey. It had recently rained, and my car often plays up after it's rained or been washed. Since then, it's been as right as rain (pun intended).
I conclude from this that damp has an effect on one or more sensor signals which causes the ECU to get its knickers twisted (I suspect the lambda sensors only because of their exposure and the exposure of their various connectors in the engine compartment).
If our systems are a bit sensitive to these signals, then the robustness of the signals needs to be sorted, changing fuel maps does not seem to be a good solution even though it may eliminate the symptoms.
It may be that the tolerance on the output of the sensors needs to be tighter than the manufacturers allowable tolerance to work satisfactorily with our systems.
I conclude from this that damp has an effect on one or more sensor signals which causes the ECU to get its knickers twisted (I suspect the lambda sensors only because of their exposure and the exposure of their various connectors in the engine compartment).
If our systems are a bit sensitive to these signals, then the robustness of the signals needs to be sorted, changing fuel maps does not seem to be a good solution even though it may eliminate the symptoms.
It may be that the tolerance on the output of the sensors needs to be tighter than the manufacturers allowable tolerance to work satisfactorily with our systems.
Thanks guys, yes I have checked the connectors to the Lamba sensors and they're as clean as a clean thing. And the road speed sensor seems to be OK, I fitted a new one in January which solved some problems I was getting, and those symptoms are different to this.
The odd thing is, the car plays up after it's got wet whilst stationary i.e. after I've washed it or it's rained whilst on the drive. I don't seem to get the problems when I'm driving through rain.
Does anyone else's do this?
The odd thing is, the car plays up after it's got wet whilst stationary i.e. after I've washed it or it's rained whilst on the drive. I don't seem to get the problems when I'm driving through rain.
Does anyone else's do this?
There is a chance that these symptoms could be caused by more than one problem, or combination of problems.
With my car, it's only occured once, and that was when it was wet. This I think, rules out any permanent arrangement like cam, AFM etc, and suggests that the most likely candidate is a control signal going a bit haywire.
I believe that the ECU works only on the values from one Look-up-table, but these values are modified by the various control signals fed back from the sensors. It is a fact that the 14 CUX changes its fuelling outputs by some kind of "learning" process - I believe that this is intended to take account of sensors with varying tolerances - but there are limitations. I have first hand experience of my ECU learning its way into a diabolical state due to an ignition problem such that when the ignition problem is fixed, the ECU got "stuck" in this diabolical state, and the simple fix was to re-set the ECU to get back to factory settings.
So it might be an idea to re-set your ECU if this is happening to you.
My understanding of the purpose of the Lambda sensor is to get the engine running in stoichiometric condition (ideal combustion where fuel is COMPLETELY burned), thus hydro-carbons are eliminated from the exhaust. But this only occurs at steady throttle and engine load conditions, the mixture is considerably richened when you plant your foot to the boards. I believe that the response time for typical modern Lambda sensors is in the region of half a second, and this period may very well correspond with the frequency of our kangaroo - what I'm getting at here is that maybe the lambda signal starts to oscillate for some unknown reason which then sets up a positive feedback (this means that when a problem starts, the feedback just adds to the problem i.e the kangaroo gets progressively worse very quickly) which then causes the kangaroo. Mine certainly felt like positive feedback behaviour.
With my car, it's only occured once, and that was when it was wet. This I think, rules out any permanent arrangement like cam, AFM etc, and suggests that the most likely candidate is a control signal going a bit haywire.
I believe that the ECU works only on the values from one Look-up-table, but these values are modified by the various control signals fed back from the sensors. It is a fact that the 14 CUX changes its fuelling outputs by some kind of "learning" process - I believe that this is intended to take account of sensors with varying tolerances - but there are limitations. I have first hand experience of my ECU learning its way into a diabolical state due to an ignition problem such that when the ignition problem is fixed, the ECU got "stuck" in this diabolical state, and the simple fix was to re-set the ECU to get back to factory settings.
So it might be an idea to re-set your ECU if this is happening to you.
My understanding of the purpose of the Lambda sensor is to get the engine running in stoichiometric condition (ideal combustion where fuel is COMPLETELY burned), thus hydro-carbons are eliminated from the exhaust. But this only occurs at steady throttle and engine load conditions, the mixture is considerably richened when you plant your foot to the boards. I believe that the response time for typical modern Lambda sensors is in the region of half a second, and this period may very well correspond with the frequency of our kangaroo - what I'm getting at here is that maybe the lambda signal starts to oscillate for some unknown reason which then sets up a positive feedback (this means that when a problem starts, the feedback just adds to the problem i.e the kangaroo gets progressively worse very quickly) which then causes the kangaroo. Mine certainly felt like positive feedback behaviour.
Let's try a slightly different approach, Griffmates.
My Kangaroo had the following characteristics;
1.-It only occurs on constant, low throttle setting.
2.-It starts with a very small kangaroo, then gets progressively worse over a very short time (2-3 seconds) to a point where it is diabolical.
3.-The Kangaroo oscillates at about half a second frequency (i.e. each power surge occurs about every half a second).
4.-It occurs at any engine speed below 2000rpm.
Please bear with me on this.
The engine power is generated by mechanical parts and fuel/air, and the power is governed by means of a number of control inputs, one of them being your right foot, but many more of them being electronic.
It seems reasonable to suppose that, if our engine power oscillates when our foot is rock steady, then one of the other inputs is oscillating. The big question is - which one(s) (bear with me Quinney, I’ll get to your cam change).
It also seems reasonable to suppose that the frequency at which the errant control oscillates is similar to our Kangaroo oscillation, otherwise, you’d expect it to oscillate at a different frequency. It also seems reasonable to suppose that the errant signal is in some closed loop control system which is the only thing that can explain the behaviour of quickly getting progressively worse (no. 2 above).
If the logic applied here is sound, then we are looking for a control input that oscillates at the frequency of our Kangaroos, and resides in a closed loop control system. If you consider this information along with the excellent point made by TRACKCAR which is that the Lambda feedback attempts to adjust the fuelling to achieve “1” when in constant throttle position (bearing in mind that this is when it works in closed loop control) whatever map you’ve got, then we have a strong suspect. The question is, what can cause it to oscillate? We need to consider the whole system and not just the sensor. For example, it could be that electrical noise gets on the sensor supply wire which then causes the sensor to wobble which in turn causes the fuel adjustment to wobble which then causes the exhaust gases to wobble which then makes the sensor wobble even more because it’s trying to correct for the exhaust gas wobble (this is known as “positive feedback” in a control system – you can think of it as like hitting a kind of resonance with no damping to remove the energy, thus the damn thing just gets worse until it shakes itself to bits).
Quinny, it’s just possible that in changing your cam, your wiring has been disturbed (the lambda wiring does go over the top of the engine I think), and this may have had the effect I described above.
I’m not arrogant enough to assert that what I’ve described IS the cause of all our Kangaroos, but there may be an inkling of an idea that you guys can develop?
My Kangaroo had the following characteristics;
1.-It only occurs on constant, low throttle setting.
2.-It starts with a very small kangaroo, then gets progressively worse over a very short time (2-3 seconds) to a point where it is diabolical.
3.-The Kangaroo oscillates at about half a second frequency (i.e. each power surge occurs about every half a second).
4.-It occurs at any engine speed below 2000rpm.
Please bear with me on this.
The engine power is generated by mechanical parts and fuel/air, and the power is governed by means of a number of control inputs, one of them being your right foot, but many more of them being electronic.
It seems reasonable to suppose that, if our engine power oscillates when our foot is rock steady, then one of the other inputs is oscillating. The big question is - which one(s) (bear with me Quinney, I’ll get to your cam change).
It also seems reasonable to suppose that the frequency at which the errant control oscillates is similar to our Kangaroo oscillation, otherwise, you’d expect it to oscillate at a different frequency. It also seems reasonable to suppose that the errant signal is in some closed loop control system which is the only thing that can explain the behaviour of quickly getting progressively worse (no. 2 above).
If the logic applied here is sound, then we are looking for a control input that oscillates at the frequency of our Kangaroos, and resides in a closed loop control system. If you consider this information along with the excellent point made by TRACKCAR which is that the Lambda feedback attempts to adjust the fuelling to achieve “1” when in constant throttle position (bearing in mind that this is when it works in closed loop control) whatever map you’ve got, then we have a strong suspect. The question is, what can cause it to oscillate? We need to consider the whole system and not just the sensor. For example, it could be that electrical noise gets on the sensor supply wire which then causes the sensor to wobble which in turn causes the fuel adjustment to wobble which then causes the exhaust gases to wobble which then makes the sensor wobble even more because it’s trying to correct for the exhaust gas wobble (this is known as “positive feedback” in a control system – you can think of it as like hitting a kind of resonance with no damping to remove the energy, thus the damn thing just gets worse until it shakes itself to bits).
Quinny, it’s just possible that in changing your cam, your wiring has been disturbed (the lambda wiring does go over the top of the engine I think), and this may have had the effect I described above.
I’m not arrogant enough to assert that what I’ve described IS the cause of all our Kangaroos, but there may be an inkling of an idea that you guys can develop?
Griffter, I think you may have mis-understood what I was saying, I was not referring to any oscillation due to the drive train, I was referring to an oscillation in a closed loop control system.
Wixer, what does I.I.R.C. mean?
It sounds as though we agree that it's most likely to be something to do with the closed loop system, and the logic you apply to your experiment where the pre-cat chip resolves the problem seems to support this theory. I wonder if it's possible to run your car with the cat chip, but somehow "fix" the feedback signal to the ECU with some electrical wizardry to make it constant? If the problem is still apparent, then you have ruled out the closed loop feedback and the problem must be down to the different fuelling. (perhaps it's what Green V8S says, although I don't understand how exhaust gases can get back into the Plenum?).If however the problem goes away, then you have eliminated the fuelling, and it must be something in the feedback loop that's causing the problem.
Wixer, what does I.I.R.C. mean?
It sounds as though we agree that it's most likely to be something to do with the closed loop system, and the logic you apply to your experiment where the pre-cat chip resolves the problem seems to support this theory. I wonder if it's possible to run your car with the cat chip, but somehow "fix" the feedback signal to the ECU with some electrical wizardry to make it constant? If the problem is still apparent, then you have ruled out the closed loop feedback and the problem must be down to the different fuelling. (perhaps it's what Green V8S says, although I don't understand how exhaust gases can get back into the Plenum?).If however the problem goes away, then you have eliminated the fuelling, and it must be something in the feedback loop that's causing the problem.
wixer said:
Secondly, Raw-sewedge is looking into the MegaSquirt injection system. This is a DIY system that you can tune from a laptop and isn't very expensive. We are planning to build an engine rig and test it over the winter. If it's successful, then we may fit it to our cars.
Crikey, you blokes are live-wires. An engine rig eh! Sounds like a very interesting experiment - I'm sure that there are a few people who would be interested in the results! Good luck!
I've just dug some info out that I got ages ago - it may help:
ECU connector pin 23 - OS (left bank) Oxygen sensor signal specification;(n.b. pin 24 is for right bank)
ignition key on (multiplug disconnected)= 0.4 to 0.5V
engine running, hot = 200 to 1000mV (switching)
throttle fully open = 0.5 to 1.0V
deceleration (fuel cut-off) = 0 to 0.5V
switching frequency = 1 sec intervals (approx)
I find the "switching" intriguing - some bright spark said earlier in the thread that the signal "should oscillate when it's working properly". I assume that the switching is controlled by the ECU and not the sensor? The fact that it's switching at approx 1 second intervals when the engine is hot means that the ECU has to deal with a switching feedback signal when it goes into closed loop mode.
Maybe some other parameter in the ECU is on it's limit of specification which then makes it difficult to cope with the switching signal - if this was true, then changing over to another ECU is likely to solve the problem - anyone done that yet?
Or maybe there's some electrical interference causing a spike on the lambda signal which is screwing up the switching timing in the ECU - it is easily possible to get 200mV, or even 1000mV coupled into a wire, especially when you've got massive voltages from the ignition in the immediate vicinity - I don't suppose that the ECU needs very much current so this is a feasible possible cause. Has anyone tried re-routing signal wires away from HT leads?
It's frustrating that we haven't yet got a solution or solutions to this problem - I guess I'm lucky that mine's only done it on one occasion, but I suppose it could happen again. It sounds like some of you boys have to live with this all the time?
ECU connector pin 23 - OS (left bank) Oxygen sensor signal specification;(n.b. pin 24 is for right bank)
ignition key on (multiplug disconnected)= 0.4 to 0.5V
engine running, hot = 200 to 1000mV (switching)
throttle fully open = 0.5 to 1.0V
deceleration (fuel cut-off) = 0 to 0.5V
switching frequency = 1 sec intervals (approx)
I find the "switching" intriguing - some bright spark said earlier in the thread that the signal "should oscillate when it's working properly". I assume that the switching is controlled by the ECU and not the sensor? The fact that it's switching at approx 1 second intervals when the engine is hot means that the ECU has to deal with a switching feedback signal when it goes into closed loop mode.
Maybe some other parameter in the ECU is on it's limit of specification which then makes it difficult to cope with the switching signal - if this was true, then changing over to another ECU is likely to solve the problem - anyone done that yet?
Or maybe there's some electrical interference causing a spike on the lambda signal which is screwing up the switching timing in the ECU - it is easily possible to get 200mV, or even 1000mV coupled into a wire, especially when you've got massive voltages from the ignition in the immediate vicinity - I don't suppose that the ECU needs very much current so this is a feasible possible cause. Has anyone tried re-routing signal wires away from HT leads?
It's frustrating that we haven't yet got a solution or solutions to this problem - I guess I'm lucky that mine's only done it on one occasion, but I suppose it could happen again. It sounds like some of you boys have to live with this all the time?
After not using the car for the last month, the little red devil gave me some serious gip over the weekend. Saturday, after six miles of perfect running it cut-out completely - not a sausage. After switching the ignition off, it started up immediately and then repeated said cutting-out after 30 seconds. It did this about 8 times, then it ran ok on constant light throttle for a few miles. Later on (different journey) it started Severely misfiring, sometimes on constant light throttle and sometimes when accelerating - oh and of course serious kangaroo-ing. Also, when it wasn't mis-firing, the power was not its usual magnificent levels. Other symptoms are; smell of unburnt fuel and extreme fuel consumption.
I found and tightened a loose plug and reset the ECU (I thought the plug may have caused combustion problems, causing the lambda feedback to screw up, causing the ECU to go into Limp Home Mode). It didn't fix it. I also spotted (in a darkened garage) some infrequent "lightening" from one plug lead to another, and also the exhaust on the same side as the plug leads was glowing (very, very dull) red around the position of the pre-cats.
Whatdya reckon?
I found and tightened a loose plug and reset the ECU (I thought the plug may have caused combustion problems, causing the lambda feedback to screw up, causing the ECU to go into Limp Home Mode). It didn't fix it. I also spotted (in a darkened garage) some infrequent "lightening" from one plug lead to another, and also the exhaust on the same side as the plug leads was glowing (very, very dull) red around the position of the pre-cats.
Whatdya reckon?
Alan, I started the thread originally, and posted the more serious problems of engine cutting out above. The problem has completely gone away, and as I recall, the action that did the trick was a new set of HT leads.
So you do not have to live with it, the only question is how long it'll take you to find it and how much it'll cost. If your leads are getting old (over three years?) I'd replace them anyway (renew the metal plug caps while you're there). If your coil's getting on a bit, replace that too. Even if you don't do this, then definitely shove yer head undre the bonnet in the pitch black of night (no lights on anywhere near) and give yourself ten minutes of staring very hard at the HT areas to see what's happening. Feint "watery blue flashes" on the leads where they are touching something can be OK (but not alway), but if you see any "spark" type of flashes then this is definitely a problem that wants fixing, although it's not necessarily the whole problem.
After this, I guess you're down to making signal measurements at the ECU or checking the fault codes logged on the ECU.
Good luck!
Steve
So you do not have to live with it, the only question is how long it'll take you to find it and how much it'll cost. If your leads are getting old (over three years?) I'd replace them anyway (renew the metal plug caps while you're there). If your coil's getting on a bit, replace that too. Even if you don't do this, then definitely shove yer head undre the bonnet in the pitch black of night (no lights on anywhere near) and give yourself ten minutes of staring very hard at the HT areas to see what's happening. Feint "watery blue flashes" on the leads where they are touching something can be OK (but not alway), but if you see any "spark" type of flashes then this is definitely a problem that wants fixing, although it's not necessarily the whole problem.
After this, I guess you're down to making signal measurements at the ECU or checking the fault codes logged on the ECU.
Good luck!
Steve
Gassing Station | Griffith | Top of Page | What's New | My Stuff