|
If I disconnect the LAmbda sensor and drive the car is logs the fault and lights up if I then clear the fault and monitor the Lambda sensor it behaves exactly the same slowly cycles at tickover falls to .1 of a volt with throttle cracked open 15 to 1700 rpm and the engine purrs no missing, just fine. Regards Peter
|
|
Is there any possibility that a fuelling problem is causing the flat line out of the Lambda or is the end of the road for the sensor. Regards Peter
|
|
could well be, (VW's do that with dissconnected duff MAF's) you could try and see if the lambda probe heater is taking any current, if it has a heater that is.......
|
|
|
|
If the lambda output cycles when the probe is disconnected, something else is causing the variation - i.e., it isn't an artefact of closed loop operation. The ECU, with a disconnected lambda probe, reverts to open loop operation, albeit with a fault code flagged up.
Under open loop operation, I would expect the probe to read very weak with no cycling when the engine is running steadily at above 2000 rpm.
The cycling at tickover is not what I would expect to see with a disconnected lambda sensor. It suggests, to me, a fault elsewhere.
After you had all of the fuel pump trouble, is the fuel pressure at the injectors now stable? If that is OK, my top suspect would be the mass air flow sensor.
Number_Cruncher
|
|
I headed there yesterday and Volvo identified the test procedure for this electronic version air mass meter. The voltages against speady rpm where correct and the static no flow value was within 0.01v I am goint to get access to a gas analyser tomorrow to see if it really is running very week although she pulls well when driven which does not imply a weak mixture. I have revisited all the pump and injector pressure conditions and they are within spec other than the blank pressure is quotes as 88 to 92 and I have a touch over 100. I do not what else to go for, I suppose it is possible that the O2 sensor failed when the car was driven 20 miles with a failed in tank pump thus fuel starvation ran weak on the cruise and burnt the last of the coating off the sensor in the Cat. Thanks for your help everyone. Regards Peter
|
More interesting tests. Engine cold Lambda heater resistance 12.2 ohms, all scoped up to the Lambda probe. Turn on open loop mode voltage reads 0.6V . 2 minutes later it starts to cycle slowly and erratically 0.15 to 0.9V Touch throttle and instantly drops to 0.15V and 10 seconds later starts to cycle again. Rev the engine and the level drops and on the overrun the level rises initially then falls to 0.15 and starts to cycle 15 seconds later. Hold the engine at 2500 rpm for 30 seconds solid0.12 volts back to idle 10 seconds later starts to cycle slowly again. When it is in cycle mode it averages 0.55V. So propane test gas on and a few second in and the voltage rises to 0.9v and engine rpm lists slightly gas off level slowly falls and 20 seconds later the sensor cycles again.
Engine running smoothly but at 1100 rpm instead of normal 900 ish. Engine doesn?t miss but clearly no idle control. I?ve checked all over for air leaks, vacuum pipes, leads feed and everything else. I am in the fortunate position of being able to swap to a spare ECU but exactly the same. Any ideas where to look next would be most welcome as I am bang my head on the garage wall here. Regards Peter
|
|
When writting the previuos post and said 1100rpm I questioned why that was so I've just been out and restarted the car and then clamped the air hose after the Idle Contol Valve and sure enough the engine tries to maintain the rpm until it suddenly stalls and you can here the ICV close ( it's the rotarary type driven by the ECU with a markspace ration pulse which with more load spends more time on then off so effectively opening the valve more. I've also clamped the feed pipe for the ICV with similiar effect. The only time 1050 rpm is used is from cold when the temp sensor tells the ecu to increase the inj pulse width and raise the tick over slightly. I've retested the dual temp values and they are fine. What is going on here. Any help welcom. Regards Peter
|
|
Hi, Yes I checked the TPS's signal at the ECU I have one with the lid off and they are fine I rechecked them 10 minutes ago and the temp sensor signals and they are fine. It started to rain so quickly put the gear away ( mains powered scope ) so that's it for a while. Is this all to do with the O2 sensor indicating it is lean thus the ecu turning up the air to obtain Lambda thus overriding the tickover value. Regards Peter
|
I'm not sure the lambda sensor is entirely to blame for your problems - as it indicates high voltage when the propane is added to the inlet giving a rich mixture.
On the lambda low is lean.
I don't understand why the voltage goes low (lean) when you blip the throttle, I'd expect the opposite to happen.
FYI this site:
www.picotech.com/auto/waveforms.html
has some interesting waveforms, with sensible comments about them.
This place:
www.lambdasensor.com
seems competitve for price.
Regards
John H
|
|
Thanks for your reply. I have been to that site already and very useful. As you have gathered my sensor does not perform like that. I have just come back from the test station where my exhaust Lambda is 1.1 but the Co at O.1 and 13ppm HC The Lamda output at 2500 rpm is fixed 0.9 Volts and only returns when returned to tick over and 10 to 15 seconds the output starts to slowly cycle. Does this finally point to a duff Lambda Sensor. Regards Peter
|
Hi Peter,
I would expect that the lambda sensor should continue to cycle while you hold the engine at 2500 or similar rpm, only faster cycling than at tickover because the transport lag between fuel being injected and the resulting exhaust gas being measured takes less time.
The fact that you aren't getting cycling at higher revs indicates that the management unit is switching back out of closed loop control, into open loop again.
As the lambda sensor is highly non-linear about lambda=1, the fact that you measure 0.9v does not necessarily mean that the mixture is very rich - it could be anywhere from ever so slightly rich to very rich!
If the cycling frequency at tickover is much slower than 1Hz, then I would be suspicious of the lambda sensor. You won't ever see a nice regular trace from the lambda sensor, but you should see cycling, or switching, that increases frequency with engine speed - although the system may resort to open loop during transients and full load operation.
Number_Cruncher
|
Thanks for your reply. I have been to that site already and very useful. As you have gathered my sensor does not perform like that. I have just come back from the test station where my exhaust Lambda is 1.1 but the Co at O.1 and 13ppm HC The Lamda output at 2500 rpm is fixed 0.9 Volts and only returns when returned to tick over and 10 to 15 seconds the output starts to slowly cycle. Does this finally point to a duff Lambda Sensor. Regards Peter
Perhaps it is time for a new sensor.
I understand that the lambda sensor can get sluggish and slow to switch in old age, and according to the people that sell them their life expectancy is less than yours.
But I'm still puzzled why the gas check at 2500 rpm was OK, as the "fixed 0.9 Volts" would indicate a rich mixture.
Oddly enough I'm in the throws of changing one myself at the moment, but it's likely to be the coming weekend before I get sufficient time to do the job "properly".
I have just got a "generic" sensor (43UKP delivered) which has shorter leads than I'd hoped for (and no connectors) and need to get the wiring joins water proof.
|
Yes I did that today but the Cat is too good and the output stayed at 0.1%. But the sensor has deteriorate to the point of only swinging between 0.3 and 0.7v and when it gets fully up to temperature it's output falls to zero. I have ordered a new one which should arrive tomorrrow. Hopefully that will be the end of the saga. I also found an intermittant in the NCTII ( Temp Sender ) signal and that was causing the enrichment that the Idle Control Valve was opening further to burn the fuel off and keep the Co down thus the 1050 to 1100 rpm. I only spotted it because I was monitoring the mark space ratio of the IDC 55-45 (0v) when the note change as did the ration to 50-50 thus more air I then went under the bonnet to starve the valve of air and as I disburbed e harness the revs dropped back. I probed the ECU( Lid Off) NCTII and it was noisey in fits, I wriggled the connector and there it was the socket click back into place firmly enough replaced the connector and it's fine. I'll report back when I've replaced the sensor and run the car for a while. Regards Peter
|
Ah, the intermittent on the temperature sensor makes a lot of sense.
If the connection is lost, the ECU doesn't know what is happening (why it didn't put the light on is interesting - is your system LH jetronic, or is it a Motronic variant? If it's LH, that makes sense because it only has very limited on board diagnostic when compared with Motronic). However, with no temp sensor signal, the ECU uses substitute values, and will swich out of closed loop control.
The idle speed will be raised in this mode, but, as the idle speed control valve by-passes the throttle but not the mass air flow sensor, the mixture isn't affected by this; i.e., the mass air flow sensor detects the extra air, and the ECU ups the fuelling accordingly.
It may be interesting to double check your lambda sensor oscilloscope readings with and without the temp sensor connected - you may find a change in the results!
How many latent faults does your Volve have??!!
Number_Cruncher
|
|
Yes I checked in the manual as well and it implies an error code and ECU fault light. I did see it come a couple of times but either went off again or was reset by a restart. I thought if was trying to tell me the O2 sensor was not switching but that normally latches and has to be cleared. Thanks for your help, lets see what a new sensor does for it. Regards Peter
|
New Lambda sensor fitted but alas still not right so here are
the result of the test done this afternoon with the new O" sensor
These ere the results from the AMM Still Air 1.377 volts Running 2.240 to 70 3.5V on open throttle and rpm rising through 3 rpm Heater 3.2 ohms.
45 degrees C 2.5 to 2.8 ms injector pulse.
85 Degrees C 2.3 to 2.45 cycling O2 0.1 to 0.85v cycling around 0.55 V
Pull regulator vacuum pipe O2 lean slow cycle Inj time 2.3 to 2.4 mostly 2.3
Vac back on 2.4 to 2.5ms 0.55V O2
1500 rpm Injector pulse widens then falls back to 2.5 ish O2 a little lean
2500 rpm O2 Stuck at 0.1 v with steady 2.5 ms Inj pulse
Tick over place car in gear slip clutch and pulse width increase out to 3 ms.
Temp sensor at ECU AOK
Throttle switch AOK and clean
Earth continuity to Lambda AOK
IDC pulse 55% off 45 on% but idle rpm not solid as it used to be.
Injector rail 48psi dead ended 105 psi
flow rate 1.1 litres per 30 seconds directly from manifold and almost 1.1 L per 30 seconds if measured in the return pipe.
Lifter pump 2.5 psi dead ended
What else can I say.
Better but lack of idle control, lumpy ar lower rpm and detectable missing at cruise but goes fine above 2.5K onload.
Any help welcome as this one has got me beat.
Regards Peter
|
Peter,
Is the lambda now switching more quickly at tick over than before? (If so, then you have done some good!), and is the lambda still not cycling as it should when revved up?
Number_Cruncher
|
|
Yes indeed it is and it's rise and fell time is much better, however it does cycle at 1500 rpm but at 2500 rpm is it locked down at .01 ish volts and it stays there only cycling again if you let the rpm fall. From this it implies that the Lambda it telling a true a story and the real fault is still at large. At 2500 if I pour in propanne the sensor quickly responds and shows a 0.9 V high as the sensor is now starved of O2. Any ideas welcome as I'm right of of them. Regards Peter
|
New Lambda sensor fitted but alas still not right so here are the results of the test done this afternoon with the new O" sensor
These are the results from the AMM Still Air 1.377 volts Running 2.240 to 70 3.5V on open throttle and rpm rising through 3 rpm Heater 3.2 ohms.
45 degrees C 2.5 to 2.8 ms injector pulse.
85 Degrees C 2.3 to 2.45 cycling O2 0.1v to 0.85v cycling around 0.55 V
Pull regulator vacuum pipe O2 lean slow cycle Inj time 2.3 to 2.4 mostly 2.3
Vac back on 2.4 to 2.5ms 0.55V average O2
1500 rpm Injector pulse widens then falls back to 2.5 ish O2 a little lean
2500 rpm O2 Stuck at 0.1 v with steady 2.5 ms Inj pulse
As the avid readers have gathered this has been a knightmare and continues to allude myself, and I've been around engines and ECU a long time, and two diagnostic technicials and the Head Mech at the Volvo main dealers. Bear with me though guys as I still need any help I can get with this oersistance and ellusive fault. Regards and Appreciation Peter
|
Peter,
When you pull the vac pipe off the regulator, did you plug the end to prevent the engine sucking air?
The reason I ask is that if you plug the pipe, so the engine isn't sucking unmetered air, the fuel pressure regulator should increase the fuel pressure in the rail, and the mixture should go rich, rather than weak as you reported.
If, however, you didn't plug the pipe, as well as the regulator changing the fuel pressure, you also allow unmetered air in, so you have two competing effects, so it is difficult to see what should happen - in other words, you would be changing two things at once.
Number_Cruncher
|
|
I pulled the reg hose to allow air in to lean it off, I thought, correct me if I am wrong that the vacuum at tick over is not high enough to back the pressure off, this is designed to lean the fuel on the over run to reduce emmissions. Thanks for you help my friend this is a long struggle. I'm starting to think I have a partially blocked ejector now and not sure witch way to go. Regards Peter
|
Hi Peter,
The regulator is set up to keep a constant pressure drop (i.e. fuel line - manifold) across the injector, so whatever the operating condition of the engine, the amount of fuel injected is as closely as possible, proportional to the amount of time the injector is held open.
If it weren't like this, the ECU would either have to have an input from a manifold pressure sensor, or calculate the manifold absolute pressure in order to apply a compensation for the fuel being "sucked" out of the injectors under partial and closed throttle operation.
So, on tickover, while the manifold vacuum is high (Manifold absolute pressure is low), the fuel pressure is reduced.
At wide open throttle, there is no manifold vacuum (Manifold absolute pressure is high - atmospheric), the fuel pressure is increased pro-rata.
If you think that you only have one or two blocked injectors, a cylinder balance test (Morse type test) might enable you to find the culprit.
Number_Cruncher
|
I've just bethought myself that if you do try a cylinder balance test, the idle speed control circuit might fight you.
So, you could try the test with the throttle just open beyond the idle switch, or disconnect either throttle switch or idle control valve to remove it.
Or, as you have the kit, you could compare the contribution of each cylinder by the change in idle control mark/space ratio as you cut each cut each cylinder out.
Number_Cruncher
|
|
Now that's a good idea Number-Cruncher. Just got back form two hours with Bosch Diagnostics gey and he has been through all the measurement and test results and can not fault the system. I do not feel such an idiot but I still have a fault. He thinks that when I was serching for the original fault the vibration os the howling HP pump may have caused the 100,000 feul filter to let loose fine debris whis has partially clogged the injectors thus the super lean engine and Lambda 1.1. We finally clamp the return pipe closed by maybe 80% and the engine cleaned up and drove much better. This also implies I am not getting enough fuel through the injectors to meet the airflow for this correct mixture. It's looks as though an ultra sonic clean, flow test and refit is required. Personally I am not convinced. Regards Peter
|
The blocked injector suggestion is a possibility. IIRC, there is a very fine, non-removable, filter on the inlet of the injector which can get clogged. Hopefully, the cleaning operation will begin with backflushing the injector to clean this filter.
In "normal" service, the usual cause for blocked injectors is people fitting the fuel filter backwards, so the fuel filter collapses inot bits rather than being forced out to the sides of the can.
Number_Cruncher
|
|
Reverse fitting is not possible on the Volvo as the two Banjo coupling are different diameters. But matbe there is some mileage in the vibration theory and the fact I thumped if a few times to see if the noise would clear. Still a long shot though I think. I can not get then cleaned and flow tested until Monday so I guess I get the weekend off. Thanks Peter
|
|
I rigged up an old pressure transducer in the tailpipe and connected it to a scope and triggered it with No1 firing. The other channel I connected to the ignition colt LT and Ran the car at normal operating temp I get an even output from all 4 pots so pulling the injector connection you can cleally see the pot that is not contributing and how much lower the pressure it is, a bit smoothed I guess by the silencer but still valid, I also measured the drop in RPM indicated my the other trace. I reconnect the injector and test no 2 and the effect is very similair, the three and then 4 all very much the same. I conclude that the fault then is either all 4 injectors are partially clogged by the same amount or non are clogged and the fault still lies elsewhere. No more time today but any ideas would be welcome. Regards Peter
|
Well, my new lambda sensor seems to have sorted out the "surging" and intermittant warning light I was suffering from.
The only comments are:
1 I ended up using a long 7/8" AF ring spanner with a bottle jack underneath the far end to slacked it, as it was too tight for me.
Obviously the wires were cut to get the spanner on.
2 as it was a cheap generic sensor I soldered on the old cables - well three of the four - the output from the sensor was a stainless steel inner which I ended up crimping, as I couldn't solder it.
Meanwhile - your problems.
I'm not sure about the Volvo specific arrangement with how the injectors are driven, but on the small FIATs I have some familiarity with there can be problems with either a safety cut-out relay going high resistance contacts, or the wiring to a ballast resistor, both of which are in series with the (single point) injector.
It seems the injector can function, but not hold open long enough for full throttle (wider mark-space ratio).
I realise your set up is multipoint, but it might be an area to investigate.
|
|
Thanks for that. My Volvo does not have a ballast resistor associates to the injectors they are directly driven by the ECU. I have spoken to Volvo this morning and they think the probability theat all four injectors are partially blocked after the in tank pump was changed is a very long shop so I will check everything again before taking the injectors out and having them ultra-sonically cleaned. Thanks Peter
|
Went to the car this morning and removed the oil vapour seperator from the block and cleaned it out although it was not very dirty, cleaned all associted pipes and reassemble. I removed the limk pipe from the AMM to the air filter and turned the pipe so it was facing upwards. Ran the engine and slowly covered the aperture to expose and air leaks either side of the throttle body no leaks anywhere. Mind you I alreadt new that but I've got nothing else left to try. Put it all back to gether and scoped it up. Ran engine till warm and O2 cycling then help the rpm at 2500 and low and behold the O2 was cycling let the engine return to tick over O2 rich, as it should be, then it cycled again. Engine tick over fine and appeared to pick up OK and rev freely. Now ly car alarm requires the key fobs to be associated to the alarm within 15 secs of power up and last time I was messing with ECU's I had not done this so I removed power, which means the ECU also losses it's 12v feed and repowered it and associated the keys. However when I restarted the car the engine was a bag of nails, no tick over control and very empty on opening the throttle. I removed power again for 10 mins and tried again, better but not as it was prior to powering down for the alarm.
I have check the ICV waveform and the O2 and they are all there but it just does not like it. Got rained off so it will have to wait. If any one has any ideas or can point me to a web site that can explain how the LH2.4 Jetronic really works and how to break the servo loop so I can see what is really happening this please put in your penies worth. Thanks Peter
|
The ECU I'm familiar with is reset to factory default by prolonged power off.
I can't find anything sensible about the LH2.4 either, but you may find it will "re-learn" its optimum settings by drving the car.
|
I have finally got my Volvo fixed. It turns out that even a Bosch specialist had trouble finding the cause but eventual after having the Injectors cleaned 3 times but they still failed to meet the flow rate they where replaced, Although this solved the leaness issue the car sat at 1300 rpm, no faults and everything seemed to test out ok. However a tap on the AMM and the engine note chances but the ECU has still gone into default mode and no ICV signal, so a replacement AMM and instant return to smooth running.
Total bill £540 inc Ouuuch Regards Peter
|
|
|
|
|
