o2 sensor malfunctioning?

[CRCinAU]

fooey - I just re-read what you said.... yeah - 0.3s is too fast for the meter... So if I understand this right, the ecm picks its AF ratio, then feedback is given to the ecm via the o2 sensor which should read ~450-500mV for a correct tune? higher than than being too rich, lower being too lean?

The ecm then changes the mixture to suit (assuming via injector opening time?) to adjust the AF ratio and we go around in circles again.

Is this basically how it happens?

 

[Cheap6]

Pretty much. The mixture, and O2 sensor voltage, will oscillate around that 450mV value though. At idle and high load the target mixture will be something other than 14.7:1 (actually stoichiometric).

The plot of the sensor output looks OK in that it is oscillating as it should but perhaps a bit slow. You can count the # of oscillations - peaks or troughs - over an interval of time and divide by the time from the scale on the bottom to get the frequency or cycle time. What is the rate at which EFILive collects data; maybe that signal is aliased anyway? Canyou zoom in?

I think that the scale for the sensor output is not mV, despite what it says on the screen - perhaps AFR? If so, it looks to be lower - richer - at idle, spike lower when the throttle is opened and return to track the 14.7:1 target off idle @ constant ~2200rpm i.e. is pretty much correct.

You can use a digital multimeter to get some of the sensor output. Backprobe the O2 sensor wires at the PCM and record - write down - the readings over a minute or two (while driving, with help). The signal will be aliased but with sufficient readings you will be able to get an average which should be ~450mV (off idle - readings at idle will be somewhat higher).

If you have some cracked hoses maybe replace them. If the vacuum reference to the fuel pressure regulator is leaking, fuel pressure will be higher => rich. The effect of any vacuum leak will reduce with load.

Coolant temp. looks to be OK but a partial list of temp. vs resistance for the sensor is in the How To on switching a V6 engine cooling fan.

The long term fuel trim value should give you an idea of whether or not the PCM is correcting for some fault. It will vary through the particular cell (particular part of the load vs rpm map as divided by a grid) and should (will) rest to zero when you clear the PCM memory.

 

[delcowizzid]

change injector time and on efilive to o2 sensor volts by right clicking on the gauge and selecting o2 voltage should flutuate beween 300mv and 600mv or so when in closed loop. at at idle and high throttle will go up into the .800mv if it just hovers down in the 300 range all the time shes dead just not dead enough to set off error codes or you have a serious fuel supply issue

 

[CRCinAU]

Pretty much. The mixture, and O2 sensor voltage, will oscillate around that 450mV value though. At idle and high load the target mixture will be something other than 14.7:1 (actually stoichiometric).

Excellent! I have to apologise if some of these are dumb questions - I am more of a pilot than a mechanic - but you get taught lots of mechanical basics when you do the training. In the air its easy - we have an EGT gauge and a mixture control to take care of all this stuff!

The plot of the sensor output looks OK in that it is oscillating as it should but perhaps a bit slow. You can count the # of oscillations - peaks or troughs - over an interval of time and divide by the time from the scale on the bottom to get the frequency or cycle time. What is the rate at which EFILive collects data; maybe that signal is aliased anyway? Canyou zoom in?

Not that I know of - however 13.3mV would agree with what the dash says. It wouldn't surprise me if the o2 sensor is completely ratshit. I haven't replaced it since I got the car (then at ~128000km, and now at about 360000km)... They don't last forever

I think that the scale for the sensor output is not mV, despite what it says on the screen - perhaps AFR? If so, it looks to be lower - richer - at idle, spike lower when the throttle is opened and return to track the 14.7:1 target off idle @ constant ~2200rpm i.e. is pretty much correct.

You can use a digital multimeter to get some of the sensor output. Backprobe the O2 sensor wires at the PCM and record - write down - the readings over a minute or two (while driving, with help). The signal will be aliased but with sufficient readings you will be able to get an average which should be ~450mV (off idle - readings at idle will be somewhat higher).

Might have to play around with this - although nearly 240000km on the one o2 sensor probably means tis time for a new one anyhow

If you have some cracked hoses maybe replace them. If the vacuum reference to the fuel pressure regulator is leaking, fuel pressure will be higher => rich. The effect of any vacuum leak will reduce with load.

You know, I didn't even think of checking this... After looking, the vacuum hoses are almost present. I think I can see daylight THROUGH a few parts of them. Off to the local supply shop to pick up half a metre or so and replace any that I see I think this could be part of the main problem!

Coolant temp. looks to be OK but a partial list of temp. vs resistance for the sensor is in the How To on switching a V6 engine cooling fan.

I don't have any reasons to suspect this sensor - but I'll have a play as well

The long term fuel trim value should give you an idea of whether or not the PCM is correcting for some fault. It will vary through the particular cell (particular part of the load vs rpm map as divided by a grid) and should (will) rest to zero when you clear the PCM memory.

I assume that is the LTFC gauge? I guess a lower value of this is better...

 

[delcowizzid]

sensors have around 70000km life expectancy really before they start to wander lower in the volts you dont notice the fuel consumption going up slowly untill they get way rooted bet you save heaps on gas

 

[CRCinAU]

change injector time and on efilive to o2 sensor volts by right clicking on the gauge and selecting o2 voltage should flutuate beween 300mv and 600mv or so when in closed loop. at at idle and high throttle will go up into the .800mv if it just hovers down in the 300 range all the time shes dead just not dead enough to set off error codes or you have a serious fuel supply issue

I am assuming open loop is using preset values while it learns by using the o2 sensor what is the correct ratio, then the system switches to closed mode for the measured readings?

I figure with the vacuum hosts being almost non-existant, this would be bad and probably the main cause of a lot of issues.

sensors have around 70000km life expectancy really before they start to wander lower in the volts you dont notice the fuel consumption going up slowly untill they get way rooted bet you save heaps on gas

This is what also makes me believe that the 13.3mV being measured may well be correct.

Thankfully, for me, the car is on LPG - so I don't usually burn that much petrol... That being said, it runs nicely on LPG, but rather average on petrol - rough idles, the lot.

 

[CRCinAU]

ok - I replaced the vacuum hose on the fuel pressure regulator, the T piece and the throttle body angle. Idle seems to be much smoother as a start. I got the following from EFILive now regarding the o2 sensor. It's interesting as it seems to just fade away to nothing...

After a while I got the CEL come on with the error for lean exhaust - I figure this is because the o2 sensor was indicating 0mV for a while... This means its probably still running rich to some degree.

 

[Cheap6]

What happens to the O2 voltage (I'm willing to accept that it is voltage now - I'm not familiar enough with EFILive ) and CEL after you either drive the car for a while or, better, clear the PCM memory and drive?

If the PCM has learned to alter the "first guess" injector pulse widths (yep, LTFC, altered by the STFC, in turn altered using feedback from the O2 sensor output) and the reason for altering the values has been removed, it will continue to use the previously learned "first guess" values for a while until it learns new values. It will do that for each of the load "cells", so it can take a little while for correction to occur in all the cells, particularly in the load cells that aren't used/visited much (high load and rpm).

Running on LPG with the mixtures "off" and a standard, petrol, memcal will mean the PCM will learn the incorrect values also, as it tries to correct for the LPG mixture without any actual control over it.

At 240K km you've got your money's worth from that O2 sensor .

 

[CRCinAU]

What happens to the O2 voltage (I'm willing to accept that it is voltage now - I'm not familiar enough with EFILive ) and CEL after you either drive the car for a while or, better, clear the PCM memory and drive?

I did all this over the weekend - ended up doing ~90 mins drive each way on petrol for testing. The CEL comes on and then turns off a bit before eventually coming back on again. It happened to me today on the way to get more vacuum hose to replace the perished ones.

If the PCM has learned to alter the "first guess" injector pulse widths (yep, LTFC, altered by the STFC, in turn altered using feedback from the O2 sensor output) and the reason for altering the values has been removed, it will continue to use the previously learned "first guess" values for a while until it learns new values. It will do that for each of the load "cells", so it can take a little while for correction to occur in all the cells, particularly in the load cells that aren't used/visited much (high load and rpm).

yeah - I ran the macro in EFILive to reset the LTFC values - so hopefully it will start to learn from scratch. I'll do it again when I put the new o2 sensor in and see what its all about. At least after replacing the vacuum hoses to the fuel pressure regulator, the exhaust has gone from smelling like "OMG, the fuel pump has been diverted to the exhaust!" to "Hmmm - its running rich". Thats a start - and hopefully with a bit of correct feedback, the computer will start figuring out correct mixtures.

Running on LPG with the mixtures "off" and a standard, petrol, memcal will mean the PCM will learn the incorrect values also, as it tries to correct for the LPG mixture without any actual control over it.

The LPG system in this was factory fitted by holden - or at least a subcontractor - so I'd hope that they got this right! From memory the memcal is on a 27cXXX EPROM - and I just happen to have a reader/writer for these (and probably quite a few spare chips) from the old radio days of programming radios via (E)EPROM. With a bit of motivation, I can probably read it out and get a bin from the chip.

At 240K km you've got your money's worth from that O2 sensor .

Yeah! I'm also thinking of replacing the injectors - it hasn't been a primary concern for me for quite a while - as I mainly use LPG where the o2 sensor and injectors are ignored completely. Actually, I now wonder - as the car has a little LPG 'computer' that is behind the glovebox with LEDs that flash on and off - as well as some vacuum solenoids under the bonnet - so it may actually tune the LPG too?

The sensor I ordered was from Rare Spares / Auto One - $46.50 or something - but I have to wait a couple of days until they get it in.... Much better than the $122 from Repco or $108 for their 'universal' sensor - and massively better than the butt-surgery that Holden wanted at $225. They all have the square plug and are all 4 wire - so that seems like a good deal to me.

 

[CRCinAU]

I also just happened to come across this which is a quite handy reference:

AIR FUEL RATIO CHART
PETROL  L.P.G   LAMBDA  USE
11.0    11.6    .75     POWER
11.8    12.4    .80     POWER
12.5    13.2    .85     POWER
13.2    14.0    .90     POWER
14.0    14.7    .95     POWER
14.7    15.5    1.00    EMISSIONS
15.4    16.3    1.05    ECONOMY
16.2    17.1    1.10    ECONOMY
16.9    17.8    1.15    ECONOMY
17.6    18.6    1.20    ECONOMY
18.4    19.4    1.25    ECONOMY