'Actual consumption' values...

[bazzbazz]

When I come up against a car with anything to do with fuel/AFR/emissions/mileage/poor response/performance or just being temperamental & bitchy I always work on the first principle of Alfas -

MAF, MAF, MAF, Bacon & eggs & MAF, MAF, MAF, MAF, MAF & Chips & MAF.

May I suggest disconnecting the MAF and see if the tangible problems you can physically notice while driving disappear , as the system will be running on the default values. This will no doubt totally pork your fuel consumption values but it will hopefully give you an idea as to if it is the culprit if any of the operational issues disappear or change.

Worth a try maybe? 

[johnl]

Thanks Bazzbazz,
that's what I need, an experienced opinion / suspicion to give me a reason to try one semi random part throw in preference to another semi random part throw...

How's this for something not right;

Yesterday in town, I had to stop on the street (held up by a car turning left into a driveway), when I see a nearby traffic light that I think is about to turn red, floor the throttle in first gear, car leaps forward enthusiastically and the display instantly goes to apparent 'maximum' (25L/100kms). I make the light (just), jab brakes and calmly turn left. From there, very gentle throttle / low rpm in first gear, second gear, third gear, (fourth gear? can't remember) and the display stays very high (over 20 L/100kms) for the next say 150 metres or so until I halt to turn right. When the car stops the display instantly drops to 2L / 100kms for maybe five seconds until traffic allows me to gently move off, at which point the display instantly jumps to about 18L / 100kms, despite only a whiff of throttle (hardly touched it, honest!). From there for maybe another 70 metres or so (till I park the car) at walking pace / light throttle the display remains no lower than about 16L /100kms (though gradually falling). I'm sure this is abberant...

Note that this was all no faster than the 60kmh speed limit. All the numbers are what I saw on the display, keeping in mind that I had to look at the road too...

A randomly selected comparison; Later that day I was driving a work vehicle (a 2.4L Camry). The 'acutual consumption' display on this car (or whatever Toyota may call it) behaves very differently. If you floor the pedal the display momentarily goes very very high (highest I've seen flash up is 70L / 100kms!!), and when you back off it only takes maybe half a second or so for the display number to plummet, how much depending on how far you've backed off. It can very quickly go from a momentarily frighteningly high number to a single digit number (back off completely and it almost instantly drops to 0L /100kms, but the 147 display also does nearly the same thing in this condition, though only down to a displayed 2L / 100kms). I assume that the 147 display really should behave more or less similarly to that in the Camry, but it it massively different. I vaguely recall driving a BA Falcon that displayed it's instantaneous consumption quite similarly to the Camry, i.e. could see hugely high numbers displayed with a heavy throttle, that then dropped rapidly and very substatially on a reduced throttle.

That this Camry can display such a monstrously high instantaneous number on it's consumption display also gives some weight to the suspicion that the highest number ever shown by the 147s 'actual cunsumption' display (never higher than 25L / 100kms) is indeed very likely to be an arbitrary number, and that the real momentary consumption (as opposed to a seemingly fictional maximum "actual consumption") can most probably be far greater than the display is capable of showing...

Regards,
John.

[johnl]

As to spurious consumption data, why or how could just that aspect be (digitally) wrong in isolation?

A question is; are the displayed numbers spurious?

If yes then this implies a problem with the calculation of the fuel useage by the ECU.

If no, then they are indeed reflecting actual real consumption (more or less), implying a problem with moment to moment fuelling, and therefore a likely problem with some data being recieved by the ECU. Just what that might be...

Regards,
John.

[poohbah]

No range calculator on my V6 156 - I still reset the trip computer every refill just out of habit even though I know I can get at least 620km per tank, driving it in peak hour traffic 4 days a week.

[johnl]

So, ignoring "actual consumption" for the moment, the "average consumption" (as calculated / displayed by the computer) is getting worse than it typically has been for quite some time. The "average consumption" has been hovering around 8.2L/100km (with 'spririted' but not insane driving) for quite a while, but recently has suddenly risen to a quite consistent 8.8L/100km (and I don't think that I've significantly, if at all, changed how hard I've been exercising the engine).

Is this reflecting the suspected 'laggy' AFR richness of the "actual consumption" (on a light throttle after having been at a wider throttle opening)? I don't know, but it's a bit suspicious. For the last few weeks I've strongly suspected that the AFR is remaining quite rich on a trailing throttle (after having been at a wider throttle opening), but then quite slowly settling down to a more reasnable AFR on a steady light throttle. If so, and assuming there actually is a transient over richness occuring, this might be why I'm not seeing evidence of a rich AFR on the plugs, i.e. any carbon deposited on the plugs (and Lambda probes) during transient richness may be burnt off during prolonged light throttle operation?

Some more careful readout observation (of "actual consumption") is suggesting (at least to me...) that the AFR may also be taking some time to enrich adequately when the throttle is snapped open (from light throttle). When the throttle is mashed open (from a steady light throttle), the engine isn't very responsive, but keep the rpm up and throttle opening higher and it is quite responsive, maybe due to the seemingly possible 'lagginess' in the AFR response to changed throttle opening(?). This is suggested by the "actual consumption" display being slow to respond with a higher reading when the throttle is opened quickly from light opening / low rpm.

Today I pulled the O2 sensors out to check them as best I could. According to the 'Lucas Lambda Catalogue" page, a rough and ready test of O2 sensor function can be done by heating the sensor probe and measuring the voltage output as follows:

".... now connect the voltmeter red probe to the signal wire (generally black) and the black probe to the earth (generally grey) and set voltmeter to .000 volts. Then with a propane torch heat up the tip of the sensor until it is cherry red. The sensor should give off a reading of nearly 1 volt in a few seconds. If it takes longer than 20 seconds to get a reasonable reading change the sensor."

So, following the directions I find that both sensors seem slow to produce respectable voltage, but perhaps more tellingly one is substantially slower than the other is. I think my next move is to change both upstream O2 sensors, and see what happens.

Anyone have any bad experience using 'universal' zirconia lambda senors on Alfas? I've used them before on other cars with no problems, and my understanding is that a 4 wire zirconia sensor should be a 4 wire zirconia sensor should be a 4 wire zirconia sensor. But I have heard random stories of universals not working well with some cars. The Lucas page also says that universals probably won't work properly for most cars after about MY2000, but then further into the text it appears that this applies to titania sensors, not zirconia sensors (as the 147 sensors are).

Regards,
John.

[Citroënbender]

Why not invest in a laptop and MES at this point?

Parts roulette is not going to return overly useful real time info and you can only go by passive inspections - vs live data.

[johnl]

CB,
Good advice.

I do have a code reader, but it won't "establish a connection" with the 147 (or the Saab for that matter). I don't own a laptop, so would have to aqcuire one (my wife has one, but she's possessively paranoid about it, that it could get stuffed up if she lets anyone use it unsupervised, so to ask would be to burn marital brownie points...). I might see if I can find a cheap obsolete used one.

I do agree that parts roulette is generally a poor strategy, but the O2 sensors are not unlikely to be the original ones (at least there is no record of them ever having been replaced), so likely due for changing anyway. That each one behaves quite differently to the other (re voltage output) when heated does suggest that at least one is not up to scratch, so I don't think replacement is an irrational move...

Regards,
John.

[Citroënbender]

A laptop running XP Pro is sufficient. It will need a memory upgrade if still "as sold". It will also benefit from installing Firefox as an aid to downloading various bits of supporting software. This should come in under $200 if you look around.

Cheap code readers are generally not worth the distraction. With good cables and a franchised version of diagnostic software you will be a long way ahead.

A oxy sensor that doesn't look sooty or have over 160K should be doing its job just fine.

[johnl]

A oxy sensor that doesn't look sooty or have over 160K should be doing its job just fine.

If I make the (I think) reasonable assumption that the sensors are probably the original fitment, then they are now 210,000km old. So, it would not be suprising if they were to be somewhat knackered (maybe more surprising if they weren't?), or if not then are not unlikely to be less than healthy soonish ...

Considering this, and that average fuel consumption appears to be on the rise, with weird transient "actual consumption" readouts, as well as there being erraticish driveabilty issues, I doubt that a pair of new sensors would be a complete waste of money. That they are not sooty in appearance does go against this rationalisation, but my suspicion is that the AFR may be reasonably OK much of the time (longer periods of steady state light throttle), so any tell tale sootiness might possibly be burned off under these conditions...? 

Regards,
John.

[Citroënbender]

It'd be a cheaper test of your premise to fit a pair of used but tested sensors, any gross deviation would be readily apparent.

[bazzbazz]

So, we didn't bother with the MAF then? (cleaning for example)

90% of the time when diagnostics throw up a O2 sensor error it is actually caused by the MAF (which only ever shows up as faulty when it's totally dead), feeding misleading data into the ECU which leads to erroneous operation and indication from the O2 sensors.

Not saying that the O2 sensors don't need replacing but you ALWAYS suspect the MAF first in these circumstances.

[johnl]

Bazz,
Yes, we did clean the MAF. I took it out 'properly', i.e removing it from the tube case to get a good look and good access for cleaning. It looked clean as a whistle, but I sprayed it thouroughly anyway. No change.

I then tried an ECU 'reset' (disconnect battery for 45 minutes, reconnect, turn to 'MAR' for 90 seconds, etc. etc.). Whether this is a true ECU reboot I'm not sure, but it seems to have worked for some people (if the internet can be believed...).

Anyway, this did have an effect, a bad one. The engine seemed more 'doughy' than it had before, but at least with some k's over a few days it has improved back to how it was.

I haven't discounted the MAF, it might still be an issue, but one thing at a time...

CB,
I agree, it would be cheaper just to throw some O2 sensors at it, and see. Business is down just now, so cash is an issue. I can't see myself paying whatever it might be for new direct replacement sensors, so it looks like some universals, or, as you suggest see if I can pick up a couple of matched low km sensors from a wrecker (needn't be from an Alfa, any 4 wire zirconia sensors should work, at least that is my understanding...).

Regards,
John.

[bazzbazz]

I then tried an ECU 'reset' (disconnect battery for 45 minutes, reconnect, turn to 'MAR' for 90 seconds, etc. etc.). Whether this is a true ECU reboot I'm not sure, but it seems to have worked for some people (if the internet can be believed...).

On your 147 that procedure will only do a Throttle Reset, it will have NO EFFECT on the Adaptive Parameters of the engine ECU. You need one of the Alfa Romeo diagnostic suites to do that.

The ONLY car that that procedure will work on is the Alfa 156 Selespeed with Bosch 3.1 ECU.

Also if you change any of the engine Management Sensors, such as O2 sensors, one needs to do a Adaptive Parameter Reset with the Diagnostics. Few people realize this.

As per MES -

If some elements such as the minimum actuator, the Lambda probe or the knock sensor are replaced, the function lets the ECU return the adjustment parameters to initial values so as to correctly manage the new component.

Bazz,
Yes, we did clean the MAF. I took it out 'properly', i.e removing it from the tube case to get a good look and good access for cleaning. It looked clean as a whistle, but I sprayed it thouroughly anyway. No change.

As you no doubt are well aware of, a clean MAF can still be a faulty MAF.

As for O2 sensors, you can pick up brand new genuine Bosch units for a good price these days -

https://www.ebay.com.au/itm/Bosch-0258006389-Oxygen-Sensor/232754279789?fits=Make%3AAlfa+Romeo%7CModel%3A147&epid=244826859&hash=item36313c756d:g:fD4AAOSwoKFa6dmj

[Citroënbender]

Worth following this auction, a durable machine you can load up with a variety of diagnostics programs.

https://www.ebay.com.au/itm/292583278633

Then you can watch one forward sensor independently of the other; reduces guesswork.

Last I looked, I had three or four used 147 oxy sensors laying around.

[johnl]

On your 147 that procedure will only do a Throttle Reset, it will have NO EFFECT on the Adaptive Parameters of the engine ECU. You need one of the Alfa Romeo diagnostic suites to do that.

Understood, thanks Bazz.

I felt it had some affect (if only on the throttle behaviour...), because the engine seemed less responsive for some time afterward, and then improved (scientifically validated with the 'buttometer'...). Placebo? I wasn't expecting it to be better after the 'reset', but was hoping it might be, but it actually seemed worse, when the worst I thought likely was no change...

Also if you change any of the engine Management Sensors, such as O2 sensors, one needs to do a Adaptive Parameter Reset with the Diagnostics. Few people realize this.

As per MES -
If some elements such as the minimum actuator, the Lambda probe or the knock sensor are replaced, the function lets the ECU return the adjustment parameters to initial values so as to correctly manage the new component.

I didn't. Does failure to reset in the prescribed manner prevent the ECU from adapting to new sensor parameters, or merely slow this down?

Whatever, it appears that electronic access will become a necessity some time quite soon...

As you no doubt are well aware of, a clean MAF can still be a faulty MAF.

Of course. I wasn't assuming that no change (as a result of cleaning) meant it must be good...

As for O2 sensors, you can pick up brand new genuine Bosch units for a good price these days -

That's a good price for a 'correct' sensor. Doesn't mean I can afford two of them just now with other large outgoings looming, unless I put it on the dreaded plastic...

I still wonder if it is actually necessary to use 'correct' O2 sensors, or whether generic sensors of the appropriate type (i.e 4 wire zirconia) will work acceptably, or any differently at all?

Regards,
John.