Advice on sporadic AFR's

[pezzler]

Just wandering if anyone could make any suggestions on how to fix the fuelling problem; I'll start with the car and spec:

 

1993, Auto, TT. Full decat, FCD (HKS), AEM EBC, IK24 plugs, Walbro pump, Nur Spec exhaust, Whifbitz SMIC, AEM AFR gauge

 

The car had been running very smootly for the last 3 years; I finished BPU'ing about 2 months ago, and it had been running pefectly - apart from it was idling at 1k on a cold start

 

I took the car to THOR to have the AFR gauge fitted, I was going to do it myself but couldnt be arsed with cutting and welding (seemed like a lot of fannying around). When I collected the car, it said on the report that they had replaced the stock o2 with the Bosch uego that came with the gauge.

I was under the impression that the sensor was going to be mounted downstream of the o2, and purely used for monitoring, but I guessed that they knew more than me about these things so I didnt question it.

 

Since then, 1 out of 3 times, the fuelling runs really lean occassionally; this normally happens when the car is warm, in "D", and coasting with my foot off of the pedal. If I coast upto traffic lights, the gauge will move upward of 14.5 to 18, and then off the scale. When it runs off of the scale, the car idles really roughly at about 750 and threatens to cut out - a blip of the throttle and the it will drop back down to 15ish, but then it will rise if I dont maintain revs

 

I cant see why it only happens occassionally??

 

Anyone have any ideas?

 

PS. Before anyone suggests it, I cant afford to take it back to THOR at the moment!

[Tricky-Ricky]

Sounds like they haven't calibrated the narrow band simulated output to the ECU properly, sorry but to me they have just been lazy and fitted it in place of the std lambda, rather than drilling and welding the bung in for the wide-band, if you gave instructions to have the wide-band as a monitor and not work as the primary lambda feedback, i would be having word with them.

 

Although the readings you seem to be getting are pretty normal, i believe the lambda feedback is used slightly differently by the ECU with autos. (maybe somebody who maps auto's can confirm this or not?)

Edited December 2, 2010 by Tricky-Ricky (see edit history)

[Swampy442]

Running lean off throttle is normal, it should be reading --- for no fuel as nothing is injected on the over run. Watch my vid and see if this is what you mean...

 

 

[johnny g]

Have they used the 0-1v switch the stock "narrow" band reading rather than the 0-5v for the wideband operation?

[Guest Stillen]

Agreed. You will see a lean condition on off decel of course but double check the calibration.

[pezzler]

sorry but to me they have just been lazy and fitted it in place of the std lambda, rather than drilling and welding the bung in for the wide-band

 

Completely agree, I left the installation instructions to read at their leisure - I just assumed that they would have fitted it purely as a monitoring sensor. Just thought that that was the way they were meant to be fitted, and that they wouldnt need me to tell them how I wanted it.

 

Might drop them an e-mail tomorrow, see if we could sort something out

Edited December 6, 2010 by pezzler (see edit history)

[pezzler]

Running lean off throttle is normal, it should be reading --- for no fuel as nothing is injected on the over run

 

My gauge is pretty much reading exactly the same as yours in the video, on decel the mixture runs lean up until ---; the only issue I have is that it runs so lean that it has a really "lumpy" idle. The needle bobs around the 750 mark, and it feels like its going to cut out - but this is not always the case, only seems to be intermittent

 

I might try resetting the ECU as a starting point, failing that, i'll contct THOR

 

Thanks for all the advice chaps

[pezzler]

Have they used the 0-1v switch the stock "narrow" band reading rather than the 0-5v for the wideband operation?

 

Not sure, i'll take a look at the harness later to see how they have spiliced it into the original loom

 

I'm not saying anything at the minute, I know the guys at THOR know their onions

Edited December 6, 2010 by pezzler (see edit history)

[Swampy442]

I found with mine once it goes to really lean then off the scale the idle is really rough, if I moved the gauge and the plug it went back to normal. Im replacing the whole lot, dont want an AFR gauge I cant trust

[pezzler]

I thought the AEM AFR gauges were one of the better ones; I guess yours is fitted the same as mine, with the UEGO used as the feedback to the ECU as well the the gauge?

[Swampy442]

Yes it is currently and the car runs fine but Ive got an Fcon to install.

[pezzler]

Just dropped Charlie at THOR an e-mail; I asked if the work carried out is covered by any kind of warranty, they're usually quite good, so fingers crossed!

[pezzler]

Incase anyone is interested in future (and they're an idiot like me). I had a a pretty quick response (2 hours!) from Pete at THOR, he assured me that the sensor is fine in this location, and that the wideband sensor supplied with the AEM gauge will simulate the the signal of the stock narrowband.

 

He also assured me (as did Swampy) that the lean mixture on deceleration is normal

 

However, I do still have the problem of the car running very lean (occassionally) at idle when warm. If i'm stopped at lights for example, the engine will idle at 750 RPM, the AFR gauge will be off the scale, dash lights flicker and the the car will feel like its going to stall. Anyone want to have a pop at this?

 

Cheers

[Tricky-Ricky]

Incase anyone is interested in future (and they're an idiot like me). I had a a pretty quick response (2 hours!) from Pete at THOR, he assured me that the sensor is fine in this location, and that the wideband sensor supplied with the AEM gauge will simulate the the signal of the stock narrowband.

 

He also assured me (as did Swampy) that the lean mixture on deceleration is normal

 

However, I do still have the problem of the car running very lean (occassionally) at idle when warm. If i'm stopped at lights for example, the engine will idle at 750 RPM, the AFR gauge will be off the scale, dash lights flicker and the the car will feel like its going to stall. Anyone want to have a pop at this?

 

Cheers

 

 

 

I was always under the impression that you where wanting and expecting the wide-band to be installed purely as a secondary monitoring device, rather than the sole lambda feedback device, and that you where not happy (and rightly so IMO) that it had been installed as it had?

Personally i would always run off the std lambda, unless your running an ECU that can recognize and utilize a wade-band lambda.

 

Anyway the description you provided gave rise to doubts of how the device was calibrated, as although the AFRs described where correct for deceleration, they where not for idle, and the fact that you where describing a near stall condition that accompanied the very lean AFRs just reinforced this, was the car doing this before the wide-band was installed?

If so there is a definite connection, other than that maybe is a coincidental dirty idle control valve disc, try cleaning this, as i said before i think the auto-box is ECU controlled, and i guess it possible that something was disturbed, or possibly still something amiss as regards lambda feedback, as you should be seeing 14.7 AFR as soon as its idling, which as i said ECU controlled using lmabda feedback.

[pezzler]

I cleaned out the IACV a couple of months ago, and as you said, the car was running perfectly fine before the stock sensor was done away with. I'm going to be calling them tomorrow to see if I can get it booked in for Friday - really not happy with how its running

It seems to be running completely shite when the throttle is closed, which would indicate an issue with the Lambda

 

I'll update when its all sorted - thanks for all the comments

[Ian C]

Don't allow them to make any adjustments to the throttle stop screw just to shim the butterfly open a bit more I'm not saying Thor would, but it's an easy bodge to cover up an underlying problem. Forewarned is forearmed and all that.

[pezzler]

Don't allow them to make any adjustments to the throttle stop screw just to shim the butterfly open a bit more I'm not saying Thor would, but it's an easy bodge to cover up an underlying problem. Forewarned is forearmed and all that.

 

Actually conisidered doing this myself as a temporary measure; decided to leave as is, and not to drive it. It goes in on tuesday, so hopefully will be sorted.

 

Thanks for the heads-up though

[pezzler]

Just for future reference...

 

I have got back from THOR; spent npretty much the whole day there. They seem to have cured the issue that I had with the car running lean, by replacing the stock Lambda, and mounting the AFR gauge sensor downstream - about 18in further down

 

Whilst there the car was running sweetly, nice and smooth through the range, no coughs or splutters. However, on the way home I was getting a pretty consistant reading of 14.5-15.5 on the motorway for about 20 mins. After 20 mins had passed, the gauge started reading quite rich figures; 12.5(ish) whilst cruising, and as low as 10.5 with the throttle closed @ 1200 RPM, and sat at idle

 

I thought better of it so I pulled over and shut the engine off for a few minutes; this seemed to do the trick. When I fired the engine back up, the gauge read around the 15 mark on idle, I pulled away and everything seemed fine until I came off of the motorway, when it decided to display a rich mixture again - 12.5 cruising, 10.5 - 11.5 on idle

 

Anyone have any ideas?

 

On the plus side; I played some XBOX, and watched series 1 of "Extras" whilst they refitted the stock o2 sensor, welded in the WB sensor for the gauge, and gave the car a general checkover - I wasnt billed a penny

[Tricky-Ricky]

18" further down would equate to the decat pipe, is that correct? i am also presuming that the sensor is the Bosch LSU4? i have known

them to suffer from overheat issues, but that usually results in reading all over the place, not just rich,but that was with the Innovate unit.

If the gauge was reading correctly initially,then i would be starting to suspect that either the sensor or the gauge has a problem, unless its down to the std lambda, but i would presume that they would have verified that was OK anyway?

Other than that, i guess its worth checking the MAP sensor just to be sure and eliminate it, although i doubt it would manifest as this problem.

 

Get the gauge+lambda sensor checked if pos, did the AFRs go rich just after a period of boost?

[pezzler]

It has been mounted in the decat pipe

 

The sensor is a UEGO LSU 4.2

 

I was thinking the stock sensor might be at fault, but then the readings are fine for around 20 mins of driving. Think i'll check the voltages tomorrow just to be sure

 

I dont even think I made any boost on the way home, It was dark and foggy! It just seemed to happen all of a sudden, nothing I did differently. Strange one, its going back in on the 7th though, going to have them stick a probe in it and run it for a bit on their dyno

[Tricky-Ricky]

The LSU 4.2 sensor is Bosch, AEM don't make lambda sensors, if it was the std lambda sensor playing up it would be showing rich or weak all the time, unlikely to be so constant, i still favor the AEM unit as the culprit.

[pezzler]

Think you may be right

 

I'm hoping that its the UEGO though, would be cheaper to replace than the gauge unit. Definitely suspect though; seems odd that when the car was relying on the UEGO for ECU feedback, the car was running like sh*te on idle. As soon as the stock o2 sensor is refitted, the car runs sweet but I get intermittently odd readings on the AFR gauge. I'm 99% sure the car is running fine, and that there is either a signal issue or poor calculation within the gauge itself

 

Just as a future word of warning for whoever may be looking into an AEM gauge - make sure the UEGO sensor is fitted downstream of the stock sensor!

 

I believe Swampy has/has had similar idling issues with his car, and his AFR gauge sensor is fitted in the same position that mine was

[Swampy442]

I do still have the issue (well did when it was on the road) but when you disconnect the AEM unit its smooth as silk. Im replacing the Uego first, I dont run an O2 sensor.

[pezzler]

I dont run an O2 sensor.

 

Thought so

 

The 2 weeks that my car was running without an o2 sensor, and using the UEGO as ECU feedback, I was having intermittent issues with rough idling/almost stalling - when hot, it kinda wanted to die at 750 rpm

 

As soon as a stock o2 was refitted, and the UEGO was used purely as a monitoring sensor, the problem was disappeared.

 

Maybe the WB sensor doesnt like being mounted so far up the downpipe?? Or maybe it doesnt like being used as a feedback sensor for the ECU. Bit odd that we have had very similar issues after having the same set-up

[Swampy442]

There is an output for simulated narrow band so I dont why you cant use it.. But Im fitting an Fcon in the new year and getting it mapped so Ill see whats recommended by the mapper.

In my particular case though when it runs rough and the AFRs go off the scale lean, if I hold the loom and move the gauge up and down it runs fine so I reckon my gauge is shafted.