1UZ-FE AFR varying between banks - A problem?

[Ian C]

Hi,

I am currently fitting and early (1992) 1UZ-Fe into a kit car to replace a Rover V8 and have found that the AFR is different between the two banks when carrying out the initial fuel set up. The engine is out of a Soarer and has had a COP upgrade using Honda K20 coils and is controlled by a Link Storm ECU. I have an Innovate DLG-1, coupled to an LC-2, to provide AFR readings from both banks to the ECU.

The initial engine runs showed a variation of approx. 13% between the AFRs measured on the nearside & offside exhaust collectors (Nearside 15.0 and Offside 13.5). I swapped the sensors and wiring side to side to eliminate the potential of errors in the Innovate or Link ECU but the difference in AFR was unchanged. I then performed a flow test on the injectors (standard 240 cc/s injectors) and found the variation between the injectors to be less than 2%. The variation in the AFR, between the two sides, is most notable at idle and decreases to between 2% & 4% at 3,000 rpm.

So the question is: Does a variation of 15% between the nearside and offside exhausts at idle constitute a problem or is it within tolerance. Any suggestions or advice would be gratefully received.

Regards,
Ian

[stevieturbo]

Firstly...so you have 2 independent widebands ?

And you are viewing or logging the info from them on either Logworks, or via the serial data stream ?

I simply would not trust the analogue outputs on the Innovate, unless you have actually manually tested and verified it's output voltages right through from say 0.5v to 5v in steps, flatlining the output each time and verifying the same reading on whatever "gauge" or display you are using.

And were the injectors tested at low duties ? What injectors are they ? Is the engine a standard engine, intake, plumbing etc ?

15% does sound like a lot, but could be affected by many things, I guess at a stretch you could swap injectors bank to bank to see if it moves. But if there are any modifications to plumbing, intake etc....that might account for some difference. Exhausts....less so.
The fact it cleans up with a bit more airflow....means there shouldn't be any major issues.

And is this running proper sequential injection ? Where is the end of injection timed ?

And I assume the engine itself is in good health ?

[ChrisJC]

Is there an air leak on the intake side?

Chris.

[DaveEFI]

On an RV8, the difference between 15:1 and 13.5:1 would be very noticeable indeed at idle. You'd not get a reliable idle at a sensible speed at 15:1.

Is the Link setup batch injection? If so a dual band scope would be useful to see if the injector drive pulses are the same to both banks. Assuming you have it wired left and right banks, rather than grouped in firing order.

A big difference like that in fuelling would also show in the plug colour, IMHO.

[Ian Anderson]

Just to throw it in the mix!
Have you checked the spark plugs on both sides and are they properly gapped?

Ian

[Ian C]

Firstly...so you have 2 independent widebands ?

And you are viewing or logging the info from them on either Logworks, or via the serial data stream ?

I simply would not trust the analogue outputs on the Innovate, unless you have actually manually tested and verified it's output voltages right through from say 0.5v to 5v in steps, flatlining the output each time and verifying the same reading on whatever "gauge" or display you are using.

And were the injectors tested at low duties ? What injectors are they ? Is the engine a standard engine, intake, plumbing etc ?

15% does sound like a lot, but could be affected by many things, I guess at a stretch you could swap injectors bank to bank to see if it moves. But if there are any modifications to plumbing, intake etc....that might account for some difference. Exhausts....less so.
The fact it cleans up with a bit more airflow....means there shouldn't be any major issues.

And is this running proper sequential injection ? Where is the end of injection timed ?

And I assume the engine itself is in good health ?

Stevie (?),

Thanks for your reply & interest, in order raised in your post:

Yes, I have used the Innovate DLG-2 gauge with the additional LC-2 unit to run wide band sensors in each side pipe just downstream from the collector.

Both AFR readings are displayed on the gauge and the analogue voltages are fed to the Link ECU and calibrated using the Innovate details. I have downloaded and installed the Innovate Logworks software but find it easier to use the Link management and keep everything together.

The readings on the Innovate gauge and the levels shown in the Link ECU software are in agreement to within approximately 0.2 AFR over the range and both show the difference, between the nearside & offside, to be of the same magnitude.

I tested the injectors with pulse widths of 2ms; 5ms; 10ms & 15ms. The largest variation in measured volume was 2% when testing using a 10ms pulse (2,000 pulses of 10ms with frequency of 50Hz - should have given a volume of 80cc measured average was 79.2cc). The 2ms test (10,000 pulses of 2ms at frequency of 50Hz to give a volume of 80cc) gave an average of volume of 47.7cc but the variation across all injectors was only 1.2%. The injectors are the ones that came with the engine (bought from a salvage facility via a well known auction site) but were fitted with new seals & filters before the engine was fitted. The engine is a standard 1992 (pre VVT) but has had the dual distributor ignition replaced with COP ignition using Honda K20 coils. The intake is currently unfiltered and consists of an 80mm silcone hose, fitted with an IAT sensor, with a 90 degree bend to the throttle body and standard intake plenum.

After the flow test on the injectors I sorted and fitted them to even out the fuel delivery between the two sides of the engine. The engine run yesterday, after refitting the injectors to "even out" the fuel delivery between the two sides, showed that the fault had swapped sides - the nearside now shows an AFR of 12.6 and the offside shows 14.7 (at 1,440 rpm).

The injectors are set up in sequential mode and are timed for the injection pulse to end 360° BTDC

The engine was given a visual check before fitting and looked to be in very good condition, sump and cylinder heads looked "golden" rather than black. The valve clearances ( gap between cam shafts and shim in the bucket above the valve) were all checked and found to be in tolerance. The engine started easily, once I had got the timing sorted, and it idles smoothly at 750 rpm and sounds good - the only (apparent) issue is the lambda difference!

Given that the problem swapped sides when I refitted the injectors would seem to indicate that one (or more?) of the injectors may be sticking or slow. This would explain the fact that the difference decreases at higher rpm where the injector pulses would be longer?

Any further help / suggestions would be appreciated.

Regards,
Ian

[Ian C]

Is there an air leak on the intake side?

Chris.

Chris,

Not that I am aware of, no noises or other indication of a leak. if it wasn't for the lambda readings I would have said it was running sweetly!

Regards,
Ian

[Ian C]

On an RV8, the difference between 15:1 and 13.5:1 would be very noticeable indeed at idle. You'd not get a reliable idle at a sensible speed at 15:1.

Is the Link setup batch injection? If so a dual band scope would be useful to see if the injector drive pulses are the same to both banks. Assuming you have it wired left and right banks, rather than grouped in firing order.

A big difference like that in fuelling would also show in the plug colour, IMHO.

Dave,

The ECU is set up for sequential injection (see reply to Stevie above) and all the injectors are connected individually to the injector driver channels on the ECU.

The suggestion to check the plugs is a good one, I did look at them way back at the start of this and nothing was obviously different. However since the problem switched sides (see reply to Stevie above) it would now appear that one (or more) injectors is either sticking or sluggish and this might show on the plug.

Thanks for the response, take care,
Ian

[Ian C]

Just to throw it in the mix!
Have you checked the spark plugs on both sides and are they properly gapped?

Ian

Ian,

Thank you for your response, yes the plugs were checked and properly gapped but are the original plugs that came with the engine so may be due replacement.

Envious of your GT40, the engine I replaced was a 3.9 with Hotwire 14CUX EFi. Do you use the Rovergauge diagnostic software on you Lucas system? - if not I would strongly recommend it!

Thanks again, take care,
ian

[DaveEFI]

With such a vast error between banks, wondering if a ColourTune would show which cylinder is causing the problem - if the plugs ain't showing it? Are the pulse widths from all the drivers near enough the same at a constant idle speed?

[Ian C]

With such a vast error between banks, wondering if a ColourTune would show which cylinder is causing the problem - if the plugs ain't showing it? Are the pulse widths from all the drivers near enough the same at a constant idle speed?

Dave,

Colourtune might be an option to help identify the cause, thank you. I am assuming the pulse widths are the same, as the problem swapped sides, after repositioning the injectors, I have eliminated the engine management from the list of possible causes - for the moment anyway!

Regards,
Ian

[stevieturbo]

Certainly with the "problem" following the injectors, it would seem there is some sort of issue with maybe 1 of them.

It would be a pain in the hole...but you could try swapping injector pairs 2 at a time across the banks.

I'd probably take the easy route and either ignore it, or buy new injectors.

Or if you knew which cylinder was the issue, just trim it up or down in the software at low loads. Problem is you need to find that cylinder/injector.

[Ian Anderson]

Just to throw it in the mix!
Have you checked the spark plugs on both sides and are they properly gapped?

Ian

Ian,

Thank you for your response, yes the plugs were checked and properly gapped but are the original plugs that came with the engine so may be due replacement.

Envious of your GT40, the engine I replaced was a 3.9 with Hotwire 14CUX EFi. Do you use the Rovergauge diagnostic software on you Lucas system? - if not I would strongly recommend it!

Thanks again, take care,
ian

Thanks Ian, I was lucky to get on the GT40 build when things were a bit cheaper. No I do not use Rovergauge, if it runs and pulls I am happy, though I still hanker for the simplicity of a carb!

Ian

[DaveEFI]

Just to throw it in the mix!
Have you checked the spark plugs on both sides and are they properly gapped?

Ian

Ian,

Thank you for your response, yes the plugs were checked and properly gapped but are the original plugs that came with the engine so may be due replacement.

Envious of your GT40, the engine I replaced was a 3.9 with Hotwire 14CUX EFi. Do you use the Rovergauge diagnostic software on you Lucas system? - if not I would strongly recommend it!

Thanks again, take care,
ian

Thanks Ian, I was lucky to get on the GT40 build when things were a bit cheaper. No I do not use Rovergauge, if it runs and pulls I am happy, though I still hanker for the simplicity of a carb!

Ian

If you had wideband sensors checking what that nice simple carb does, you'd likely find similar errors.

With the problem being most obvious at low demand, could you simply have a leaking injector? Does the rail hold pressure, engine stopped?

[Ian C]

. . . If you had wideband sensors checking what that nice simple carb does, you'd likely find similar errors.

With the problem being most obvious at low demand, could you simply have a leaking injector? Does the rail hold pressure, engine stopped?

Dave,

Yes, the fuel rail holds pressure pretty well though it does drop over time it take hours rather than minutes.

After the problem switched sides, when the injectors were repositioned to try and equalise the fuel flow between banks following the flow test, I was fairly confident that I had a sticking / sluggish injector. Four of the injectors had swapped sides when they were repositioned, I swapped two back over and did an engine run - no change. Swapped the remaining two back over, did another engine run and again no change! So now all the injectors are back on the same side of the engine as they started but the it is now the nearside exhaust that is showing as being leaner than the offside whereas at the start it was the offside that was running leaner!?

Starting to doubt my sanity now and wondering if I am chasing shadows. It seems weird for the problem to switch sides following the repositioning of the injectors but to remain when they were changed back. the difference is still between 13% and 15% at idle but drops to less than 5% when run at over 2,500 rpm.

Any suggestion would be welcome as I am running out of options.

Take care,
Ian