Idle Air Control valve causing my Lean under elctrical load?

I've been tearing my hair out trying to find whats causing my Lambda to go lean when the fans/lights etc are on at idle (Toyota 3sgte with stock 3 wire idle air contol valve)

I may have stumbled on the culprit would it make sense that the IAC would get lazy when theres a high electrical load on the engine? I'd assume it'll be sticking open to go lean.

How lean does it go? (Both absolute and relative to your target Lambda).

I would expect the IAC to open up as the load is increased to keep the same idle speed.

The problem may be that the IAC air isn't metered. What are you using for the load axis?

Or the decrease in battery voltage isn't compensated in the injector setup, so less fuel is delivered when the voltage drops. What do you have for injector dead times relative to battery voltage?

How lean does it go? (Both absolute and relative to your target Lambda).

I'm trying to tune my idle area and just seem to be going around in circles. I've just fitted a new set of 1000cc injectors and I'm not confident that the deadtimes and short pulsewidth data are accurate (at least for my setup). Lambda can swing 0.4 either way seeming depend on which way the wind is blowing at the time (at idle).

I would expect the IAC to open up as the load is increased to keep the same idle speed.

currenly have set IACV to open loop, just to give me as rough idle without adjusting duty cycle.

The problem may be that the IAC air isn't metered. What are you using for the load axis?

Load axis is MAP so i'de imagine the air entering should be accounted for.

Or

the decrease in battery voltage isn't compensated in the injector

setup, so less fuel is delivered when the voltage drops. What do you

have for injector dead times relative to battery voltage?

The battery voltage itself isn't dropping much and pretty much settles back to where it was before the fans kick in. I'm honestly at the point of just re wireing the whole engine harness as I just seem to be chasing my tail for months now. The problem seems to have gotten worse since changing to the new 1000cc injectors (previously 540cc) so I'm assuming the small pulse width is having a bigger effect. It's just frustrating when I paid for a set of injectors for the sole purpose of ruling out the injectors as the problem.

Is there a particular order to approach tuning in modelled mode? There seems to be alot of tables that can affect what ve numbers are inputted but something sureley has to be tuned first.

The first thing, is always tune the fuel when the engine is a narrow "normal operating range". Say 80-90 degC. Don't fall into the trap of adjusting fuel just after startup -- just ignore that for now and only tune the fuel map when the temp is warm. Once you have idle control working and can maintain a steady idle following startup, then you tune the Coolant Temperature Compensation as the engine moves from cold to normal operating temp.

Also, make sure any closed loop fueling is turned off, you don't want it fighting you while you try to make changes.

If you have idle control don't enable that until you've got the fuel correct -- you can use the IAC setting to cause different cells to be used, or adjust a manual throttle stop while tuning the idle area of the fuel map.

Make sure you are using all the table resolution -- some systems have more resolution available you just need to specify the properties of the table or value.

Thanks David, I've had a little play this afternoon and loged basically the car idling and you can see the big swings in lambda as the fan come on. I also saw the fueling richen out a bit as IAT increased (when fans weren't running) so took the opportunity to tweek that a few points closer to ect so i'm hoping now charge temp at idle is set.

I'm still thinking my deadtimes are out (despite been given deadtime data when I bought them) as If i'm say 10% rich and I adjust my fueling by 10% it's swings to something like 20%.

I got told to switch to batch fire injection as that should highlight when deadtimes are set but I don't think that is working with my injectors running such small pulsewidths it just runs off the scale rich

if switching to batch results in too rich, then your deadtimes are too large, try decreasing them. That also correlates if you are running sequential and remove 10% fuel but it reduces by 20%, another indication of too large of a deadtime.

You will then need to re-tune your VE in the running area, but that is an excellent way to set the deadtimes.

Also, your ECT compensations will need to be revisited, since +5% should really be +5% not something less.

See, I've tried switching back and forth from sequention and batch and no matter what I did with the deadtimes I couldnt get them to be anywear close to each other. I'll try a few datalogs tomorrow and see if i can show what I mean. The only thing I can think, is it possible with the injector pulsewidth being so small they are going to be even smaller again in batch fire and fall below what the injector can physically opperate?

I think off the top of my head in sequential i was at around 0.6ms (effective) and would batch mean I'd be down to around 0.3?

A crude test you can do to validate deadtimes are in the ballpark is as follows:

Set up engine at a fast idle around 2000RPM (just set idle target to 2000 if you have it in closed loop), allow temps to stabilise.

Set the whole lambda target table to 1.0, adjust fuel table so that the engine is running with a measured lambda of 1.0.

Change target lambda to 0.90, if dead times are correct the measured lambda will also change to 0.90. Do the same test in the opposite direction - say target 1.05 or whatever it is happy to go to without misfire. If measured lambda doesn't follow target lambda then tweak deadtimes until it does.

Note this is not a perfect test as injector non-linearity and charge cooling effect etc plays a part, but it will usually give you a good validation.

Thanks Adam, I've usually got a window of about 5 mins from getting into operating range before the fans start cutting in and everything starts getting thrown out of whack but i'll give it a bash. If I ask for 0.90 and it overshoots would that mean I need to increse or decrease the dead time?

Is your fuel pressure regulator referenced to your manifold pressure?

What does the fuel pressure trace look like?

I don't understand how at 33:20 in the screen just above, the effective pulse width only rises about .1ms, yet the overall pulse width rises at least .25 ms. This could only happen if the deadtime were increased, and the MAP values is what has changed at this time. Could the ECU be calculating a change in differential fuel pressure, even when you should have a constant differential fuel pressure if the Fuel Pressure Regulator is referenced from the Manifold Pressure. Possible setup issue?

Fuel pressure is set to 3 bar differential and a fuel pressure sensor fitted. To be honest by the end of the log I probably was messing on with Dead times / electrical load, that's was the big jump will be. I turned all the lights, heaters and fans on to drop the voltage down. which in turn sent the lambda values all over.