another IAT correction disscusion

I'm wondering about a few things arount IAT correction.

If I'm correct, Andre suggested a correction value of 2.5% per every 10degree change in IAT regarding to the ideal gas law and fitting the IAT sensor just before the throttle body. That's what I also made good results. The old tratitional LINK base maps has 2.0% correction.

Yesterday I made some calculations and found the ideal gas law says the density change ~3.5% per 10degree temperatur change. Where does this discrepancy come from?

Regarding to LINK G4+ ECU:

The help file before firmware 5.2 says, that the recommendet IAT location is just before the throttle. From 5.2 on, the suggested position is after the throttle body. Especially with the use of the the charge temperature estimation equation modelle, I can't comprehend why placing the IAT sensor after the throttle should be better? As the most know, heat soak on long idle periods are the biggest problem with IAT sensor placed there. With the fuel equation calculation, we can bias the calculation more to engine side if the IAT are getting to less hot if we placed it before the throttle. In the case of fittment in the intake manifold the IAT sensor getts often to hot (heat soaked) and to much fuel is removed. Biasing the charge modell in the direction of ET will make it even worse.

According to LINK the fuel equation (both modelled and traditional with charge temp estimation turned on) adjusts the air density exactly as per the ideal gas law.

I agree in theorie. In practice the reason for placing the IAT sensor just before the throttle body is, that the manfold gets hot if there is only a low air flow inside (e.g idle). As longer the engine idles, as hotter it gets. The problem is that, the sensor gets "heat soaked" from the manifold and reads higher than the air is. That's because the air has to less time while it pass through the manifold to heat up to the same temeperatur as the sensor reads. 

So in practise I often have to lower the IAT correction value on idle to prevent the engine from leaning out on long idle periods. That's the reason I asked. I think the Ideal gas law correction which is ~3.5% substracts to much fuel in this situation. Anyway I have to make more testing on the new modelled fuel equation. It's good that theres is still the IAT correction table, in the case it should be needed.

Regardless of all the theory above. I played today 2-3h with the LINK G4+modelled fuel and the charge temp approximation table for the first time. I can't get my head around, how I should tune this table best.

Does anyone have pratical tips how I can get it dealt in in time?

It's a Subaru GC8 with top mount intercooler and IAT sensor placed in the intercooler just before the throttle body. With dyno fan of it's quite easy to bring IAT temps up. I tried to adjust the charge temp table so, that AFR are constant within 20°C and 70°C IAT.

I found that as more I bias the charge temp map to ECT, as less fuel is taken out a higher temps.

The process takes quite alot of time because, I had to heath up the IAT several times for any load point, and retune the fuel map.

I thought about, if it would be the better way, to first tune the fuel Map at 70°C IAT temps and than adjust charge temp map at 20°C to get the right AFR's? Because the influence of charge temp approximation map value is as less as the ECT and IAT are closer to each other.

If the charge temp approximation map is dealt in, I'm wondering how cold start will look like?

Maybe I'm doing it totaly wrong and someone can bring some light in the dark?

Hi Adrian. If you calculate the change in air density with temperature, you do end up at about 3 degrees per 10 deg C step. I have found that this ends up being a little aggressive and 2.5 degrees per 10 deg C in my experience has given more stable results. If you are using an ECU that uses a VE fuel model then the handling of the IAT is taken care of in the main fuel model.

Now to discuss the sensor placement. Your point is valid that with a conventional ECU, placing the IAT sensor pre throttle is definitely superior - As you've found out, the heat soak can have far too much impact on the fuel equation forcing a lean condition at idle when the manifold is heat soaked. With a charge temp approximation table however this can be dealt with separately. I'll explain:

It sounds like you have a good handle on the way the charge temp approximation works but maybe you have missed an important aspect. In my own testing, I've ended up biassing the charge temp reasonably heavily towards ECT at and around idle. My 86 for example is using a value of about 60-70% in the idle areas (60-70% weighting towards ECT). The net effect of this is that the fuel model will do two things:

1. The charge temp value will be lower at cold start so the fuel model will actually do some of the warm up enrichment for you in the back ground.

2. At normal operating temp the charge temp will be much higher than a conventional IAT sensor will read - Perhaps 60 degrees C or more. It will also be reasonably constant compared to IAT alone. This means that you will actually need to increase the values in the main efficiency table around idle to get back to your target lambda value.

So far I haven't formed a fool proof technique of how to approach tuning the charge temp estimate table. It's complex because you can make changes in more than one place to achieve what appears to be the correct outcome. What I can say is that in our 86 I got much better consistency with the high ECT bias. We are considering a webinar on how to deal with this table but right now I need some more time to do some testing.

A Webinar would be great based on this. I can understand that it would be difficult to try and come up with a "lesson plan" due to it being quite different from engine to engine, sensor positions etc.

I have a RB26 with a high end IAT sensor in the factory position and I am still trying to get this right. It is also quite difficult because I'm running a Link G4+ with Modelled multi fuel, one fuel map, and BP98/E85.

Looking forward to see what you guys can come up with.