Mitigating AT sensor heat soak effects

[mrsaturn7085]

Car is a 2006 USDM Subaru STI 2.5L running MOTEC M800 plug-in. AT sensor is located in the intake manifold behind the throttle (JDM sensor/manifold location).

I've noticed some issues lately with the AT sensor becoming heat soaked during a hot-restart after sitting for 20-30 minutes. When heavily soaked (70-80 deg C) the sensor takes ~10 minutes idling to cool down (all the while running at ~1.10 lambda and struggling to stay running). Normal driving cools the sensor in 2-3 minutes.

When the motor was tuned, the AT fuel comp. table was set as a 3D table with the axis being AT vs. MAP. Fuel followed the ideal gas law from 60 kPa up; from 20-40 kPa, the fuel enleanment was reduced above 40 deg C (40 deg C is my zero-point). First problem was obvious - the fuel enleanment reduction is not reduced enough. My second concern is that the AT vs. MAP would be better off as AT vs. TP. Since my motor is able to reach 20-40 kPa during cruise conditions, I feel that it would be difficult to maintain accurate fueling cruise with the AT fuel comp. set in this fashion.

I have enclosed a few images showing my previous and proposed AT fuel comp. tables and I wanted to hear some opinions on if this is a correct move, or if there are better practices for mitigating the heat soak effects.

[Rude Engineering]

The correct amount of air temp compensation will depend on the location of the sensor but the values you have look higher than I would normally use. I would normally say 1.5% trim per10 degrees but capped at 5% trim at 50 degrees. Although the air density reduces with temperature I think that if things are getting that hot I would prefer to have the engine running richer. (of course this is application dependent but I am talking about normal conditions). The standard MoTeC correction is as follows and will work well in most applications:

-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100
16% 13% 10% 8% 6% 4% 2% 0% -2% -4% -6% -8% -10% -12% -12% -12%

If you are having problems with heat soak on the sensor and relocating the sensor is not an option I would look at the plausibility of your inlet temperature reaching this temperature, is it likely to reach this temperature under any other conditions? What are the "normal" operating parameters of your engine and is it really going to cause problems if you apply less trim at this temperature?

[mrsaturn7085]

Normal AT is 40 deg C, post-I/C and post-throttle.

MOTEC base map AT fuel comp. is centered on 20 deg C, which is better suited to N/A applications. Furthermore - WRX9-10 white papers indicate the AT sensor used for plug-in applications was the factory pre-turbo AT sensor (pin B18), rather than a post-turbo/throttle/intercooler sensor; as such, the MOTEC base map compensation tables would have been fairly mild as the sensor could not accurately meter the temperature of the air entering the cylinders.

My compensation table is based on the ideal gas law. Dulling the comp. in the high and low end may make sense, but the air density correction in my table is scientifically accurate.

My heat soak issue is taking place after a hot-restart - AT rises to the ET level since no airflow is present in the intake manifold. This is the ONLY time there is a major issue.

Sitting at running idle causes AT to rise a little, but ~5 deg C tops - nothing like the hot restart issue I am encountering.

[Rude Engineering]

I understand the science behind but the sensor is not actually located at the point where the air is consumed by the engine. Bosch use a similar correction to what I have suggested and I have found it usually works well (of course depending on the sensor location. From what you have said though it sounds like you could solve your problem by using a compensation that effectively ignores the high temperature value (or partially ignores it). As this condition is only present post hot re-start it will not have a detrimental effect on the running of the engine.

[mrsaturn7085]

AT sensor is located between cylinder 3/4 runners inside the intake manifold - short of being next to a valve it's as close as I'm gonna get to the actual air charge temp.

I'm considering a engine run time based compensation but haven't played with the table to see what I can do - basically make the AT comp. zero above 40 deg C until the engine has been running 2-3 min or so - a 4D table would sure be nice here... maybe I can utilize one of the generic fuel 'Comp 1/2' tables to counter the effects of 40+ deg C compensation based on low engine run time...

EDIT: Like this -

[Rude Engineering]

I was going to suggest that but the only problem is that from what you said earlier the effect of the heat soak reduces quicker if you drive the car (as opposed to the engine idling). This is the reason I thought that plausibility was a better option... It could work ok though!

[mrsaturn7085]

I was going to suggest that but the only problem is that from what you said earlier the effect of the heat soak reduces quicker if you drive the car (as opposed to the engine idling). This is the reason I thought that plausibility was a better option... It could work ok though!

It does, yes - 2 minute of driving usually resolves the issue compared to ~6 min of idling. I tried to scale the Run Time axis appropriately.