heat soak on IAT

[snillet]

I´ve after many many miles of driving identified a heat soak problem on my IAT sensor. It is located in the intlet manifold on cylinder 1 just before the head (about 10cm, or 4 inches before the intlet valve).

When "half warm" being stopped at a store or something and i go out to start it up again it goes WAAAAAAYYY rich until i´ve driven it at mid (2500:ish) rpm:s under load for a little while.

Also when i´m stuck in traffic jams and slowly moves forward, the longer i stay there the hotter the spot where the IAT sensor is located, gets.

And the car gets really lumpy and awkward (spells like that ?) to drive.

I´ve noticed that at all those times the IAT goes WAY up to about the same as the CLT.

I´ve noticed that on my other car the IAT is located just BEFORE the throttle housing.

That position would NOT give such a bad heat soak of the sensor, but will the system still work as intended ?

When i saw the installation on the other car i got the idea that the important factor might be the temperature of the air entering the throttle housing ?

Please advise me if you have ideas

[superchargingmachine]

I think a few people have had this problem. I have the same idle AFR issue and have struggled with it for some time.

In my opinion since you will set your VE table up with a fixed sensor location the absolute value of the air temperature is not critical. Since the VE will be adjusted for the sensor location(hotter closer to the engine or colder farther from the engine) air temperature. But, you definately need to be measuring air temp not the manifold or block temp as happens when the sensor is heat soaked.

I think for non boosted engines the easiest way to avoid heat soak issues is to measure air temperature out ahead of the throttle body or in the air cleaner.

On boosted engines since the turbo or supercharger adds heat to the air, that we want to adjust for, it is more difficult to avoid heat soaking the sensor since the sensor will be closer to the engine.

Remember that if you move the sensor you may have to adjust you VE table slightly.

[keithmac]

I`ve been working on a code using ego feedback to adjust the VE at idle, this eliminates any flutuation due to IAT heatsoak. I`ve found idle to be the worste affected when the IAT is heatsoaked, supposed closed loop ego correction could be used just as well but I don`t like giving the ego correction more than 10% authority at idle.

Works a bit like autotune in MT, updates the VE every 1 second (ram value). I think I may have found the point at which the code is completely full though (bootloader problem?), still invesigating..

[superchargingmachine]

I agree idle is the worst condition for heat soak.

I tried closed loop at idle but didn't have much success. But, I didn't put a lot of time to put into it.

I was considering using 2 IAT's and switching to one infront of the turbo when I am off boost with an output. I figured this would also cure the problem.

I would like to try your code if you get it working. I'm not sure about the code being completely full but I think it has been close for a long time.

[keithmac]

It needs a fair bit of testing first, I`m trying to see if the main code is actually too full (I have a big wedge of other stuff in there as well) or if it`s a mistake in my coding (complies fine and it`s nothing fancy..).

[chicksdigwagons]

On boosted engines since the turbo or supercharger adds heat to the air, that we want to adjust for, it is more difficult to avoid heat soaking the sensor since the sensor will be closer to the engine.

Remember that if you move the sensor you may have to adjust you VE table slightly.

I don't think I'm following you here. A boosted application should not suffer any more or any less from heat soak of the IAT sensor assuming identical locations. You just need to pick a point where engine heat has little effect on the sensor. The intake manifold can be a bad location for two reasons. It has more thermal mass than typical intake piping, and it also is usually not isolated from the head thermally. After a car sits idling after being run, the heat from the head soaks into the manifold now that the air (both internal and external) is not cooling it. This increase in manifold temp does not directly coorelate to IAT. This would happen regardless of boost or not.

I would wager to say the intake piping immediately before the TB is ideal for most engines. I would also suspect that some IAT sensors are more suceptable to heat soak than others. The open element variety is likely superior to the closed style in this respect.

[snillet]

I´m running closed loop on idle and in order to make it work somewhat good i have to set the authority to +/-35% , to make way for the differences in temperature when heatsoak happens.

But at that time the system really is working under "screwed" conditions...cause of the heat soak .

So i´ll move the sensor.

When the heat soak is really BAAAD, it takes quite a bit of driving , about 1kilometer before the sensor cools down again as it is now. Which is no good.

[superchargingmachine]

On boosted engines since the turbo or supercharger adds heat to the air, that we want to adjust for, it is more difficult to avoid heat soaking the sensor since the sensor will be closer to the engine.

Remember that if you move the sensor you may have to adjust you VE table slightly.

I don't think I'm following you here. A boosted application should not suffer any more or any less from heat soak of the IAT sensor assuming identical locations. You just need to pick a point where engine heat has little effect on the sensor. The intake manifold can be a bad location for two reasons. It has more thermal mass than typical intake piping, and it also is usually not isolated from the head thermally. After a car sits idling after being run, the heat from the head soaks into the manifold now that the air (both internal and external) is not cooling it. This increase in manifold temp does not directly coorelate to IAT. This would happen regardless of boost or not.

I would wager to say the intake piping immediately before the TB is ideal for most engines. I would also suspect that some IAT sensors are more suceptable to heat soak than others. The open element variety is likely superior to the closed style in this respect.

I think we are basically saying the same thing. In boosted applications you want to measure the heat added by the turbo or supercharger so this limits your sensor locations. I have had my GM open element sensor in the intake and in the tube that runs between the turbo and intake and both heat soak in my application. Part of my problem is that my car is mid engined and my 2.5" tube from turbo to intake doesn't have much air velocity at idle on a 1.6L engine.

Does anyone know of a sensor that is either smalll that doesn't have a metal body that will absorb and hold heat?

[snillet]

Problem is that it´s not only the body of the sensor but the radiating heat from surrounding metal surfaces that heats it up and the airflow is to slow at idle to coll it down to the same temp as the airflow.

So you need a SMALL sensor with quick responsetime, i´m thinking of purchasing the NTC sensor as a high spec electronic component and place it at a spot where heat soak is LESS likely to happen.

[superchargingmachine]

Sorry, what is NTC?

[keithmac]

It stands for negative temperature coefficient, most Bosch and GM temp sensors are ntc.

I got the autotuning idleVE working tonight, works very well. Set it up so if ego goes under 95% or over 105% the VE is adjusted by 1 point, ego is checked once per second to aviod it running away with itself. Should work for IAT heatsoaking problems.

Maybe an idea to have a VE multiplier that checks what the target AFR should be and what it actually is, then shift the whole table up or down instead? Only problem with all this is the car will be reliant on the Wideband sensor, I suppose if you only modify up to 80-90kpa it should be fine? Or just leave the ego correction to it.

[snillet]

Sorry, what is NTC?

The type of temperature dependant resistor that is inside the IAT sensor.

The component itself rarely costs more then $1 , and a high spec one might cost up to $3 but that is a REALLY good sensor, the ones you find in the IAT sensors from the automanufacturers are usually the cheap kind.

The more expensive ones react quicker to changes in temperature, and often have smaller bodies, therefore they MIGHT, i say might...., be lsee prone to heat soak problems.

[PSIG]

I can't find the reference now, but someone used the NTC sensor from Digi-Key. They placed it in the middle of the pressurized intake duct at the end of a light plastic tube that had very low mass. Almost instant response even in slow air after hot start was reported.

Someone else wrapped the insulated legs and leads around a wire (or was it a fiberglass cord?) that stretched across the throttle body inlet with the sensor in the center. The number of low-mass mountings are endless based on your materials and fabricating ability.

David

[superchargingmachine]

After someone was kind enough to tell me what NTC was I did a little research. NTC's are basically Resistance Temperature Devices(RTD) and it seems that there are many available. I would agree there are many ways a guy could mount one. There are even extremely small RTD's, which should help even more. My only concern would be making something to hold the sensor that won't allow it to break loose and go through the engine.

It seems like the OEM's must have come up with a different solution. Either that or there should be a good sensor already available. I guess it could be that they spend a lot of time figuring the best location for a sensor too???

[chicksdigwagons]

It seems like the OEM's must have come up with a different solution. Either that or there should be a good sensor already available. I guess it could be that they spend a lot of time figuring the best location for a sensor too???

How often do the OEMs really come up with the best or even a decent solution to problems? Largely cost and convenience of manufacture take front seat to innovation and ideal solutions. As someone mentioned earlier they use the cheapest components they can to provide the minimum service life they deem acceptable.

If you've ever looked closely at a GM caged sensor, you can clearly see its just a $.50 thermistor inside a little plastic cage. So for the price of a thermistor, threaded brass plug and a some goop you could recreate the sensor for $3. Being there are much better and more responsive units availible, for $5 you could create a superior sensor. Now use a plastic NPT plug and now we're drasticly improving the design.

Again, I would have a concern that this could potentially break off and end up in the engine, however, I have seen OEMs use a non-caged design. Namely in snowmobiles, something that likely sees far greater vibration than a car.

So I don't think its fair to say the OEMs have really spent that much time working these things through, they just care about the bottom line. Case and point, a lot of the design constraints were dependant on using these sensors for both Intake air and coolant. I think we can all agree that the qualities necessary or ideal for a coolant temp sensor is not the same as for an intake air sensor.

[superchargingmachine]

Maybe my comment wasn't worded correctly. All I was saying is that they have somehow worked around the heat soak/idle issue.

I am sure you are correct that with a little thought and a good RTD($5) one could make a far superior sensor that wouls be safe.

[woh]

NTC's are basically Resistance Temperature Devices(RTD)

NTC is not the same as an RTD. The terminology is usually NTC Thermistor. An NTC thermistor temperature sensor commonly come in 2k, 5k, 10k etc values at 25ºC. They have a Negative Temperature Coefficient, are ceramic. and they have a high change in resistance per deg C

read about it here:
http://www.ysitemperature.com/

RTD is a metal (platinum) temperature sensor. Usually a low value like 100 ohms and has very small changes per deg C.

[superchargingmachine]

Thanks for the correction:-) I guess when I saw that they work on resistance I thought they were the same thing.

I will have a read to see what I can learn.

[subie4me]

On boosted engines since the turbo or supercharger adds heat to the air, that we want to adjust for, it is more difficult to avoid heat soaking the sensor since the sensor will be closer to the engine.

Remember that if you move the sensor you may have to adjust you VE table slightly.

As I understand it, as the air gets hotter it is less dense. This requires less fuel to get a desired AFR. Problem is , that hotter air makes a turboed motor more prone to detonation. The last thing I want to do is take away fuel due to the hot air and further compound the problem. The extra fuel helps to prevent detonation. For that reason, I have NEVER used the IAT sensor on any of the megasquirts that I've done. I also didn't use it on the link plus that I have on an offroad car. Maybe you need it on a street car, but you do not need it for offroad use.

[keithmac]

That`s an interesting angle on it! Maybe a part solution would be to have the IAT reading ambient air temp to compensate for winter when the car needs more fuel on warmup etc?