Talon blow thru is (was prior to speed density patches being widely used ) popular with he Subaru . It draw by factory but I have a spare subie maf as I had a 07 wrx with sti swap.. Kept the td04 though

 

Nothing is said about whether the test was done in flowing fluid. Convection, i.e., forced flow, plays a big role in sensor response time. The GM IAT sensor element is fairly small, and I expect that under high air flow conditions, the response time is way faster than 15 seconds. Placing the element so that face of the element sees the air flow would help too. Also, keep in mind that IAT fine tunes AFR, and IATs should not change by a huge amount in a relatively small period of time.

 

I think this is why speed-density actually works. The IAT is only a very small part of the total calculation. But this is probably why Aaron's method also works.
I don't really know how fast your air temp changes when you go from, say, cruise to WOT. And I still wouldn't be sure about that even if I was logging with an IAT sensor because I wouldn't know how much my log values were being affected by the lag of the sensor!

For sure moving air will push the sensor faster than still air, and AEM should have stated if their test oil was stirred or not.
Never the less, kudos to AEM for having the guts to even print a spec at all for response time. I imagine most of the other aftermarket suppliers would rather avoid the issue.
For fast moving air the uicp is the place. Still, I don't think you could ever get an IAT to respond fast enough to be what I would call "real-time" data. It's laggy data. So where Aaron, in that other thread, said that using an IAT has "all sorts of little issures" and "There is no such thing as true speed density" - I would count slow response as being one of the little issues, and one reason why there is no such thing as true speed density.

I'm going to post a better (I mean more info) thermal response spec that I found. This is for a TI family of temp sensor chips. This info won't be accurate for the specific IAT sensors we would be using, but the shape of the curves, the basic characteristics, should be similar. The thing I couldn't find in this TI spec is a definition of their thermal time constant - where AEM says theirs is at 63.2%, TI doesn't say what %. Here's their spec:





and this from a different page in the spec:

 

Following the same criteria that AEM used, the TI sensor would have a response time of ~1 minute in still air and 1 second is a stirred oil bath. The huge difference is in-part due to forced flow and in-part due to oil having a much higher heat capacity than air.

I still think the response time of the GM sensor is not nearly as poor as what the AEM spec sheet suggests. I think its in the ~2 sec range for fast moving air. I have a very low time constant type K thermocouple that I can install in virtually the same location as my GM sensor and can log both simultaneously. I'll have to see about doing that soon.

 

This is probably about right. Even for the TI sensor above, in Fig 11, the "time constant" gets down to about 2 sec at 2000 FPM (where it is kind of leveling off). Then the question is what % of final value is the "time constant" figured at. Looking at that a couple different ways -
One way is combining fig 11 and fig 12 for 0 FPS / Still Air, the faster curve (To-92) is at about 20 sec in fig 11, 20 seconds on fig 12 is at about 40% final value, hmm, not a very high percent.
The other way is combining their fig 13 for stirred oil with the spec given in the chart at the bottom of post 33 where it says time constant for stirred oil is 1 sec. Looking at fig 13, 1 second is at about 70%.
So probably we could say TI's "Time Constant" is figured at somewhere between 40% and 70%.

Love it when a good plan comes together!

Anyway, 2 seconds in fast moving air to reach about 2/3 of the true final value (delta actually) - my guess - for this particular sensor.

 

But you guys surmise that's well enough for fuel metering ?

I wonder how the maf iat stacks up?

 

Well I guess it's good enough because a number of people seem to like their results with "speed-density". But I can't have any hard-core opinion about it having never run it myself. I just think it's a good thing to know about, and interesting, and might cause you to rethink some things.
For example, if you are going to run a wmi nozzle into your uicp, where would you put it? Would you put the nozzle closer to the throttle body than the IAT, or farther away? I don't know. I think it would be up to the person who is going to tune the beast. They would know how they want to compensate everything and what works for them. But now I know to expect that if I put the IAT closer to the TB than the nozzle, I'm not going to expect the IAT to really pick up that first step downward in air temp instantly when the nozzle comes on. It will take a sec or 2 for the IAT to converge, and it never really will converge because it will always be lagging behind whatever is happening during these transients.

 

Vipec iat. Appears internally similar to echlin/gmiat/aem .

But is 1/8 npt.

 

I looked at this sensor a while ago and agree that its internally identical to the GM IAT but is packaged as 1/8" NPT instead of 3/8" NPT. There are websites out there claiming that this is a super fast response time sensor with the implication that its better than anything else out there including the GM IAT.

 

Thanks for mentioning the Vipec IAT. But good-grief, I went to the Vipec web site looking for it and I couldn't find it there! What the heck?

Then I got looking at the configuration manual for the Vipec V44 and V88. In the IAT section it says this:

"It is very important on a turbocharged/supercharged engine that the air temperature sensor can react fast enough to track the rapidly changing temperature. For this reason, an open element sensor is required. The recommended sensor is a Bosch 0 280 130 085."

Well that's what they say.

BTW thanks again for mentioning Vipec because I just noticed that the AEM EMS #30-6300 for my car (1990 Talon) is not listed anymore on the AEM web site and it's freakin me out a little bit. Trying to find out right now if that is just a boo boo or what. The Vipec manual is nice - that's a real manual.

Here's a pic somebody has taken of the business end of one of these Bosch sensors:

 

I use a air temp sensor from a Triumph motorcycle on my own car. It is a decent amount faster than the GM sensor. It believe the part number is T1290510.

 

The GM is an open element sensor as well. That Bosch sensor is a bit scary to be using - there's nothing to contain it if it lets loose.

Here's a picture of the Triumph one:

 

the vipec one claims to be BSPt. be careful about that. They are different taper's and may not like to be installed in NPt. . . Just in case putting that out there.

 

ok. I was just trying to point out that it looks to be the same sensor with slightly different packaging.

 

Could you guys give me a link to that vipec IAT sensor? I can't find it. All I can find is mention of it in older forums like from 2009, 2010.