Since our cars have horrible AIT sensor heat soak issues, on top of having extremely slowly reacting sensors, it seems like there would be a good benefit to have a more accurate, faster reacting system. Many relocate their sensors to an area less prone to heat soaking (away from the UIM). This may help, but doesn't address the real problem IMO.

I did some experimenting tonight and it proved to me that the stock sensor is very sluggish when it comes to registering temp changes. I blew a hair dryer into the primary port of the UIM and kept an eye on the temp and voltage readings from my commander. The temps did go up fast. After blowing compressed air into the port to clear out all the hot air I just blew in there, the sensor dropped only a few degrees. Then took about a minute to return to reading ambient temp. It dropped steadily throughout the minute, but there was no actual change of the temp in the UIM.

So my plan is to use a fast reacting thermocouple, placed in the elbow, wired to a simple circuit translator board. Then the translator sends a 0-5V signal to the ECU.
http://www.simplecircuitboards.com/Thermocouples.html

Output calibration wouldn't be hard at all to figure out. We could just heat the stock sensor up to various temperatures, and log the temperatures as well as the voltages (sensor check screen) via the commander.

The ECU interprets a 0-5V signal. The sensor is a PTC (positive temperature coefficient) type, meaning as temp goes up, so does resistance. So I'm guessing the ECU sends a 5V reference signal, and uses the returned signal as it's temp signal. The stock AIT sensor has a green wire and a brown/black stripe wire. I'm not sure which one is which, but that's not hard to figure out.

So the output from the translator box spliced to whichever wire is the temp signal, and done.

Thoughts?

 

Sounds like an improvement.

What direct effect would it have on how the car operates?

Dave

 

It probably took a while to cool due to heat radiating from the surrounding material (UIM) as the aluminum normalized in temp to its surroundings.

I look forward to seeing your results.

 

The stock AIT sensor is a NTC (Negative temperature coefficient) thermistor: when the temp is high the sensor's resistance is low. When finding the resistance vs temp curve, it's a good idea to keep the sensor at the same temperature for a few minutes to let it stabilize. Here are some numbers from the stock IAT sensor from my car:

Temp (C) / Resistance (ohms)
-10 / 9570
9 / 3960
25 / 2080
40 / 1175
50 / 832
70 / 440
80 / 321


I'd be very interested to learn the temp vs resistance data for other OEM sensors, I've only tested mine.



By the way, the GM IAT sensor has an open element sensor, it's not encased in brass like our OEM one is. It responds much more quickly. I'm not the biggest fan of GM connectors (difficult to de-pin them) but the sensor is pretty good.

-s-

 

I have a datalog in this thread comparing the temperature change rates of the stock in the greddy elbow and a GM open element sensor..

https://www.rx7club.com/showthread.p...&highlight=ait

 

Could you explain what the columns mean fritts? I'm a little confused while looking at your log.

Do you have the GM open element sensor signal being read by the ECU? Or did you just hook it up for datalogging/comparison data?

 

Same scale (75-150 F) used for both the GM AIT (green trace) and the OEM sensor (red trace). Pink trace is manifold pressure, blue is RPM. GM sensor is very noisy, it looks like one of the wires was not connected well enough (temp even spikes up to a nonsensically high number later in the datalog, that usually indicates a broken or disconnected wire).

Looks like the GM sensor measures an increase in temperature before the OEM sensor, and it gets up to a higher temperature as well.

The smaller sensing element has less mass, it will heat up more quickly than the larger OEM sensor.


-s-

 

Thanks for the data. Very interesting stuff. Did you get the noisy readings figured out?

Seems like the GM sensor works differently than our stock sensor, as others have thought about using the open element sensor as an input to the ECU, but aren't able to. It's a 2 wire sensor, probably also an NTC thermistor, so I'm guessing it's a calibration issue. Maybe this sensor with some kind of translator box would be just as good a setup.

 

Since we use a speed density air measuring system (MAP sensor), knowing the density of the air which changes with temperatures is vital for calculating fuel injection. So with more accurate readings, our ECUs have a better input in which to base fuel.

 

Not my data, not my car. Just thought you might appreciate seeing a graph of what fritts posted.


With the EMS that I have, changing from the Mazda sensor to a GM sensor is as easy as wiring it in place of the stock sensor and selecting 'GM AIT sensor' from a wizard. If I somehow come across a sensor that isn't in the wizard I can just punch in the voltage vs. temp data myself.

There must be a way to make the GM sensor work with the PFC, because fritts was using a PFC to create the datalog that he posted.

 

Looks like the GM IAT sensor could easily be scaled/conditioned to match the Mazda one with a simple resistive pad. If someone wants to send me a GM one (that I can keep) I'll design the circuit and post it on the board here for everyone's benefit.

 

Its the PFC datalogging the extra AIT sensor. I was doing intercooler efficiency for a while with two GM sensors. The stock AIT was only used for the ecu as I wanted WOT accurate readings that the stock sensor would never end up showing.

 

There is a chapter in the Logworks3 Manual that has info on thermistors as commonly used in cars. It's chapter 4.2 "Thermistor Temp Sensor Configuration"...

http://www.innovatemotorsports.com/s...ks3_Manual.pdf

It looks like all that is needed is to add a resistor or replace the resistor (if a smaller resistance is needed) from the circuit to use the GM open element IAT sensor. I hope somebody figures this out and posts the how to as heat soak lean conditions are probably a contributing factor in many blown rotary engines. I may do a little experimenting myself when I find the time. Good luck.

 

A mentioned, moving the AIT sensor to the IC outlet or TB elbow is the first way to reduce heat soak. It works great on my FMIC outlet.

An overlooked way to improve the stock sensor's reaction time is to carefully remove the black plastic coating on the sensor tip until you reach the light brown colored covering of the sensor.

I posted this a few years back in the PFC forum and it does speed up the responce of the sensor.

 

I was thinking the same thing, just because the hot air was removed briefly doesn't mean the manifold wasn't radiating heat onto the sensor still, and it would take a couple minutes for that to cool down.

Not to say that the stock sensor isn't slow, just kind of musing over any real gains you'll see from a different sensor if placed in the same area...

*subscribed .

 

looks like alot of potential here if we get it right.

hc

 

fritts removed some of the plastic from the stock AIT sensor above. The GM AIT sensor was still way more responsive. I wonder if the GM sensor is faster because it's more exposed yet or because it's made of a different material. digikey.com has thermistors that are a lot faster reacting than others. If the GM sensor is faster from being exposed and made of the same material, then theoretically removing all the plastic from the stock sensor should give the same response time as the GM sensor. Some experimenting may be required. I think we would all like to have the fastest sensor as is financially reasonable and at $10-$15 the GM IAT sensor offers a lot for the money if it can be retrofitted easy enough.