Fast acting IAT

Red Raspberry · Sep 23, 2012

I was testing the actual value of the air in the intake and comparing it to the OBDII IAT value. What I found is that the IAT is fairly accurate as to the running temp of the air in the intake, but is is very slow to react to sudden temperature changes.

If I floored it the IAT would take quite a bit of time to change. But the test probe I had showed an immediate drop in temperature.

This was all done with mid 50F ambient temps and the IAT showing ~70F or 15F above ambient. The two temp probes read within 2F of each other when at the same ambient temps.

Anyway, it was surprising to see a big drop to near ambient with in a couple seconds yet the IAT did not budge. Only after the turbo had heated up the air a bit and temps coming out of the IC were around 70F did the IAT start to fall, then go back up, There seems to be a great deal of latency in the factory reading.

So you get a big slug of cold air but the ECU does not see it to adjust the fuel or spark. I would think theres some neck snapping being left on the table here.

I searched and came across a thread about a Fast Acting IAT.

Fast acting IAT sensor [Archive] - Club Protege

They had success using the Triumph 955 IAT sensor.

So I need to see if the resistance readings are the same as ours. They probably are.

itslikeanegg · Sep 24, 2012

Since you were a participant in the high iat thread, do you believe the issue is a slow acting sensor rather than delayed readings due to sensor placement? Do IAT's recorded after the sensor relocation trick still appear to be delayed or will they be fairly immediate?

Red Raspberry · Sep 24, 2012

One of two things. It is either the plastic body surrounding the thermistor or it is a programming delay. I'm betting it is the plastic body at the tip retaining heat.

itslikeanegg · Sep 24, 2012

Given that we've heard positive results with members who have done the IAT relocation trick (logged improvement in timing advance) which combined with your results indicates that the relocation still reads accurate IAT's, is it not possible that the relocation trick removes the effects of a slower heating/cooling manifold on the sensor? The relocation alone seems to improve reaction to temperature changes.

Red Raspberry · Sep 24, 2012

I haven't seen any logs of the IAT relocation that include secondary verification of the IAT temperature reading. I do not think giving the ECU false information is an improvement.

From what I have seen the OEM IAT is relatively accurate, just slow to respond to any temp changes. I cannot see how when using the same IAT module this changes due to the location.

cris90 · Sep 24, 2012

do u think haltech has this delay?

Red Raspberry · Sep 24, 2012

I don't know if there is any OEM delay. It is only supposition.

I've programed quite a bit of automated industrial temp control circuits. A buffer is usually used as the temp swings can cause to much unneeded "churning", though it is usually just a few seconds.

I really think it is the way the MAP/IAT is made with the outer plastic sheathing over the thermistor.

Just guessing but I bet you could unsheathe the thermistor and get good results.

cris90 · Sep 24, 2012

i wonder, but my meth kit had a instant change. when meth started running my IAT dropped, no delay. Could it be because the water when it sprays against the thermostat it allows to heat trasfer through the plastic faster?

itslikeanegg · Sep 24, 2012

I think cold fluids allow for a pretty efficient heat exchange.

Red Raspberry · Sep 24, 2012

Yes you are seeing the effect of evaporation with the meth. And this would be evidence that it is indeed the plastic that is holding the heat.

BTW I scored one of the Triumph 955 IAT sensors.

ozonegc · Sep 27, 2012

nice i shall be watching this thread closely

Red Raspberry · Sep 27, 2012

I got rid of one of the large vacuum hoses in the brake booster line. It is just one large hose now, the one with the check valve in it (very important). I hoped the steel vacuum nipples were a metric thread but they are a 1/8" - 28 BSTP thread. The IAT I'm getting is a M10x1.25 thread.

So either I have to tap and drill a new hole or I try to tap the already threaded hole for the vacuum nipple. Both the 1/8 -28 and the m10x1.25 take the same size hole. Maybe the I can get the aluminum to mash up enough to retap it with the metric thread? Need something to do a practice tap with.....

danperezsax · Sep 27, 2012

I'm sure you already know this, but for the love of God please be careful drilling or even tapping your IM. Take it off the car and make sure you buff or file out any hanging filings/shavings, and try to polish/clean the inside as best you can before reinstalling. I'd hate to hear that a shaving got into a cylinder and wreaked havoc!

Red Raspberry · Sep 27, 2012

Bah, it's aluminum.

Red Raspberry · Sep 27, 2012

Actually the threads are so similar and the OEM ones are so deep that it will more than likely just mash the threads into the new pattern. A minor amount of fine aluminum chips maybe because the bottom of the OEM threads are tapered in. I'll use some grease on the tap to catch the shavings.

GenisisPrincess · Sep 27, 2012

Try blowing air over the thermistor using a fan. I tried getting another cars reading at Idle to decrease using a AC vent hose to airbox and it did not budge at idle with the amount of idle air moving over it. Made no sense to me then but now it might have more to do with movement over the wire. Just a idea....

Red Raspberry · Sep 29, 2012

I got the sensor today. A two wire connector with the spring wire clip like what is on the oil temp sensor clips right on.

It does work very fast when connected to an ohm meter.

I figured out too that if I take the TB off I can easily clean up any threading debris.

Tim@GencoupeStore · Sep 29, 2012

Interested to see what you find

Red Raspberry · Sep 29, 2012

I'm thinking it will help with lean tip in. And maybe a lower timing cut when going WOT.

Tufast · Sep 29, 2012

Good work...subd

SteveJohns29 · Sep 30, 2012

subd.

Red Raspberry · Sep 30, 2012

Here's a break down of the temp readings and the actual ohm readings. I'm going to mount the sensor and connect it to an ohm meter and drive around a while comparing the readings to the OEM temp sensor before splicing it into the MAP wiring.

Here's the Hyundai specs for the IATS.

°C___°F___kΩ

-40(-40) 40.93 ~ 48.35

-20(-4) 13.89 ~ 16.03

0(32) 5.38 ~ 6.09

10(50) 3.48 ~ 3.90

20(68) 2.31 ~ 2.57

40(104) 1.08 ~ 1.21

50(122) 1.56 ~ 1.74

60(140) 0.54 ~ 0.62

80(176) 0.29 ~ 0.34

2.2K Thermistor Output Table

°F___°C___Ohms
-7 -21.67 24063
-5 -20.56 22552
-3 -19.44 21134
-1 -18.33 19826
1 -17.22 18607
3 -16.11 17471
5 -15.00 16412
7 -13.89 15424
9 -12.78 14502
11 -11.67 13641
13 -10.56 12837
15 -9.44 12079
17 -8.33 11376
19 -7.22 10719
21 -6.11 10104
23 -5.00 9529
25 -3.89 8989
27 -2.78 8484
29 -1.67 8011
31 -0.56 7566
33 0.56 7146
35 1.67 6755
37 2.78 6388
39 3.89 6043
41 5.00 5719
43 6.11 5414
45 7.22 5128
47 8.33 4858
49 9.44 4605
51 10.56 4364
53 11.67 4139
55 12.78 3927
57 13.89 3727
59 15.00 3539
61 16.11 3361
63 17.22 3194
65 18.33 3035
67 19.44 2886
69 20.56 2743
71 21.67 2610
73 22.78 2484
75 23.89 2365
77 25.00 2252
79 26.11 2145
81 27.22 2044
83 28.33 1949
85 29.44 1858
87 30.56 1771
89 31.67 1690
91 32.78 1613
93 33.89 1540
95 35.00 1471
97 36.11 1405
99 37.22 1342
101 38.33 1283
103 39.44 1226
105 40.56 1172
107 41.67 1122
109 42.78 1073
111 43.89 1027
113 45.00 983
115 46.11 942
117 47.22 902
119 48.33 864
121 49.44 828
123 50.56 794
125 51.67 761
127 52.78 730
129 53.89 700
131 55.00 672
133 56.11 645
135 57.22 619
137 58.33 595
139 59.44 571
141 60.56 549
143 61.67 528
145 62.78 507
147 63.89 488
149 65.00 469
151 66.11 451
153 67.22 434
155 68.33 418
157 69.44 402
159 70.56 387
161 71.67 373
163 72.78 359
165 73.89 346
167 75.00 333
169 76.11 321
171 77.22 310
173 78.33 299
175 79.44 288
177 80.56 278
179 81.67 268
181 82.78 259
183 83.89 250
185 85.00 241
187 86.11 233

nyangineer · Feb 20, 2015

Were these values posted online somewhere or did you calculate them? The triumph IAT sensors I got are supposed to be the same part number but one says made in china and the other in the UK so I want to make sure I have the right calibration.

cris90 · Sep 30, 2012

Soi was thinking... anyone think that maybe the slow acting IAT is a double bladed sword? Ok it doesn't tell our ECU that the air is cool fast enough to advance timing BUT! it also doesn't tell our ECU the air got warmer say during idle after a hard pull

Red Raspberry · Oct 1, 2012

If I go WOT the air coming in is near ambient so it is denser than what the IAT is saying. More air needs more fuel which can take more timing.

If the IAT is saying the air is hotter than ambient and you are not using a lot of throttle there really is little change in the temp but once again the air coming is is denser than what the IAT is saying.

It's really a short window that this would benefit, say 10 seconds total. I don't think it would help with part throttle tip in as there just is not that much air that comes in. But at WOT and 15 lbs of boost there is alot of air coming in.

Red Raspberry · Sep 30, 2012

I don't think it makes any difference unless you are opening the throttle. IMHO it's the tip in and the WOT areas that this makes a difference. The tip in will happen way before the OEM IAT sees any temp change and there for bases timing on the slow IAT reading. This will cut it more than it should for peak power.

At WOT I think it is more important as this will be for a 5-10 second period. I want to see what the timing drops to when I mash the gas at about 2500 rpm and 4th gear.

Here's a graph of a Mazda OEM IAT compared to the Triumph 955 one.

Red Raspberry · Sep 30, 2012

It's installed

So I put it in. Realtively easy to do as far as mounting. It is not yet connected to the ECU. I'm waiting to get some test done on the road first.

First I took off the TB so I could get in there to clean it up and to see where the senor ended up when installed.

Using a paper towel I laid that in there to catch any chips from the m10-1.25 tap.

As stated earlier what is stock is a British Standard Taper Pipe thread of 1/8"-28, very similar to the NPT of 1/8"-27. It is the same size drilled hole as the m10 1.25 tap when drilling a new hole. Being that it is a tapered thread then at the bottom of the tap there would be some new metal to cut threads into.

Then put a bit of thread sealant on the sensor.

Here's a shot of the MAP and the new IAT with my test probe from inside the manifold.

Then I connected an ohm meter to the new IAT, got my OBDII reader and the temp test probe reader all together so I could monitor them as I started it up and let the engine heat up. I let it run for 30 minutes to get everything as hot as it normally would be under idle conditions.

What I saw was the OEM IAT and the new one were very close as best I could tell with the OEM IAT having such a broad rage before changing to the next temperature reading. The new IAT changed rapidly and smoothly, I assume the OEM one is too, what the ECU does internally IDK. The test probe was always off and when guesstimating the temp reading from the ohm treading the new one read a couple degrees hotter, as did the OEM IAT.

Here it is from the top installed. It takes a pigtail just like the stock BOV solenoid uses.

cris90 · Oct 1, 2012

dude thats nuts....

Red Raspberry · Oct 1, 2012

As expected the new IAT when going WOT the resistance drops quickly 200-300 ohms, within 2 seconds. In ~10 seconds the temps starts to rise again. None of this is show by the OEM IAT.

So the next thing to do, and actually what I should have done first, is see how fast the OEM IAT reacts at the sensor. I have a feeling the ECU is only pulling data from the IAT ever 30 seconds or so which may be the cause of the data lag. If the OEM IAT at the connector reacts as fast as the 955 one then that is the case.

itslikeanegg · Oct 1, 2012

Good work bud! Very interested to see how quickly the OEM IAT sensor reacts before the ECU gets its hands on the data.

Red Raspberry · Oct 1, 2012

Yes I just need to figure out how to do that. I'm thinking of connecting both up to the ECU input wire with a double throw switch so I can switch them on the fly to see if there is any difference.

It's really hard to declare any accuracy as to the actual temperature numbers at this time. There may be corrections added to the OEM signal in the ECU logic.

itslikeanegg · Oct 1, 2012

I think at the very least if a particular temperature equates to an equivalent voltage figure for both sensors you've got a pretty good start. If the ECU isn't contributing to any meaningful delay and the response issue is tied to the sensor this research can help quite a bit. If the delay is a result of ECU programming then we're a bit screwed.

xbeyong · Oct 1, 2012

Fast acting IAT

Our Vendors

Top Contributors this Month

Recommended Communities