tici

iTrader: (1)

Hi,
I'm now running SD and I see pretty high variation in the fuel trims depending on the IAT value.
So far it looks like that the IAT has to see the real outside temperature and not that one of the air flowing into the engine (I'm blown!)

Higher IAT values than reality produce positive fuel trims changes. It's logical: high temperature means the engine sucks less oxygen, the PCM injects less fuel, the O2's complain and the PCM adds fuel = positive fuel trims.

Where do you SD guys install the IAT?
In my case it looks like the sensor is reading the heat of the surrounding parts (plastic connector) and not only from the air flowing around the sensor.

With the real outside temperature (for example 77F) the fuel trims are +/-2%, in traffic when the engine bay gets hot the IAT reads 120F and the fuel trims are around +10%

miami993c297

iTrader: (6)

Hi tici,

I chose a way in my IAT sensor positioning and it is close to work pretty well in many conditions in my application. Having an accurate VE Table is the key.

One parameter to consider deeply is the BIAS Table associate, mine is still the one you find in the stock bin file, I hope Marcin to come back with something soon.







Christian

jimmyblue

iTrader: (11)

What I would like to know, is what the IAT is meant to really
represent. This involves not only location but what kind of
mysterious fiddling the PCM does to the IAT data.

Is it meant to really be outside air? Air before the throttle
(same pressure, maybe warmer)? Air after the throttle (lower
or same pressure, expanded pressure as seen by MAP)?

The latter, inside-manifold temperature, seems the most
sensible for straightforward SD computation. But I'm not
inclined to assume GM did it simple & sensible. If they did
though, then this location (or similar) is ideal. Hope to know
someday.

SSpdDmon

iTrader: (10)

I'd argue that it's the temp of the AIR entering the filter. Based on it's factory location, I'm fairly confident in that answer. It's stuck in the lid to protect it from the elements. The problem is, we go to an aftermarket lid that moves the IAT closer to another heatsoaked component (the MAF). There's no longer a below connecting the MAF to the lid....it's a direct connection. I have moved my IAT to just above/behind the washer fluid tank and it's reading underhood temps entering the lid. Temp readings have dropped and are much quicker to react once the car is moving. My VE tuning efforts seem to be much more consistent....although not perfect (the charge temp blending table seems to be off still). Once I (or Marcin) determine(s) a solid way to dial in the charge temp blending table, it should help some more with consistency.



I wonder what would happen if I could permanently set the IAT reading to a certain number based on a resistor of some sort??? I'd like to see what happens as the outside air changes, but IAT stays the same. My problem is, as IAT's rise, my car actually leans out in the lower airflow regions. I wonder if keeping IAT constant would reverse the results?

RedHardSupra

iTrader: (1)

keeping IAT constant vs regulating IAT without really having a full physics based model is the same argument as the which clock is better, the one that's completely stuck (shows the right time twice a day) vs a clock that drifts off a second per day (it's not perfect but at least in the ballpark for quite a while).

I'd rather go with regulating it because we have at least a good idea of what the curve should be. having it stuck at some arbitrary value (even if it's the perfect average of effective IATs) it means it will be too hot half the time and too cold the other half.

my biggest concern right now is can we tune bias without tuning VE at the same time? afterall, bias affects airflow, airflow affects bias. also, a different bias curve would change the resulting VE numbers, which would effectively change airflow, which again will affect bias...and round and round we go.

currently i know how to _correctly_ alter VE (using BEN's or AFR%Error methods IS WRONG!) and i know how to adjust BIAS, separately. I'm working on combining the two, but that's some rather advanced math, and i'm missing about 20 math classes to that level.

My90Iroc

iTrader: (26)

I understand your logic. The problem is we don't know if gm took into account the heatsoak of components or increasing temp of incoming air. In later years (other models of course) GM moved the iat into that heat soaked maf. I'd have to believe they accounted for that change somehow. Maybe in the bias. Who's got an ls2 bias table to compare?

My other question has to do with what you said about the temp readings dropping faster when moving. That would be natural since the airflow from the car moving would take the heat off the sensor. Is that large mass of air moving over the sensor actually indicative of the "intake air?"
The intake air is either moving through the SS hood which has to be hot or coming from around the radiator.
I'm not trying to take away from your method or the time you've spent testing, I'm just curious. One day we'll get a handle on how the computer is doing its calculations.

SSpdDmon

iTrader: (10)

Airflow entering the lid either comes from behind the black, plastic grille (rain baffled by black styrofoam) or through the SS hood (if equiped) on the Camaros. The radiator airflow is blocked off by shrouds that you can see sitting behind the washer fluid tank. As soon as the car starts moving, cool air starts pushing under-hood, hot air back out of the cowl or under the car. I don't have the FRA mod done, so that's the only air entering my intake.

A simple test: take an LS1 Camaro out for a drive in the afternoon. After 20 mins or so of preferably stop and go traffic, stop and pop the hood. Feel the top of the lid where the IAT sits just in front of the MAF....and then, feel the black, plastic shroud just behind the washer fluid tank. It'll be a significant difference. The way mine is setup, the IAT is floating in mid air just above/behind the washer fluid tank. There's nothing near by that is heat soaked to throw off the temp reading. The result, more accurate IAT temps. 2 inches away from the only opening of the lid, I'd say it's about as accurate as you can get.

jimmyblue

iTrader: (11)

While the MAF body may be heat soaked, the thermistor
is stood off by its leads and the body has the benefit of
the concentrated air stream (unlike my Blackwing lid original
location, a backwater corner with minimal air velocity). I
see reasonable IATs even idling parked, other than a few
seconds after cranking. Talking about the 5-wire MAF here.

In my case there is only a few degrees' difference between
in the throat, and totally outside the car. But still I wonder
about things like, when you drop 60kPa across the throttle
blade, the MAP knows but is the temperature used, really
right (on the other side of the "expansion valve" from where
the pressure is being taken)?

You can have accurate temperature data anywhere you
locate the sensor, but very few of them will be what the
PCM is set up to expect, as far as relation to the MAP
quantity. That is my question, which of the positions
along the intale tract are consistent with the workings
of the airflow model?

RedHardSupra

iTrader: (1)

the stock one? i mean, that's what the factory have set it up for, right?

RedHardSupra

iTrader: (1)

the IAT itself might be more accurate, but that's not the temp SD calcs will use, unless you disable the complex temp blending model. in order for your 'improved' IAT to really work you gotta let the computer know what's going on. if the new placement of IAT really allows for the final TEMP to be much closer to IAT, then the bias curve would be closer to 0. however, by how much and where remains to be calculated during calibration.

tici

iTrader: (1)

Interesting answers!

Theoretically the PCM should use the IAT in the gas equation (pV=nRT).
A change in temperature means a proportional increase / decrease in the amount (mass) of air (and oxygen) sucked in by the engine.

I tricked the IAT using a potentiomenter and observed what happens with the fuel trims in SD: the PCM makes more or less the expected changes, but there are other variables. I'm sure the ECT is part of the calculation.
I think it's a more complex thing than just physics... probably measured values better than theory.

Just an example:
at 25°C (298K) I see 0% fuel trims correction
at 60°C (333K) I should see +10.5% fuel trims correction (298/333*100 = 89.5%)
This means the PCM is calculating a fuel need = 89.5% of what it would be at 25°C, which is wrong because it's tricked by the potentiometer, therefore the O2's are adding 10.5% fuel.
In reality the correction fluctuates between +7% and +10% depending on the ECT...

tici

iTrader: (1)

OK just to explain why I want to try SD:
my Vortech setup recirculates the unused air back to the blower inlet. This heats up the air A LOT!
Letting escape this air I cool the blower and the charge air is cooler too.
If I let the unused air escape the system I have to run SD or the MAF will report an huge amount of air to the PCM.
Even if I still have the stock 1 bar system the AFR at WOT is correct: don't ask me how but it works

This works well if cell #22 is around 0%, but as long the other fuel trims jump up and down depending on the IAT cell #22 will be a random number, that's why I'm trying to unserstand the relationship VE / IAT / ECT / and so on.

SSpdDmon

iTrader: (10)

My point is simple. You have an IAT and and ECT to work with. After the air enters the filter, it naturally becomes warmer as it passes over heat soaked components (lid, MAF, throttle body, intake manifold, and heads). The amount of heat that is absorbed in the process is related to airflow. The less time spent in a heatsoaked environment, the less heat absorbed. That's your charge temp blending settings right there.

Now, if you start with a bogus IAT, you'll find yourself chasing your tail when tuning the VE. There will be situations where the VE table is right (low airflow extended cruise on the highway) and situations where it's off (the same low airflow stop and go traffic) or vice versa. Simply put: garbage_in=garbage_out. If you start with an accurate IAT reading, then all that's left to determine is how much heat is actually absorbed for every 10grams/second, which can be manipulated using the blend settings. ECT should be relatively constant +/- 10*F. When you're only using a fraction of ECT, that small variance shouldn't be noticeable.

Again, this is related to N/A applications.

jimmyblue

iTrader: (11)

If you are going to do the speed density calc using IAT from
ambient air and MAP from expanded air, you would have to do
a gas-law correction for the discrepancy. PV=nRT is all about
-a- constrained sample.

So either GM corrects for the expansion across the throttle
(tricky, possible) in using MAP from one side and temp from
the other, or they just wing it and don't care because "the
MAF will get us past the EPA" (simple, and my suspicion).
Use the IAT as a "good enough" ambient sensor and don't
worry about what goes on at unsteady-MAP conditions?

SSpdDmon

iTrader: (10)

Another point.....at the track on Friday, my buddy logged a couple of passes for me. 1st, IAT in stock location. IAT's were 126*F leaving the tree and dropped to mid~90's by the time he let off the throttle after crossing the traps. For the 2nd pass, I gave him my IAT extension and sensor to relocate it to just above/behind the washer fluid tank. IAT's were high 80's leaving the tree and low 80's on the other end of the track. Ambient temp was roughly mid~upper 70's. He's now buying his own extension.

tici

iTrader: (1)

How accurate is the fuel metering for a stock SD system like TPI or the erlier LT1's?
We (LSx) are now playing with a system that uses SD for some particular situatons and it may be OK if it's not 100% accurate.

What about a system that uses it all the time?

My90Iroc

iTrader: (26)

That's interesting. What were his a/f logs like? What were his times?

I think an interesting experiment would be to put the iat sensor through the air filter so it's inside the air lid but not soaking up the heat from the lid.

BTW, where did you get the iat extension from?

SSpdDmon

iTrader: (10)

I didn't have the WB hooked up. However, being the tuner, I know it has a tendency to lean out slightly when the IAT's are inflated. Times were skewed as well because we were having 2>3 shift problems, which ended up being either a slipping trans or a VSS error due to the fact that I tuned it rolling on 25" tires and not the 26" slicks he had on it. If you could stick the IAT through the filter without letting other dirt particles in, I'd assume that would be a good spot as long as it's far enough way from anything retaining heat.

The extension came from www.casperselectronics.com.

ringram

Stock ECM blends IAT and ECT for charge temp as RHS says, IAT is not directly used, but only as part of the actual airtemp calculation along with ECT. If you move IAT then you need to relook at your charge blending table.

HD6.0l

While in SD watching AFRs with Wide Band unplug the IAT sensor , it goes rich , or disable IAT , ETC blending , goes rich .
So moving the IAT will cause a rich attitude in the tune and then you tune around it , did it really accomplish anything except causing a rich condition that needed to be calibrated out .