Vlad yours and Pats results testing heatsoak are different.

maybe the FFS heats the air all the time? I dunno.

As for your comment about leanness, thats cos they are fairly untuned and pat has gone too far the other way.

Andrew my point is that Pat is right in saying that if the air changes temp dramamtically from intercooler end tank to cylinder it would be better to compensate accurately and we have no idea how much it does change temperature.

Dann

But there's no way to compensate accurately for things like valve temperature (at least not with equipment that is available to consumers or probably even prosumers like you/I). If I had to guess, I would say that OEMs use computer modeling to guesstimate at that and then use cylinder pressure data to validate that as best they can. Nobody is suggesting there's a way to accurately compensate for things like cylinder head or valve temps with the bits we have available.

What Pat is suggesting is that there's an air temperature difference between the outlet of the IC and the air in the coldside pipe. What Joe and I are saying is that's absolutely ridiculous given the speed of the air and the orifice through which it's traveling.

This is one of those interesting situations in which you are 100% correct, and yet that knowledge is utterly useless.

When someone invents a technology which allows us to cheaply and accurately measure the temperature of the air inside the combustion chamber, after the inlet valves and yet prior to the ignition event, we can re-visit this conversation.

Until then, since we are stuck dealing with sensor technologies which are highly susceptible to measurement error induced by case temperature, and in which the magnitude of said error is greater than the temperature gain of the air as it travels 2-3 feet in a mostly non-thermally-conductive environment, the best we can do is the locate said sensor in an area which minimizes externally-induced heating of the case.

And, lest someone be led to believe that I'm writing this with anything even resembling a sour attitude, here's a kitten hugging a potato:

https://www.miataturbo.net/attachmen...ine=1458172071

Data and testing needed to prove your outrageous claim that the kitten is indeed hugging the potato

Fwiw I'm not knowledgeable in tuning and may have incorrect information. But thinking on a physics level...

To go along with what was said about where to measure IAT, where does it end, intake, vavles etc. Wouldn't you want the colder IAT reading to comp for extra fuel? Even just that smaller amount %?? With the amount of flow going into the intake, IAT is miniscule in difference when it's pressurized after the intercooler. Tiny amounts of changes, micromanagement tuning for accuracy? Why not put IAT in the 4 intakes to each cylinder and get an average?

Point being, does it matter.

It doesn't matter, you can tune out the intake manifold and valve heat because they should be near constant. As we all know, ideal gas law means nothing. If I try and correct air density like that I'll be rich as fuck at night and spitting stalling lean at lunchtime.

Put it outside the engine bay. My slow MS2 probably can't even shift gair in small enough percentages to matter, and I doubt a ms3 is much better.

When the air is really moving, on power, the sensor should react much much faster. I don't care enough to check, because gair doesn't need to shift hardly at all on my tune, I nerfed the hell out of "ideal gas law."

To put the bullshit to end finally I made a video.

Shitty production. But quality data. Temp skyrockets with external heat source. Even when its not being applied.


Data points @ :30, 1:10, and 2:48

Good for testing it and making a video. :) But your test is flawed if you think the sensor body alone is causing that 50*f temperature rise.

The engine was not running, so there was no airflow at all across the sensor. Thus it's just measuring whatever the air temperature of the air in the intercooler tank is. You heated the sensor/intercooler end tank with a heat gun multiple times. I would guess after heating the end tank of the intercooler with a heat gun that long, it was above ambient by a lot, and your sensor was showing that. Not surprised it showed a 50*F rise, that's completely reasonable and I would guess the air inside that tank was about 50*F hotter than ambient after heating that with a heat gun.

I run the stock AIT correction curve unchanged, and my car does not run rich or lean on hot restart, it runs at target.

I've already answered the heatsoaking it part. Test for yourself. Use a infrared temperature gun on a black surface for example to get a second reading. My intercooler pipe was black when I tested it, so I could measure the pipe temp, and air temp too. On a shutdown after restart condition, if I turn key on, engine off, and check the AIT temp vs pipe temp, they were the same. Cold start, same. etc.

EDIT: IR temp gun not heat gun....

At WOT, yeah there's negligible difference I agree 100%. Air's moving too fast to gain much heat from the walls of the pipe.

After a 15 minute heat soak from shutdown, you still maintain there is no difference? None? I pull off the highway and park at a store, come back 20 minutes later the air in that pipe isn't heatsoaked? Or the pipe itself? If I start the car and let it idle, will the cool air from the intercooler be heated any by the hot pipe?

You insist this does not happen, but your opinion violates simple thermodynamic laws.

You should test this yourself and post the data if you want to see what difference it makes. Put a sensor in the intercooler end tank, another in the plenum or near throttle body. Use an IR temp gun on a black surface (black painted, or put electrical tape over surface if not black to get accurate reading) and test and see. I can tell you from common sense that when you shutdown after driving it, the engine heat soaks everything in the engine bay.

Here's how I did it.

https://www.miataturbo.net/attachmen...ine=1458185493


Drilled tiny hole, installed thermcoulple, RTV to seal it, Zip ties to hold it/provide strain relief. Ran to 8CH EGT box.

https://www.miataturbo.net/attachmen...ine=1458185493

Alright. When my new battery comes in I'll do another test.

If you want a definitive sensor test, get a piece of sheet cardboard and poke a hole in it, stick the air temp sensor element through the hole so that the element is in free air and separated from the body of the sensor. Then put a fan blowing on the element, and a heat gun blowing on the body of the sensor. Idea test would be to have the fan running on the element for a couple minutes to let that stabilize at whatever the fans air temp is, then note that and hit it with a heat gun.

your stock ecu controls the ait curve?
or by stock you mean whatever came with the ms3-pro?

I never said it wasn't. Stop being deliberately obtuse.

I don't think it's heated substantially, no. I think I've stated that. Feel free to change my opinion with math/data, though. At idle, the air in that pipe is cycled three times a second. Can you quantify how much heat is transferred to the intake charge through that pipe (let's say it's a 2.5"x.080"x24" pipe) in that period of time?

Actually.... yes it does. :likecat:

But no that's not what I meant in that post. I run the default MS3 settings for AIT correction on the MS3.

Glad we can agree heat transfer is real and happening. I have no desire to change your opinion on what substantial means to you. I wrote down some temperature data last year, but to be honest it's probably gone by now and I'm not going to make it up on memory. Yes I could quantify the heat transfer via some 1D approximations using thermo 1 math, but I'm not typing all that out.

I run N/A so a few variables are removed. I get heat soak on my air temp sensor which is in a runner on a Skunk2 intake manifold. It will take 30 - 60 seconds of decent airflow over the sensor to drop the temps close to ambient like running on the track during a warm up lap... sitting idle won't change it much. Once running I see very little change in air temp throughout the run regardless of the temp of the motor (and also IM) and then when I pull in to the pits and airflow over the sensor is reduced, I see the temp rise fairly rapidly as the sensor heat soaks from a heated IM.

In my case on a 90 degree F day when the engine is already warm from a previous run, I'll see temps whilst sitting idle of around 140 degree F before dropping down to 100 degree F whilst I'm running. Pulling into the pits I'll see up to 160 F if I leave the car idle.

My sensor is an OEM Honda with a thin insulating gasket, 8" from the head in the runner. It can't be reading actual air temp because if that was the case, the air temp would drop immediately on WOT. Instead is slowly drops as the air flow over it cools the sensor.

This would be the same wherever the sensor is located... Runner, Plenum, intercooler.

do they?


I'd say it's like 10% and those are just NB owners and people who still put the AIT behind the radiator.

BEGi claimed they measured a 20°F temp drop between the turbo outlet and intercooler inlet with their aluminum pipes at 6psi on a 95°F day measured at a constant for 25sec done 12 times in a row.

If one eats a gallon of ice cream, then installs a silicone IC outlet pipe or SS or insulates his, he will barf and then never worry about heating the charge back up -- but the reheating couldn't be as much as the heat losses, since the outlet temps of the IC are much closer to the underhood temps.

I know I get lean on hot starts and there are a few threads dedicated to finding out the reason why this happens. NB owners seem to think it's the returnless fuel system while others credit bad AIT placement. I've got my AIT sensor in the IC endtank and see high 15 to low 17 AFRs on restart on my 1.6L car. It settles itself down if I let it idle for a bit or run it up to 2k rpms. I digress.

Great thread, quality information and knowledge expressed in here. :likecat:

did you attempt to tune your MAT corrections table or leave the default values?

I've played with it a bit but I am still taking data so that I can more accurately address the percentage correction. I don't want to derail this thread but I used a lot of the information in this thread for help. It may be a bit dated now though.

95% of the 10% is what I meant :giggle:

my advice is to zero out corrections from table completely and tune it last.

a quick look at datalogs can tell you a lot.

example, lets say you start the car and idle and your injector pw is 1.8ms at idle and you're lean. then after 60sec of idle, youre now sotich. did the pw change? did the ait change? what changed? stuff like that.

I'll give that a try. Thanks.

for returnless nb's: the fuel temp changed. so the AIT readings aren't really useful to log, however you artificially manipulate ASE corr and taper to get rid of the issue.

but anyways, back on topic lol

have kitty brain.

Engine running. Same result

Thankyou.

So you gotta get the sensor in direct cooling airlfow like the front of the IC.

this is why we say to get it away from the radiator.

and if youre running MS, zero out your AIT corrections. The default curve from B&G if youre not using a basemap is based on the Ideal Gas Law, problem is it doesnt take in account for things like shown in the video above. So it pulls fuel based on an AIT reading that isnt accurate.

Tune that curve based on what you actually need when you're actually driving the car.



what would happen if you did that same test, but held the throttle open at like 2-3K?

Have you been approached to help film the next Cloverfield movie? You got the shakey camera action down.

Also you might have an issue with oil temp readings if your gauge only goes between -1 and 1.

Oil Temp: See the latest video in my thread.

Brain: I'm running some updated table I found. To cancel out what they do.

I can do idle test mode and get ul around 2k. I'll try that again today.

aren't you on ms3? 100% across would cancel it out.


this is what I start with:


https://www.miataturbo.net/attachmen...ine=1458479554

Yep. That's the exact one I'm using.

Thank you. I already believed, but thank you.

What I really have to determine is if the WI nozzle will skew the results more if it is before or after the IAT sensor. The injection will obviously cool the charge and that should be measured, but the water or meth will probably damage the sensor and cause in an inaccurate reading. How do people do this?

it should be atomized by that point and not even be an issue. the GM sensor element can be submerged in water and not fail.

meth/water can drastically reduce intake temps and you really want to know, as close as possible, temps you're actually ingesting.

In my opinion the IAT sensor is meant to be a point of reference only, since obviously you can't read the charge swirling in the combustion chamber. I always have the same argument with one of my good engineer friends. I think that depending where between your IC and TB you mount it your VE table will look slightly different. Just as long as you don't mount it behind your radiator.
And hot starts are part of the reason there are so many start up features.
:2cents:

So to further a bit on the water/meth injection idea. Would you want to place the nozzle before the AIT sensor? For instance would placing it in the "hot side" of the intercooler endtank be a good place for it? I know some cars come OEM with intercooler sprayers like the EVO and some Japanese spec STIs but that is external and I've not seen people run it like I mentioned before. Usually it is always near the throttle body.

Those intercooler sprayers spray water on the intercooler. Not into it.

Yeah, I corrected my statement. You are just mad fast.

Knowing absolutely nothing about water injection, I'd expect you'd want the water nozzle as close to the outlet of the compressor as possible, give the water plenty of time to change phase and soak up heat.

And IAT always goes post intercooler, either endtank or plumbing directly after endtank, outside engine bay. As expected, Aidans test shows that in steady even air temps the case of the sensor will skew temps bad. There's nowhere else you can put it.

the water will provide biggest effect inside the combustion chamber
and I wouldn't want it cooling down the sensor anyway, rather be running too safe than artificially "safe"

Based upon studies done by NACA (the predecessor to NASA) during the 1940s, the optimum place for the water nozzle, in terms of combating knock due to insufficient octane, is inside the combustion chamber ala direct-injection.

This is hard to do.

I believe in the runners is the second best place. From my minimal water/meth injection research. Look up direct port meth injection.

The key point being simply that the water does most of its work inside the chamber during the compression cycle. Relatively little evaporative cooling occurs in the intake plumbing or the plenum.

You want as much of the water as possible to still be finely atomized when it enters the chamber. The further upstream you inject it, the more likely it is to pool out of atomization into larger droplets, especially if an intercooler is present.

yup, exactly where I was going with that

that's why I'm always a bit surprised (and skeptical) when people claim their AIT's magically plummet with w/i and no intercooler

Well the water probably cools the sensor...

and they think their charge is legitimately below ambient

My GM IAT is in the cold side of the intercooler and my water/meth injection is about 6" from the throttle body.

I'm sure water is probably ok to spray in front of the IAT sensor, but I have a feeling meth probably isn't that great of an idea.

When I'm spraying meth on the streets, I run a special tune for it. And when I'm running only water like at the track, I run a much more mild tune.

you dont e-cool?

how exactly would that work?

I d-cool

^He actually V-cools but he has a thick accent.

Evaporative cooling on the sensor element? Just spitballin here

I would be concerned about measuring the air as being warm and then having it actually be cooler and more dense once it is then cooled by the injection. This could inadvertently make the mixture more lean.

hmm, so basically it sucks either way.

got it :giggle:

unless it's pretty much repeatable and consistent. then the fuel map would just reflect it.

Again, there's very little cooling effect happening outside of the combustion chamber. I don't have the data in front of me, it's somewhere in a binder at home full of old photocopied test reports from WW2.

Any cooling (and thus, density increase) which happens inside the chamber during the compression cycle is irrelevant from the standpoint of fuel metering, as it's a sealed volume. More oxygen cannot enter past the closed intake valves.