GM IAT sensor / Easytherm
 

Hello guys,

I've been battling an ever-changing tune due to IAT temperature.

I finally took the time to do some empirical testing and it appears that at least in colder temps (around 28 to 50 degrees IAT), my MS thinks its around 12*F colder than it really is.

Due to this, I decided to remove the IAT from the car and test it as per DIY's website by immersing in cold water, medium temp, and boiling water. I took a few extra data points and came up with the attached graph. The higher values were nearly spot-on compared to the expected value, but there was some deviation in lower temps.

So my questions are:

1) does that look 'off' enough to create the issues i'm seeing?
2) if i were use easytherm to generate a new curve based on these measurements, does it matter that my lowest temp is 19*F vs the -4*F used in the provided .inc file?
3) has anyone else used the standard GM IAT sensor but re-easythermed for accuracy?

look about half way down the first post, which settings did you use?

https://www.miataturbo.net/forum/t7000/

I dont mean to sound dense, but which which settings?

I got the MS from braineack, he burned the HiRes10G firmware onto it, which is also the project i open when running megatune.

I think you may be asking about "Step 3", in which case i guess braineack chose option 2 (gm iat, mazda clt). He did not change the CLT range though as suggested there.

I have the same problem. When I program the IAT with the diy settings in tunerstudio it goes from being spot on with the clt sensor to reading about 15 degrees cold. When it is warm outside it seems to read correctly. I've thought about leaving the default tunerstudio settings for the IAT, but I'm afraid it might not be correct when the temp warms up.

Let me know if these IAT settings fix your problem. I'll try it tomorrow

I haven't tried to re-burn yet, I'm waiting to hear from anyone else that's tried this and had some success.

damnit. i tried to flash the MS but i see now that i screwed it up.. I created new INC files using easytherm and then just hit the download to MS button... it downloaded 350 lines then quit..

UHOH now I can't seem to get it to respond on the COM port, it looks pretty dead.

Now that the original hires firmware is gone, do i have to get in there and set that boot jumper?? ( i really dont want to, its buried w/the stock ECU and its cold outside)

i got the MS apart, shorted the boot jumper, and was able to re-flash the original firmware i had on there.

Then i got adventurous after reading about another person who pretty much made the same exact mistake as me :) (not miata though..) I selected *custom*, not hires10g in easytherm, and was able to generate a new S19 file combining the original firmware with my new sensor values.

I burned it to the MS and, ............ it's no better and maybe worse than before so i'm a little at a loss, as i used the values i measured by hand :(

Anyway, i'll try running the car tomorrow and repeat my tests to see what's going on.

-Tomaj

P.S. i think i'm going to get one of those IR thermometers .. any thoughts?

1. no.
2. no.
3. no.

In that case, my only other thought is that i should measure the resistance at the megasquirt vs. at the sensor, to see if the wiring is introducing any additoinal appreciable resistance.

Does that sound reasonable?

I doubt it is; the bias resistors for the AIT and CLT only have a +/- error value of 1%, BUT, you never know how much that also offsets the curve, so a true reading might help.

did i build you a parallel or standalone harness?

You built it as parallel.

Regarding the bias resistor, what would the effect be if i had the wrong value in easytherm? I used the value you had provided in your .inc (and also confirmed in another thread), but is it possible that you put a different physical resistor in?

in the same vein, is it easy to find this resistor on the board and do a visual check?

-tomaj

Remove both r4 and r7.

Redo the easytherm firmware using 2625 for the CLT bias and 3475 for the AIT bias.

They are the two blue resistors, cant miss them.

I took off the two blue resistors (identical) from r4 and r7.

i reflashed with the bias resistor values you supplied and the result is pretty much that nothing changed..

my car hasn't been on in a day and a half and shows: CLT 31/32 and IAT shows 24*F. Local weather is reported at 33 on google.com

-Tomaj

did you copy the new .inc files into your project folder's mtcfg folder?

i let easytherm do it, verified by checking file mod date in the folder.

FWIW i checked the resistance with the sensor in the car, it read around 9500 ohms at the current temperature. On the 'official' resistance curve chart i found, it lists 32 degrees as 9450, so it seems completely accurate up to the AFM harness.

I will next try to figure out which inputs on the MS harness are for the IAT and test the resistance on the DB37 to see if there's some issue there.

My current thinking is that the easytherm et al was correct all along but i have a wiring problem leading to increased resistance... a bad solder joint at the harness? ( i had to fuck with my harness a bit a while ago)

-Tomaj

i tested the resistance at the AFM harness as stated before, to be correct... however as soon as i plug into the AFM harness, and test the resistance again (either there or at the Db37 connector) i get around 2625 ohm... is that a coincidence that this number is the same as the bias resistor value you mentioned for my CLT sensor?

i'm not really sure what's supposed to happen once i hook into the AFM harness, as i guess at that point it's sharing the IAT sensor value with the stock ECU??

well hardy harr harr (in a pirate voice... not sure why, i guess i'm just excited)

on a hunch i figured that 2625 wasn't just a fluke so i plugged that into easytherm as the bias resistor value for the IAT sensor.

i flashed, and........ drumroll please....... 100% correct!!!!

Brain, Joe, anyone -- please confirm if this statement is correct:
You can get your bias resistor value (if running parallel? and with no r4 / r7?) by unplugging the MS and testing the resistance across the IAT at the MAF or the DB37.



Forgot to add -- thanks for the help Braineack!

That didn't work for me. To get the values I gave you took a long process of trial and error a few years back when we were first figuring out the values.

I know for absolute certainty that using 1280 for CLT and 1450 for AIT for the bias with R4 and R7 installed on a parallel install gives you accurate values. My CLT and AIT values would be within 1-2°F with the car sitting overnight on a 20*F morning or 80*F afternoon.

Therefore, if you do the math for parallel resistors and you know the value for the one 2490 and need to find X, it's simple algebra.

solve for x:
1280 = 1 / (1/2490 + 1/x)

a:
x = 2625


Regarding fining the resistance with a multimeter:

When I took the resistance reading of the stock ECU plugged into a boomslang harness and read the resistance between ground and pin 20 (AIT) I got a value of 3630.

http://boostedmiata.com/MS/random/bias_1600_ecu.jpg

If I plugged in a MS with R4 installed, the value now read 1840ohms.

http://boostedmiata.com/MS/random/bias_ms_and_ecu.jpg

Using those numbers gave me no where close to accurate values. The numbers I gave you do and are what I personally use.

I'll ask again, once you flashed your new easythermed .s19 file, did you move the .inc files that easytherm created, based on these new bias values, into your /mtcfg (MT) or /inc (TS) folder?


What you can do, without having to flash the MS over and over, is a somewhat easy process. Open easytherm and create a new .s19 and .inc files. Changing the bias of the incorrect reading up/down 200-300 ohms at a time. Each time you create a new file, copy the .inc files into your /mtcfg folder. Now turn the car to ON and read the values with your laptop. There is no need to flash the MS as your laptop is converting the resistance values straight off your .inc files not the MS unit. Once you come up with a bias that gets the AIT and CLT to read alike, you can then finally flash the MS with that .s19 based off those values and the MS unit will read the values your happy with.

Scott,

I will try to digest the content of your e-mail tonight.. Regarding your question of moving the inc files, i just verified that they are in the mtcfg subfolder of the project that I opened with megatune.

-Tomaj

and it says:

My matfactor.inc in the hires/mtcfg directory says:

; Generated using EasyTherm4.exe,
;
; Air Temp Thermistor Comment Field
;
; ***** CAUTION - NON STD BIAS RESISTOR REQUIRED *****
; File generated for use with 2625 ohm resistor at R4
;
; Steinhart-Hart coefficients: A= 1.505975E-03 B=2.249965E-04 C= 1.247472E-07
;
; Input Data: Temp, degF Resistance
; -4 28681
; 86 2238
; 210 177

FWIW i set the bias resistor value back to the value you recommended earlier (36XX ohms, too lazy to check atm). The temp came up to at 38 degrees vs 60 degrees with the 2625 bias setting.

I just confirmed again that the resistance at the sensor was correct (4640 ohms at 60 degrees), so i am satisfied (kinda).

The only remaining bothersome problem is that when i try my "advanced->coolant related air density" trick, it still appears that the MS somehow thinks it's 5 degrees off, but that's still a lot better than being 17 degrees off as it was before the reflash..

Scott -- can you try this for me??
1) get the current MAT
2) in advanced-> coolant related air density
3) set one of your temp values to one degree less than MAT noted in step 1, set value to 100%
4) set next temp to MAT, set value to 50 %
5) set next temp to MAT+1 degrees for example, value = 50%
6) set next temp to MAT+2 degrees, value = 100%
(this is to create a vertically sloped area of 50% enrichment)

If your MS is correct, your car should probably stall when u burn that.. right? For me, it doesn't.. I have to shift my temp values by 5 degrees to get it to stall out, which to me indicates the MS thinks its a different temp than is displayed in MT.

-tomaj

P.S. -- is there any way to get the AIRDENFACTOR ADC value as seen by the MS?

oh the numbers in that table have never been correct for me. i think they are off by 10*F or so.

I have my table zeroed off until 95*F where I add in some fuel to battle heat soak.

lol that never bothered you? :giggle:

Then again you're over in florida so you probably stay in the flat part of the enrichment curve and wouldn't see the same negative effects that I do.

you mean VA. i moved my AIT sensor out of the engine bay so i dont get stupid heat soak issues and dont even really need the corrections table.

Hey, I always put you in FL for some reason.. Anyways, regarding the iat correction table, even tho you are not using it, the underlying problem is stil there no?

IOW the temp being off is just a symptom, the problem being that the Ms. Rhinks it's a different temp

okay, well, I tired those numbers out...the ones I gave you without R4 and R7 and I had issues with the AIT reading about 20*F too low myself. I'll have new numbers to try tomorrow for you.

lol the numbers i used seem to be working ok.. it was off by 2 degrees today when i first got in the car, but don't let that stop you :) it would be great ot have a real fix.

-tomaj

what did you end up using so I can try them out today?

2625 for both. Seems OK but can't be sure until we see larger overnight temp changes.

tomaj

I'll probably check in the morning. I have my MSI united plugged in now that I have had the same settings for 4 years. I know the temps are accurate, it's always right around ambient anytime I check overnight and I've tested the clt temps versus a thermometer in the water and versus the thermoswitch.

my MSII unit has r4 and r7 removed, but my laptop is failing on me hardcore with it's battery and I haven't had a chance to go make sure the numbers match yet.

I got the same results as you. There must be a second pull-up on the AIT circuit in the stock ECU that throws off the formula for just two resistors. Using 2625 for the CLT and 2750 for the AIT got me 35*F for both sensors this morning. My MS-I unit had 36*F CLT and 34*F AIT.

I might reinstall R4 and redo the value because I notice the CEL for the MAF was staying on in this configuration. I think the GM sensor is too far outside the OEM range and it's confusing the ECU, with the bias resistor in MS, it goes to a more happy place. I get that CEL every now and again on my MSI with R7 and R4, but had it my entire drive this morning on my MSII unit without R4 or R7. I'll test that theory out tomorrow.

any luck?

kept it as is

that being 2625 and 2750?

actually 2625 for both. the temps are damn close from what I can gander. it was 31*F out when I went out to my car, the CLT showed 33*F and the AIT 29*F. It's close enough and with MS-II I swear they display changes that are more minute and update faster on the screen (probably cause the scale is larger than 255 lines), so a small % error in the resistance scales might be showing up.

I wish I had my thermoswitch still installed so I could test the CLT sensor at 207*F when the fans trigger, but I know my stock gauge creeps to 12 o' clock at 218*F and that seemed to be spot on when I tested that over the weekend.

It's pretty close to the ballpark, and makes sense that the values match for both, those with 94-99s found theirs were both 1450ohms. So I'm just going to stick with it. CEL never came back, so i dunno what the deal with that was the first day.

I might test both sensors in water and see if they are off a bit from the published scaling, but that requires effort. I did my CLT sensor a while back and wrote it down somewhere. But I think I remember it being very close to what the service manual stats.