When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
The MAT compensation tables for the PW calculation in the $8D mask has misleading titles. The purpose of the two 'MAT' tables is to provide an inverse temperature term for the PW calculation.
Colder air has a greater density. This needs to be accounted for in the PW calculation. The inverse temperature term provided by the 'MAT Compensation' routines is based on both CTS and MAT.
The amount of CTS vs. MAT that is used can be varied. One table is a model of the intake tract heating of the incoming air. IOW, how hot does the air become just from its travels through the intake. The engine coolant temperature affects this.
When the MAT is relocated and becomes an IAT sensor the table calibration is no longer correct.
When the intake design is changed (say to a LT1 or single plane manifold) the table calibration is no longer correct.
There is also an option bit involved. At location $018, b4 is used to select either the intake tract heating model table, or to just use the MAT/IAT as is. If this bit is set to a 1, then the intake tract model table is used. Else, if set to a 0, then the MAT/IAT is used as is.
Using a MAT as is would probably work OK. Using an IAT as is would not. There is going to be some heating of the air from the intake tract. For a decent tune this needs to be accounted for.
Here is the Intake Tract Heating Model table (with the original title):
Note the table is based on airflow. The less time the air is within the manifold the less heat it will absorb from the intake tract.
This table is used to combine the MAT/IAT sensor value with the CTS value. Here is the math:
T = ((CTS - MAT/IAT) * MULT) / 128 + MAT/IAT
What is being done is a portion of the CTS value is being added to the MAT/IAT value. This provides a new temperature term (T). In simpler words: take the difference between the air temperature and the coolant temperature, take a percentage of this difference and add it back to the air temperature.
The table value for 0 gms/sec airflow is 128 (MULT), in affect the T term is going to be the same as the coolant temperature (CTS).
As the airflow increases less of the coolant value (CTS) is added to the air value (MAT/IAT).
Here is an example at 64 gms/sec, 10°c MAT/IAT and 85°C CTS. From the table the MULT term is 27 at 64 gms/sec:
T = ((85 - 10) * 27) / 128 + 10 = 26°c
The same temperatures at 128 gms/sec gives this:
T = ((85 - 10) * 16) / 128 + 10 = 19°c
The MULT term is changed to 16 (from the table) and the final temperature term (T) is lower at 19°c.
What is interesting about the above examples is that a MAT sensor with a engine at 85°c is probably not going to be at 10°c. Well, maybe if the ambient air is at -50°c. That is where the stock calibration falls apart with a relocated MAT to an IAT.
A MAT would be about (SWAG here) 50°c when the engine is up to temperature. Lets run the first example with that:
T = ((85 - 50) * 27) / 128 + 50 = 57°c
Quite a difference from the T = 26°c of the first example. In this case relocating the MAT will cause the engine to run richer and the BLMs to drop. The table will need to be recalibrated.
The last table is easy. The values in this table shouldn't really need to be changed. Here it is:
Easy enough, use the temperature term from the previous calculation and look up the inverse temperature term (~T) to be used in the PW calculation.
The column labled 'MAT Deg c' is really the T term. Which can and should be a combination of the MAT/IAT and CTS values.
In closing, as others have stated the Intake Tract Heating Model table is the key to recalibration. With a better understanding of how the table works it should be easier to do a better tune.
My thoughts on the stock calibration shows that at low air flows the stock long tube TPI setup adds a lot of heat. This can be seen by the full CTS temperature term at 0 gms/sec
Even at higher air flows with a heat soaked MAT, the Intake Tract Heating Model will move the T term to a high value.
An intake with shorter runners will heat the air less then a TPI. There is simply less surface area and less time for the air to pick up heat as it travels through the manifold.
RBob, I will try and dig up a post I made on this subject a few years ago when I discovered that the MAT Inverse Lookup Table needed to be modified when I relocated my MAT. In fact, with a LOT of trial and error, I even came up with a revised MAT table that people could use to correct for a relocated MAT sensor.
I always mentioned that I had "no clue" how the table actually functioned and that my modifications were based on "using my car as a test bench" in for around 6 months of driving.
Given your information, I will deeply analyze it and try to use the revised MAT table I derived for a relocated MAT sensor to see if I can finally determine who my revision affected the calculation and why it works.
So for all of us out there like me who are just beginning to grasp this stuff......the delta mult. table is used to calculate how much hotter the air in the cylinder is, than the air that the MAT is reading, and the comp. counts table pulls fuel based on this.
In other words, CTS and MAT are plugged into the equation (which is referenced to airflow in the mult. table) to find a calculated "cylinder air temp," then that value is referenced in the comp. counts table to see how many counts to pull from the injector pulse.
Originally posted by TheGreatJ So for all of us out there like me who are just beginning to grasp this stuff......the delta mult. table is used to calculate how much hotter the air in the cylinder is, than the air that the MAT is reading, and the comp. counts table pulls fuel based on this.
Correct.
In other words, CTS and MAT are plugged into the equation (which is referenced to airflow in the mult. table) to find a calculated "cylinder air temp," then that value is referenced in the comp. counts table to see how many counts to pull from the injector pulse.
Right????
Correct up 'til the 'counts to pull'. The ~T term is used in the calculation of the mass of air in the cylinder. It compensates for the air density change due to temperature.
That portion of the PW calc, also using VE% and MAP, spits out the airflow (gms/sec) and the mass of air in one cylinder (gms/cylinder).
From there the code uses the commanded FAR and injector flow to calculate a PW (from the gms/cylinder term).
To take this one step further, you mentioned "cylinder air temp". In reality, this is the end goal. However, it isn't really what is being calculated. In order to calculate the cylinder air temperature additional parameters would need to be taken into account. Both frictional heating as the air passes the valve and heat picked up from the cylinder and head walls would need to be modeled.
By some degree (ohh, bad pun), these additional heating factors would then be offset by the cooling from the evaporation of the fuel. If the heating and cooling from these items is steady state, IOW, it happens pretty much the same each time, every time, then there is no need to model them.
It can be compensated for by the VE table. Whereas a difference in ambient temperature can not.
Assuming you unselect $0018, b4, does the 'Temperature Compensation vs. CTS - MAT' become the only table involved in PW compensation? If that is the case, there would be no T term, so is the temperature colum the MAT reading as given by the MAT sensor or CTS minus MAT like the title says? When using the heating tract model, I don't understand where the GM/sec value is coming from, how do we as the tuner know what multiplier to reference, all I gather from the examples is 64 and 128 were pulled at random. Why does everyone say there is no need to alter the 'Temperature Compensation vs. CTS - MAT' table when using the heating tract model? I always had it in my head that is the engine was tuned under different MAT conditions from the factory calibration this would need to be changed due to the different baseline. I have personally never touched either of the two tables because I never grasped the concept of tuning them, and in turn, have never had a consistant tune. I would finally like to nail this down.
I love it when guys want to KNOW what stuff does BEFORE changing them. That is exactly how it's supposed to be. Enough of "giving it what it wants." I think the faster way is "know what you're giving it, see how it likes it" "is that OK?" Iterate a few times for "what it likes BEST" and "WHY!!!!!!!"
Bracing.... But heck, it's true, but maybe stated a lil better. The important part is "know" and "why."
If I had a choice:
Enrichment vs ECT vs MAP.
Enrichment vs IAT vs MAP.
That takes a little more visual/analytical interpretation to avoid "what ifs."
I currently make do with
Enrichment vs ECT vs MAP
and
Enrichment vs IAT.
Seems NOOOO problem at all on this lil pos engine that passes EURO3 and starts and drives absolutely fine at -40C. I'm sure a hotrod has much less requirement.
oops, had a phone call, hope that didn't lock up the thread for that 1.5hrs.
With $0018, b4 unselected the T term is whatever the MAT/IAT value is:
T = MAT/IAT
For the air flow term it is RAM location $6B. Will need to patch the ALDL stream to include it. I checked and am surprised it was not already there. Then it is a matter of matching up that value with the table. This is a value that the ECM calculates internally.
As for the 'Temperature Compensation vs. CTS - MAT' table, I can't speak for everyone. I stated that 'the values in this table shouldn't really need to be changed.' Which if you find that large changes are being made, then something else may not be correct.
Awesome, I think I finally understand now that I have the Gm/S calculation. I got rid of PROM ID in the ALDL stream so Datamaster will display the ECM's airflow calc as well as Tunerpro. I plan on trying both using the intake tract heating model and not using it and will ultimately use whatever I can get most consistant through the ambient temp range here in NW Indiana which is 0-100F.
I love it when guys want to KNOW what stuff does BEFORE changing them.
That is exactly how it's supposed to be.
Enough of "giving it what it wants." I think the faster way is "know what you're giving it, see how it likes it" "is that OK?" Iterate a few times for "what it likes BEST" and "WHY!!!!!!!"
Ya, life is neat if there's folks that can answer everyone's questions, but if folks don't do anything because there's no one to answer their questions, then it's all been academic.
Maybe to you, but, then again, it's just your opinion.
Are you serious?.
Your playing word/ ego games.
If you want to play on words, then you need to add the comment about taking notes, and looking for trends.
*Why* would in your *crusade*, also mean understanding ALL of the engine dynamics, and ALL the interactions of the air/ fuel as it's injested, and reacted. There are folks that are absolute geniuses that have spent life times on trying to get a handle on what you've treated as a nothing more then a witty saying.
Let me try making it really simple for you, the statement about making an engine happy, means having the tuner put his wants/ thoughts/ ego, to the side, and objectively look at what's making a given engine *happy*. In making the engine happy, one has to have a clue about what's going on. having a total understanding of Obert, Glass, Heywood, Ricardo, is secondary to understanding what's making the engine happy. Happy isn't guessing either, it's about testing/ experimenting, and analysing the results. I hope this is clear enough for you to understand what was meant by keeping the engine happy, since you obviously missed it the first time around.
Not to argue, and I didn't miss the point, but making it happy can lead to ***** houses and bars. If what you're giving it looks really messed up, maybe some piece of the puzzle is drunk and passed out.
There's a rubik's cube like puzzle going on in each ECU, because these are currently running on tables and simple algorithms, hardly models. Sure, there are ways to get it to be happy under a simple set of criteria, but once you throw strange driving disturbances, or even normal driving circumstances at it, one can think "darn, that change that was happy under that test completely messes up this situation" because one didn't know what was really supposed to be changed to fix X problem, and maybe the stock code CAN'T fix X problem.
I think people should learn WHAT to change and WHY, BEFORE changing it. Yes this requires notes, research, etc... Sorry I didn't make a 100% complete argument, nor did I intend to.
I think some insight to the madness is more important to most than iteration after iteration that waste gas and sanity. Someone had to do the iterations, but not everyone. This is a place to share.
Rbob, does this $0018 correspond to a table/switch in CATS? Is it the same thing as the switch that turns on/off the INV MAT table that is presently available in CATS?
I'm running a Miniram with the sensor mounted in the air ducting. But if I'm running the wrong MAT PW mode, then I'd like to change it if I could.
Rbob, does this $0018 correspond to a table/switch in CATS? Is it the same thing as the switch that turns on/off the INV MAT table that is presently available in CATS?
I'm running a Miniram with the sensor mounted in the air ducting. But if I'm running the wrong MAT PW mode, then I'd like to change it if I could.
The location $0018 is an option word:
There is also an option bit involved. At location $018, b4 is used to select either the intake tract heating model table, or to just use the MAT/IAT as is. If this bit is set to a 1, then the intake tract model table is used. Else, if set to a 0, then the MAT/IAT is used as is.
It is option word 3. It may be listed as such in TunerCAT, not sure. It is then bit 4 within that option. These may be broken out separately.
I would run with that bit set, to use both the CTS & IAT blended together. This will give the best result.
Note the table is based on airflow. The less time the air is within the manifold the less heat it will absorb from the intake tract.
This table is used to combine the MAT/IAT sensor value with the CTS value. Here is the math:
T = ((CTS - MAT/IAT) * MULT) / 128 + MAT/IAT
What is being done is a portion of the CTS value is being added to the MAT/IAT value. This provides a new temperature term (T). In simpler words: take the difference between the air temperature and the coolant temperature, take a percentage of this difference and add it back to the air temperature.
The table value for 0 gms/sec airflow is 128 (MULT), in affect the T term is going to be the same as the coolant temperature (CTS).
As the airflow increases less of the coolant value (CTS) is added to the air value (MAT/IAT).
Here is an example at 64 gms/sec, 10°c MAT/IAT and 85°C CTS. From the table the MULT term is 27 at 64 gms/sec:
T = ((85 - 10) * 27) / 128 + 10 = 26°c
The same temperatures at 128 gms/sec gives this:
T = ((85 - 10) * 16) / 128 + 10 = 19°c
The MULT term is changed to 16 (from the table) and the final temperature term (T) is lower at 19°c.
What is interesting about the above examples is that a MAT sensor with a engine at 85°c is probably not going to be at 10°c. Well, maybe if the ambient air is at -50°c. That is where the stock calibration falls apart with a relocated MAT to an IAT.
A MAT would be about (SWAG here) 50°c when the engine is up to temperature. Lets run the first example with that:
T = ((85 - 50) * 27) / 128 + 50 = 57°c
Quite a difference from the T = 26°c of the first example. In this case relocating the MAT will cause the engine to run richer and the BLMs to drop. The table will need to be recalibrated.
The last table is easy. The values in this table shouldn't really need to be changed. Here it is:
Easy enough, use the temperature term from the previous calculation and look up the inverse temperature term (~T) to be used in the PW calculation.
The column labled 'MAT Deg c' is really the T term. Which can and should be a combination of the MAT/IAT and CTS values.
In closing, as others have stated the Intake Tract Heating Model table is the key to recalibration. With a better understanding of how the table works it should be easier to do a better tune.
My thoughts on the stock calibration shows that at low air flows the stock long tube TPI setup adds a lot of heat. This can be seen by the full CTS temperature term at 0 gms/sec
Even at higher air flows with a heat soaked MAT, the Intake Tract Heating Model will move the T term to a high value.
An intake with shorter runners will heat the air less then a TPI. There is simply less surface area and less time for the air to pick up heat as it travels through the manifold.
RBob.
i'm struggling with air temp hard core... so i was wondering what .bin you used to get the numbers in the code box above? i've got about 4 different bins i've downloaded from places and all of them have different numbers in the grams of air flow table than yours, but the same as eachother.
i'll attach what mine look like, my intake is pretty much stock, just a little porting and bigger runners. so i wouldn't expect the temp to be so off so much but after a heat soak restart(say 50*c) the A/F is almost 2 points learner than it is at say 30*c. takes a while for the intake temp to return to 30*(which is what i try to turn at) i've tried changing the MAT Compensation Counts Vs. MAT table in that temp rang but as your post says, it doesn't do much if anything. and i don't want to go nuts in there covering for another prob.
so here's my table, the b4 bit is checked so it should be using coolent as well. which doesn't very much at all on this car. it has a 180 stat and runs about that all the time. even after a heat soak restart it returns to 180 very quickly.
oh and this is a manual .bin, incase that makes a difference.
i'm struggling with air temp hard core... so i was wondering what .bin you used to get the numbers in the code box above? i've got about 4 different bins i've downloaded from places and all of them have different numbers in the grams of air flow table than yours, but the same as each other.
AXCN, '91 y-body with 6-speed. The differences are not that large, be sure to change the BIN value to the multiplier value as shown by TP.
AXCN, '91 y-body with 6-speed. The differences are not that large, be sure to change the BIN value to the multiplier value as shown by TP.
RBob.
your talking about this right.. making the air flow show up in my logs.
For the air flow term it is RAM location $6B. Will need to patch the ALDL stream to include it. I checked and am surprised it was not already there. Then it is a matter of matching up that value with the table. This is a value that the ECM calculates internally.
yeah.. i don't know how to do that lol. i'll search for adding things to my aldl logs and see if it makes sense.
my question is, and it since i'm struggling with this so hard. the question my not even make sense. but if its a correction vr air flow, should i leave the slope alone since it seems to based on temp, not air flow. meaning it doesn't go lean at higher rpm where it would flow more. it goes lean everywhere till the high temps drop back down.
RBob, I will try and dig up a post I made on this subject a few years ago when I discovered that the MAT Inverse Lookup Table needed to be modified when I relocated my MAT. In fact, with a LOT of trial and error, I even came up with a revised MAT table that people could use to correct for a relocated MAT sensor.
I always mentioned that I had "no clue" how the table actually functioned and that my modifications were based on "using my car as a test bench" in for around 6 months of driving.
Given your information, I will deeply analyze it and try to use the revised MAT table I derived for a relocated MAT sensor to see if I can finally determine who my revision affected the calculation and why it works.
Can you send me the revised MAT table you created?
Thanks to PP, this thread got revived and I picked up on it. Although I am working with an LT-5 it is an SD motor. Myself and many ZR-1 drivers DO move the MAT from the airhorn of the LT-5 to somewhere closer to the cold air intake area. Mine is behind the air filter which is sucking air through the radiator shroud. Most, if not all of us, don't do anything about the tables.
However, the only tables I have been able to find in the .xdf deal w
Density v Baro and gm/sec, and a MAT Compensation v gms/cyl. I have attached a screen shot from TPRT. I am trying to correlate these tables with the ones discussed in this thread. The values in the tables are stock.
AXCN, '91 y-body with 6-speed. The differences are not that large, be sure to change the BIN value to the multiplier value as shown by TP.
RBob.
ok, so are you saying all i need to do is use the revised MAT table by traxion and everything should work ok? I have a relocated sensor in the airbox now. Let me know.
ok, so are you saying all i need to do is use the revised MAT table by traxion and everything should work ok? I have a relocated sensor in the airbox now. Let me know.
I don't know, never tried Traxion's table values. There is also Grim Reapers values someplace too.
I'd like to try the GR or Traxion values now that my MAT is positioned as an IAT but I'm having difficulty translating the table(s) in my $D0A mask with the tables
displayed in this thread. Any help interpreting them would be most appreciated.
I attached a copy of my table in a previous post.
Just an update. I decided to play around w the MAT tables especially since the MAT is now an IAT. All I can say is WOW!. I have a WBO2 on the car, and the AFR is so much more consistent. The AFR doesn't vary as much as before. At times, it appears to just sit there at 14.7 but then it will swing although the variation is much narrower. Idle quality is smoother also. The change was very slight. A couple of % points. Basically, I thought that the MAT values needed to be moved further down the C temp scale since cooler incoming air would correspond to a warmer temp at the manifold inlet. I guessed right!
When the intake design is changed (say to a LT1 or single plane manifold) the table calibration is no longer correct.
RBob.
So I had contacted John at Tunercat a while back and he indicated this option word location corresponds to the Enable Inv MAT compensation counts table calibration switch.
I'm confused now. I just went and looked at the $DA3 LT1's (BDLD M6 and BDZL A4) and they're both showing the Inv lookup table enabled. I know they don't have the IAT sensor in the intake manifold.
Their tables are different than the $8D tables though (and I threw in Glenn's table for kicks).
So if the LT1's are also running the Inv look up table enabled and they're clearly in the IAT category, when exactly would you want to deactivate that table?
So I took the car out out for a run yesterday. I had put in the $DA3 LT1 MAT values and it made a noticeable improvement in throttle response and overall pep. Definitely made the conversion worth the effort!
But the curious thing were the temperature readouts themselves and it got me to thinking...
I have an ambient temperature readout in my rear view mirror (see below)
The sensor for this is located right in front of the air filter under the front bumper. Before the car even starts up, there's a built-in ~4°F delta between the mirror temperature and the scanner readout of the MAT sensor. Attributable to simple variability in the two devices. Though keeping the built-in difference in mind....
So when I'm driving 75 mph on the freeway, I would expect no more than that same 4°F difference, yet I get a ~12°F difference given the amount of air flow passing over the two sensors. That got me to fundamentally thinking more thermally... particularly self-heating effects (given RBob's explanation in the opening of this thread)
The plastic body MAT sensor is sitting in a rubber grommet, in a rubber air tube elbow, which constitutes a very lousy thermal path for self-heating. Intuitively it should be very well insulated against self-heating. Despite that, we know the sensor element will still pick up some small amount of self-heating from it's surroundings. However, with significant air flow it should be able to lose heat to the air stream faster than it's self heating, which means it should cool down to the air stream temperature. But yet, I'm still getting that significant temperature difference, which implies significant self heating. So why would that be?
So I started thinking about what else could be putting heat into the sensor element and then it dawned on me: the freaking wire harness. In my setup the wire harness runs from the firewall, under the intake manifold, and drops down in front of the fan. It's a long run of copper wire with a lot of surface area to pick up heat (which there is plenty of). Since copper is a superb conductor of heat, it's going to very easily conduct that heat to the nearest heatsink. With heat always flowing from hot to cold, and the air stream cooling off the sensor end of the wire via the air cooling off the sensor element itself, my guess is that the heat in the rest of the wire is seeing the cooler end and flowing to the sensor element, thereby causing self-heating from the inside out.
My next step is to re-route the wire harness off the engine and see if I can get rid of that temperature delta. It may not seem like a lot, but I'm also thinking that when driving around town, with less air flow, the difference is going to be significantly greater (and much more difficult, if not impossible, to quantify), which would ultimately prevent me from taking full advantage of the colder air coming in.
BTW, I also thought about electrical resistance vs temperature, but that appears to be only a very minor contributor...
For copper... R(T) = .1 * ( 1 + 0.0039 (T - 20) ) where T is the measured wire temperature in °C and 20 is the baseline temperature 20°C , above which every degree results in a .39% increase in resistance, and the .1 ohm is what I measure at room temp
So for 180°F (82°C) steady state engine temp (assuming the wire reaches that temperature)
R(T) = .1 * (1 + .0039*(82-20)) = .124 or, in other words, a .024 ohm actual increase in resistance on the circuit. Negligible in my opinion.
ULTM8Z,
So I'm curious if your BLM's are 128 after changing to the BDLD or BDZL table values?
Did you have to retune your VE's if not...
just looking for an update overall...
Wow... ressurrecting an old thread! I do have 128 pretty much every where now. Yes, I did have to retune somewhat to get that though. Changing to the BDZL tables altered the fuel delivery enough to require changing the VE tables.
It was worth it though. I remember seeing a noticeable improvement in power and driveability.
That's great news... I'm going to try it. Do you think after getting the VE tables right BLM's were good with large temp changes. Example at 45 degress on a cool night and still good on a hot day in the 80's or 90's??
My issue with my MAT (which is mounted in the air path in front of the throttle body) is because I have an intake similar to a miniram with a solid roller that I rev to 7000 rpm's the tune wasn't consistent. The changes in temp would throw me off in my tune from a cold night to a hot summer day. I was hoping there was a solution out there since the LT1's revved pretty high... This maybe it!!!
Thanks,
Sal
That's great news... I'm going to try it. Do you think after getting the VE tables right BLM's were good with large temp changes. Example at 45 degress on a cool night and still good on a hot day in the 80's or 90's??
My issue with my MAT (which is mounted in the air path in front of the throttle body) is because I have an intake similar to a miniram with a solid roller that I rev to 7000 rpm's the tune wasn't consistent. The changes in temp would throw me off in my tune from a cold night to a hot summer day. I was hoping there was a solution out there since the LT1's revved pretty high... This maybe it!!!
Thanks,
Sal
Mine seems pretty consistent these days. Hopefully it works out for you.
Wow... ressurrecting an old thread! I do have 128 pretty much every where now. Yes, I did have to retune somewhat to get that though. Changing to the BDZL tables altered the fuel delivery enough to require changing the VE tables.
It was worth it though. I remember seeing a noticeable improvement in power and driveability.
I did something similar to you in that I used my outside temp display to help tune the IAT table. The temp display sensor is located in the radiator shroud inlet w the IAT being a bit further up the intake tract and after the air filter. Now I get more consistent BLMs regardless of outdoor temps. I find that my outside temps will rise as the car sits in traffic. Not sure if it's the rad heat or heat coming from pavement/other car exhausts.
The MAT compensation tables for the PW calculation in the $8D mask has misleading titles. The purpose of the two 'MAT' tables is to provide an inverse temperature term for the PW calculation.
Colder air has a greater density. This needs to be accounted for in the PW calculation. The inverse temperature term provided by the 'MAT Compensation' routines is based on both CTS and MAT.
The amount of CTS vs. MAT that is used can be varied. One table is a model of the intake tract heating of the incoming air. IOW, how hot does the air become just from its travels through the intake. The engine coolant temperature affects this.
When the MAT is relocated and becomes an IAT sensor the table calibration is no longer correct.
When the intake design is changed (say to a LT1 or single plane manifold) the table calibration is no longer correct.
There is also an option bit involved. At location $018, b4 is used to select either the intake tract heating model table, or to just use the MAT/IAT as is. If this bit is set to a 1, then the intake tract model table is used. Else, if set to a 0, then the MAT/IAT is used as is.
Using a MAT as is would probably work OK. Using an IAT as is would not. There is going to be some heating of the air from the intake tract. For a decent tune this needs to be accounted for.
Here is the Intake Tract Heating Model table (with the original title):
Note the table is based on airflow. The less time the air is within the manifold the less heat it will absorb from the intake tract.
This table is used to combine the MAT/IAT sensor value with the CTS value. Here is the math:
T = ((CTS - MAT/IAT) * MULT) / 128 + MAT/IAT
What is being done is a portion of the CTS value is being added to the MAT/IAT value. This provides a new temperature term (T). In simpler words: take the difference between the air temperature and the coolant temperature, take a percentage of this difference and add it back to the air temperature.
The table value for 0 gms/sec airflow is 128 (MULT), in affect the T term is going to be the same as the coolant temperature (CTS).
As the airflow increases less of the coolant value (CTS) is added to the air value (MAT/IAT).
Here is an example at 64 gms/sec, 10°c MAT/IAT and 85°C CTS. From the table the MULT term is 27 at 64 gms/sec:
T = ((85 - 10) * 27) / 128 + 10 = 26°c
The same temperatures at 128 gms/sec gives this:
T = ((85 - 10) * 16) / 128 + 10 = 19°c
The MULT term is changed to 16 (from the table) and the final temperature term (T) is lower at 19°c.
What is interesting about the above examples is that a MAT sensor with a engine at 85°c is probably not going to be at 10°c. Well, maybe if the ambient air is at -50°c. That is where the stock calibration falls apart with a relocated MAT to an IAT.
A MAT would be about (SWAG here) 50°c when the engine is up to temperature. Lets run the first example with that:
T = ((85 - 50) * 27) / 128 + 50 = 57°c
Quite a difference from the T = 26°c of the first example. In this case relocating the MAT will cause the engine to run richer and the BLMs to drop. The table will need to be recalibrated.
The last table is easy. The values in this table shouldn't really need to be changed. Here it is:
Easy enough, use the temperature term from the previous calculation and look up the inverse temperature term (~T) to be used in the PW calculation.
The column labled 'MAT Deg c' is really the T term. Which can and should be a combination of the MAT/IAT and CTS values.
In closing, as others have stated the Intake Tract Heating Model table is the key to recalibration. With a better understanding of how the table works it should be easier to do a better tune.
My thoughts on the stock calibration shows that at low air flows the stock long tube TPI setup adds a lot of heat. This can be seen by the full CTS temperature term at 0 gms/sec
Even at higher air flows with a heat soaked MAT, the Intake Tract Heating Model will move the T term to a high value.
An intake with shorter runners will heat the air less then a TPI. There is simply less surface area and less time for the air to pick up heat as it travels through the manifold.
RBob.
Care to explain how the $8D mask measures volume of air?
That may not be the correct question. The volume of air drawn into the cylinder is the volume of the cylinder. That’s not measured at all. It’s pre-defined by the value entered for the engine displacement constant.
What the air temperature is being used for is to calculate the mass of air inside that volume after the piston reaches BDC, which is a function of the air density at the given temperature;
That calculation process is what is being explained in this thread.
That may not be the correct question. The volume of air drawn into the cylinder is the volume of the cylinder. That’s not measured at all. It’s pre-defined by the value entered for the engine displacement constant.
What the air temperature is being used for is to calculate the mass of air inside that volume after the piston reaches BDC, which is a function of the air density at the given temperature;
That calculation process is what is being explained in this thread.
Where in the $8D mask is grams per second used? Yes there's a cylinder volume assignment and it measures vacuum in KPA but no where does it measure incoming air. The code used in the OP's original post is not from the $8D mask.
Last edited by RBob; 11-06-2017 at 12:09 PM.
Reason: removed excess quoting
It's used in the Inv MAT table to look up the multiplier.
Though maybe it doesn’t need to be measured directly. Perhaps it is a derived number based on VE (pre-programmed), displacement (pre-programmed), engine RPM (measured) and air density for the given air temperature (measured) and barometric pressure ahead of the throttle body (not sure how this is determined… though if sea level is assumed then you’d probably see any error due to higher altitude show up as a fueling correction resulting in a departure from 128 BLM ideal).
But given those values, I can pretty easily do a hand-calculation to determine the air flow in g/sec. I suppose if I can do it, so can the ECM…
The ’93 LT1’s are doing the same thing I believe as they only had MAP sensors and are also using the same Inv MAT table.
Where in the $8D mask is grams per second used? Yes there's a cylinder volume assignment and it measures vacuum in KPA but no where does it measure incoming air. The code used in the OP's original post is not from the $8D mask.
The airflow in gms/sec is used in many places. You should have checked the ANHT hac before posting.
As for the code in my first post of this thread, it is used in $8D. Again, a look at the hac would show this.
I would assume that the MAT compensation tables basically reflect changes in air density?
If that's the case, why would the slopes be different between different cars? Is my assumption incorrect?
I'm just talking about the slope of the curves, not the differences in values at a given temperature...
Yes they reflect air density (kg/m3) and is part of INJ BPW calculation. This is the formula for $8D:
Would that depend on where the MAT (IAT) is located? In a TPI, it seems the MAT is more a "Plenum Temperature" sensor than a MAT. Maybe each BIN are specific to an engine set-up. Maybe it was improved just like modern updates to software.
I guess it all depends where the air temperature data is acquired in the intake track. Then, the intake track characteristics would play a role (heat transfer rate). So 2 different engines, most likely, will have different air intake characteristics. Also, air velocity and quantity might play a role. Bigger engine = more air = less % of heat transfer = more density.
I guess it's all about estimating air density based on air temperature readings at a certain location in the intake track vs air velocity (g/sec).
I have relocated the MAT to the airbox entry instead of under the plenum. By doing so there is less variation in MAT readings and heat soaking is cleared quickly. My MAT tables had to be completely retuned, so maybe I got a different curve slope. OEM curve was like ANHT.
Here it is:
Now this table was tuned in conjunction with MAT Inverse Term Lookup Delta Mult. vs Airflow. It is tuned to max out pass a certain level of heat soaking. The calculated final MAT is what the ecm will use for air density purposes at WOT. This was necessary to stabilize real AFR in PE when IAC sensor is heat soaked. It limits the amount of fuel pulled at WOT. The BLM/INT will take care of air density variation when in CL. At idle, on hot engine, outside air temperature as little effect on air density in cylinder.
I disabled my Inv MAT table altogether. With the MAT sensor no longer mounted in a metal plenum on the engine, didn't make sense to run that table. The air spends so little time in the metal intake (short runner Miniram) and I have a non metallic air tube so I didn't see the air having enough time to get preheated.
I can see the Inv MAT table being dependent on engine and air intake configuration as well as sensor location, but to me air density is air density regardless of anything except temperature and pressure.
I wonder why GM would be trying to factor this in?
Who knows eh?
I did keep them enable even if relocated the IAT to air box. I’m trying to remember why! Definitely had to retune them. I believe no matter where you get your reading from IAT there is still a difference between idle and WOT heat transfer. So I might have tuned it that way, basically adding fuel as the air flow increases.
I find that relocating IAT gave me ease of tuning not have to worry about large variation in MAT readings. And also, sensor heat soaking dissipates more rapidly.
If you wanted to move the BLMs by say, 4 points richer by altering the MAT compensation table at a particular temp, how much would you move the MAT table value?
IOW, Is there a correlation/formula between MAT counts and BLM? Like on the VE table, you'd simply scale the VE cells by the ratio of your BLM against 128. Is there something analogous for the MAT table?
If you wanted to move the BLMs by say, 4 points richer by altering the MAT compensation table at a particular temp, how much would you move the MAT table value?
IOW, Is there a correlation/formula between MAT counts and BLM? Like on the VE table, you'd simply scale the VE cells by the ratio of your BLM against 128. Is there something analogous for the MAT table?
I am going to try answering this. Not easy to explain.
Yes there is a certain correlation (if we can call it that) between the MAT "count" and the INJ BPW, before BLM/INT correction.
First, here's the formula for INJ BPW on $8D:
Second: "Air (Kg/m3)" is the calculated air density based on MAT input. Now we know the MAT input is modified before being used in the above formula. The "Inverse Term Lookup Delta Multi. vs Airflow" (table 0x80E) can play a role if enable. The MAT input is adjusted using 0x80E and CT input. Formula is the following:
CT - MAT = Delta
ECM uses an airflow g/sec calculation based on this formula: (INJ Flow Rate/60/60)*1000/2.20462 = Air Flow in g/sec Cyl. Vol. in ml * 4 * VE% * MAP/Baro * ((Baro*1000)/(287.058*(MAT+273.15)))/1000*RPM/60
It then looks up the g/sec using 0x80E which returns a multiplier
Adjusted MAT = MAT input + (Delta*multiplier)
The "Adjusted MAT" correspond to counts in table 0x7FD. Each count correspond to a temperature value hardcoded in the BIN. It is not linear. Here's the graph:
What ever value is returned by the count entered in the table 0x7FD, I call it "Final MAT". This is the value used in the above formula, Air (Kg/m3).
Third: So you can see how many moving parts there are:
If Air (Kg/m3) is adjusted +/- x% it will results in the same +/- y% in INJ BPW before BLM/INT.
It has to be Air (kg/m3) delta, not Deg.C delta.
It should be an inverse relationship between Deg.C and Air (Kg/m3). The hotter the less density.
Air(Kg/m3) = (Baro*1000)/(287.058*(Final MAT+273.15))*0.99 (Humidity is locked at 50% for simplicity).
So BLM would have to move in opposite direction from "y", at the same rate
Problem is, the BLM will adjust based on the INT. So INT would react first to keep INJ BPW stable. Then, BLM would adjust if thresholds were met.
BLM can only +/- by 1, which is +/- 0.8%
Fourth: Problem is, the nonlinear regression above. Every +/- count does not always = the same +/- "Final MAT", thus not the same +/- Air(Kg/m3). So you have to know what impact each count have on Air(Kg/m3).
Example:
Increasing counts from 30 to 40 = 4Deg.C to 11Deg.C = 7 delta
Increasing counts from 110 to 120 = 42Deg.C to 46Deg.C = 4 delta
Increasing counts from 30 to 40 = 1.282Kg/m3 to 1.250Kg/m3 = -0.032 delta
Increasing counts from 110 to 120 = 1.127Kg/m3 to 1.113Kg/m3 = -0.014 delta
Increasing counts from 30 to 40: x = -2.5%
Increasing counts from 110 to 120: x = -1.2%
Fifth: Example:
Everything else being equal
You increase counts from 30 to 40
"Final MAT" increases from 4Deg.C to 11Deg.C
Air (Kg/m3) decreases from 1.282Kg/m3 to 1.250Kg/m3
x = -2.496%
y = -2.496%
INT has to go up 2.496%
If INT is at 128, it needs to go to 131.2, rounded to 131 or 132, don't know
Probably not enough to trigger BLM adjustment (needs +4 INT for a certain duration, I believe)
Hope this helps. Took about an hour 1/2 to write and test.
11-01-2004, 05:35 PM
