Big Blue's Transformation

[Gary Lewis]

Since I'd found a bad O2 sensor I realized that the closed-loop file I'd sent Ben Head was probably not good, so I sent him the latest log file just to make sure he didn't work on a bad one.Got a note back with a new tune he wanted me to try, and from the title of it I think he's turned adaptive learning off. He had me unhook the battery cable for a few minutes to reset the ECU, then load the tune, and drive it. I've just done that an emailed him the results.But he also said "The tune looks really close. Is there anything it is doing you don't like?" I responded that the MIL came back on and it looks like the EGR may not be fully disabled. And that I'm trying to get the idle where I want it, but overall it is running good.We shall see what he comes back with.

I may have found something, and I hope I can explain it. The parameters "SW For OBD2 MIL" and "SW For OBD2 Test" show as the "units" being "BITMASK". To an old programmer that tells me that the mask determines if the MIL can be turned on for each of the functions and if the test is even to be run.

 

In the comments there's a table that shows what each bit is for, and there are 8 bits. That suggests the code is octal, and while I spent many years reading octal for PDP8's I turned to this decoder to make sure I was doing it correctly.

 

Hopefully this makes sense to y'all. The numbers like "247" and "243" are the numbers in the tune, and you can see how they decode. And it seems to say that both my tune and Ben's are set up for the misfire test to be run and for it to light up the MIL when it doesn't. And both are set up to light the MIL if the purge test fails, but the purge test isn't to be run.

 

What I'm thinking is that this is an easier way to turn things on and off than what I've been doing.

 

[Gary Lewis]

Since I'd found a bad O2 sensor I realized that the closed-loop file I'd sent Ben Head was probably not good, so I sent him the latest log file just to make sure he didn't work on a bad one.

Got a note back with a new tune he wanted me to try, and from the title of it I think he's turned adaptive learning off. He had me unhook the battery cable for a few minutes to reset the ECU, then load the tune, and drive it. I've just done that an emailed him the results.

But he also said "The tune looks really close. Is there anything it is doing you don't like?" I responded that the MIL came back on and it looks like the EGR may not be fully disabled. And that I'm trying to get the idle where I want it, but overall it is running good.

We shall see what he comes back with. http://www.garysgaragemahal.com/uploads/6/5/8/7/65879365/crossed-fingers-20-pixel_orig.png

Ben got back with me this afternoon and sent another tune. I'll write it to the ECU in the morning and go log some more data and send it to Ben.

As I've said, one of the nice features of Binary Editor is the ability to compare tunes. And here's a comparo of the tune I've been using, which is a stock M/T 460 tune with my changes for EGR, etc with Ben's. Here's a summary of what I think he's done, but there are lots of details available if I double-click the rows from Mass Air Flow Curve down.

• Highest Air Charge Temp (ACT) and Engine Coolant Temp (ECT) To Enable Adaptive Learning: I don't think these values will ever come into play, but he raised them.

• Idle Speed In Drive & Neutral: Not sure these come into play either since it is a manual transmission, but we shall see.

• Lowest ACT To Enable Adaptive Learning: I think he locked it out of adaptive learning with this parameter. That allows him to dial it in w/o the computer changing things.

• Lowest Allowed Spark At Idle: He raised the minimum spark advance at idle by 7 degrees. It will be interesting to see what happens. However, I currently have the initial timing at 13 and I'll dial it back to the stock value of 10, so the change is only 4 degrees.

• Lowest ECT To Enable Adaptive Learning: Again he's locked adaptive learning out.

• Mass Air Flow Curve: He's raised the curve a little bit pretty much across the board.

• WOT Fuel Multiplier: He changed the values across the board. But the comments in this table say "This should never be utilized except in the case where the calibrator is testing fuel delivery thresholds." So I'm sure that is what Ben is doing.

• Base Open Loop Fuel Table Stabilized: From a load value of .6 up the table values have been changed. And the comments say "Comanded fueling after the vehicle has reached the stable ECT for a predetermined time."

• Spark Borderline Knock Threshold: Looks like he bumped light load/low RPM timing up by 1 degree, and heavy load/high RPM by 1/2 degree.

[Gary Lewis]

Ben got back with me this afternoon and sent another tune. I'll write it to the ECU in the morning and go log some more data and send it to Ben.

As I've said, one of the nice features of Binary Editor is the ability to compare tunes. And here's a comparo of the tune I've been using, which is a stock M/T 460 tune with my changes for EGR, etc with Ben's. Here's a summary of what I think he's done, but there are lots of details available if I double-click the rows from Mass Air Flow Curve down.

• Highest Air Charge Temp (ACT) and Engine Coolant Temp (ECT) To Enable Adaptive Learning: I don't think these values will ever come into play, but he raised them.

• Idle Speed In Drive & Neutral: Not sure these come into play either since it is a manual transmission, but we shall see.

• Lowest ACT To Enable Adaptive Learning: I think he locked it out of adaptive learning with this parameter. That allows him to dial it in w/o the computer changing things.

• Lowest Allowed Spark At Idle: He raised the minimum spark advance at idle by 7 degrees. It will be interesting to see what happens. However, I currently have the initial timing at 13 and I'll dial it back to the stock value of 10, so the change is only 4 degrees.

• Lowest ECT To Enable Adaptive Learning: Again he's locked adaptive learning out.

• Mass Air Flow Curve: He's raised the curve a little bit pretty much across the board.

• WOT Fuel Multiplier: He changed the values across the board. But the comments in this table say "This should never be utilized except in the case where the calibrator is testing fuel delivery thresholds." So I'm sure that is what Ben is doing.

• Base Open Loop Fuel Table Stabilized: From a load value of .6 up the table values have been changed. And the comments say "Comanded fueling after the vehicle has reached the stable ECT for a predetermined time."

• Spark Borderline Knock Threshold: Looks like he bumped light load/low RPM timing up by 1 degree, and heavy load/high RPM by 1/2 degree.

I ran the tune Ben sent me last night and sent him the log. He sent another one and I ran it and sent him the log. His response was "Everything looks really good now. If the CEL comes on you need to get a stored code."

So I asked him if we are done as I'm not sure adaptive is turned on. We'll see what he says.

And, the CEL has come back on. As I told him earlier, I can clear it but it then comes back on later and neither BE nor my app finds a code. So I think he is suggesting it is a stored code that needs to be cleared. Not sure how to find that, but I guess I'll go see.

Also, this happened today and it is all John's fault. He didn't like the steering wheel cocked about 1/8th turn CCW and suggested I fix it. But, what he didn't know is that one of the puller holes was stripped out. So I spent considerable time tapping that hole 3/8-16 as opposed to the 5/16-24 it should have been. Man, is that metal tough! And, it is hard to get to when the wheel is installed. But, of course, you can't remove the wheel until you tap it. Catch-22.

[Gary Lewis]

I may have found something, and I hope I can explain it. The parameters "SW For OBD2 MIL" and "SW For OBD2 Test" show as the "units" being "BITMASK". To an old programmer that tells me that the mask determines if the MIL can be turned on for each of the functions and if the test is even to be run.

 

In the comments there's a table that shows what each bit is for, and there are 8 bits. That suggests the code is octal, and while I spent many years reading octal for PDP8's I turned to this decoder to make sure I was doing it correctly.

 

Hopefully this makes sense to y'all. The numbers like "247" and "243" are the numbers in the tune, and you can see how they decode. And it seems to say that both my tune and Ben's are set up for the misfire test to be run and for it to light up the MIL when it doesn't. And both are set up to light the MIL if the purge test fails, but the purge test isn't to be run.

 

What I'm thinking is that this is an easier way to turn things on and off than what I've been doing.

 

Got a note back from Ben this morning in answer to my question regarding re-enabling adaptive learning. He said "No, I leave adaptive off on almost all tunes I do. Once the MAF is correct the closed loop will have instant correction. If you want to turn it on ok, but this is most likely your bad idle issues. Adaptive learns bad habits with the cam at idle."I asked if I could turn it off if I don't like it, but that I expected to have to clear memory by pulling the power cable for a few minutes, and he said "Yes". But I decided to run it the way it is for a while to see how I like it.However, the check engine light was back on and I cannot figure out why. There are no pending codes and neither BE nor the app on my phone can find anything. So I wrote the "042" mask into both the "SW For OBD2 MIL" and "SW For OBD2 Test" parameters and we took the truck to Owasso and back, which is about 40 miles round trip, with several stops. And the CEL stayed off. So maybe I'm onto something re the mask.But, the idle isn't where I want it yet. Ben left my 648 base idle in, but has 728 dialed in for Drive and Neutral, so I'll drop those 'cause it is idling about there. However, it still runs up to 1500 for several seconds on "cold" starts, so I need to sort that as well. But obviously the issue isn't with adaptive learning as it isn't on now.Anyway, we are getting there!

[grumpin]

Got a note back from Ben this morning in answer to my question regarding re-enabling adaptive learning. He said "No, I leave adaptive off on almost all tunes I do. Once the MAF is correct the closed loop will have instant correction. If you want to turn it on ok, but this is most likely your bad idle issues. Adaptive learns bad habits with the cam at idle."

I asked if I could turn it off if I don't like it, but that I expected to have to clear memory by pulling the power cable for a few minutes, and he said "Yes". But I decided to run it the way it is for a while to see how I like it.

However, the check engine light was back on and I cannot figure out why. There are no pending codes and neither BE nor the app on my phone can find anything. So I wrote the "042" mask into both the "SW For OBD2 MIL" and "SW For OBD2 Test" parameters and we took the truck to Owasso and back, which is about 40 miles round trip, with several stops. And the CEL stayed off. So maybe I'm onto something re the mask.

But, the idle isn't where I want it yet. Ben left my 648 base idle in, but has 728 dialed in for Drive and Neutral, so I'll drop those 'cause it is idling about there. However, it still runs up to 1500 for several seconds on "cold" starts, so I need to sort that as well. But obviously the issue isn't with adaptive learning as it isn't on now.

Anyway, we are getting there!

Good news! Man, you’re getting that figured out!

[Gary Lewis]

Good news! Man, you’re getting that figured out!

I wish I could agree, Dane. After seemingly figuring one thing out I've run into something else, and I'm wondering if I caused it.

After installing Ben's tune but with the bit mask of 042 added, we took the run to Owasso with no problems. With that I decided to change the Drive and Neutral idle speeds from 728 to 648. When I started the truck up after that the AFR went initially quite lean, like 17:1, and then when it went into closed-loop it went to 12.6:1.

I compared the tunes and the only differences were those two parameters, so I put the previous tune back in, which was Ben's tune of Bae 3 with the 042 mask. Started it up and it appears to have gone up around 17:1 although for the first 30 seconds the wide-band doesn't report, although just after it started to report the AFR was at 17 for a second. And then it came down to 12.6 until 111 seconds of the log, at which point it jumped up to 17+, and even reached a high of 19:1.

You can see all of that on the chart below, with yellow being AFR, green is RPM, and purple is the idle speed control solenoid's duty cycle.

So tomorrow I'm going to go back to Ben's tune w/o the mask change and see what happens there. And I'm going to be meticulous with my notes as I'm quite confused right now.

[85lebaront2]

Got a note back from Ben this morning in answer to my question regarding re-enabling adaptive learning. He said "No, I leave adaptive off on almost all tunes I do. Once the MAF is correct the closed loop will have instant correction. If you want to turn it on ok, but this is most likely your bad idle issues. Adaptive learns bad habits with the cam at idle."

I asked if I could turn it off if I don't like it, but that I expected to have to clear memory by pulling the power cable for a few minutes, and he said "Yes". But I decided to run it the way it is for a while to see how I like it.

However, the check engine light was back on and I cannot figure out why. There are no pending codes and neither BE nor the app on my phone can find anything. So I wrote the "042" mask into both the "SW For OBD2 MIL" and "SW For OBD2 Test" parameters and we took the truck to Owasso and back, which is about 40 miles round trip, with several stops. And the CEL stayed off. So maybe I'm onto something re the mask.

But, the idle isn't where I want it yet. Ben left my 648 base idle in, but has 728 dialed in for Drive and Neutral, so I'll drop those 'cause it is idling about there. However, it still runs up to 1500 for several seconds on "cold" starts, so I need to sort that as well. But obviously the issue isn't with adaptive learning as it isn't on now.

Anyway, we are getting there!

Gary, if you simply pull the EEC power keep alive power fuse, it will clear the memory which is what Ben wants. Once that is done it takes a little bit for the EEC to relearn the basic parameters. Darth does strange things after a battery disconnect or fuse pull.

[Gary Lewis]

Gary, if you simply pull the EEC power keep alive power fuse, it will clear the memory which is what Ben wants. Once that is done it takes a little bit for the EEC to relearn the basic parameters. Darth does strange things after a battery disconnect or fuse pull.

Bill - There's no learning going on as Ben has it turned off. I think I'll turn it on tomorrow and see what happens. But there certainly are some strange things going on. Maybe that's why?

[85lebaront2]

Bill - There's no learning going on as Ben has it turned off. I think I'll turn it on tomorrow and see what happens. But there certainly are some strange things going on. Maybe that's why?

It's not the adaptive learning it is the EEC picking up the basic parameters, specifically things like "how high a duty cycle is needed on the IAC for 1000 rpm".

[Gary Lewis]

It's not the adaptive learning it is the EEC picking up the basic parameters, specifically things like "how high a duty cycle is needed on the IAC for 1000 rpm".

I'm not sure what the difference is between "adaptive learning" and what you are describing, Bill. But I can see that there are a WHOLE BUNCH of parameters in the tune with "adaptive" in the title or the comment, as shown below. (There are actually two more, but they have to do with the transmission and don't apply to BB.)

I think what I will do is to enable adaptive learning tomorrow and drive it a bit to see what happens. That's because what I'm seeing isn't logical as far as I can tell, so maybe the "learning" can fix it?