Big Blue's Transformation

[Gary Lewis]

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?

I reset the parameters that were preventing adaptive learning and wrote that tune to the ECU.

• Lowest ACT To Enable Adaptive Learning: From Ben's 246F to the stock -20F

• Lowest ECT To Enable Adaptive Learning: From Ben's 240F to the stock 150F

Then I started it and took it for a drive. The idle AFR seemed better but it'll take time to learn. In fact, the ECT hadn't gotten to 150F before I left so learning wasn't enabled. However, when I got back it appears to have learned something. I'm letting it cool down now so will start it up in a bit and see how it idles then.

But I've been looking at the log and it is interesting. The section below is while on cruise pulling a hill at 65 MPH. You can see the colors in the legend to the left, and the values for each parameter are from where the vertical white line is.

Note that each trace has its own scale, and the one we are on is TP - Throttle Position. IOW, you can't compare AFR, which is the blue line, to LAMBSE2 which is the white line as they are on different scales. However, AFR, which is from the wide-band gauge, is at 14.66:1 while LAMBSE2, which is commanded AFR for the left bank, is 15.14 at that instant.

But notice that Throttle Position is 2.02 and the commanded AFR is headed down almost immediately. Apparently the ECU triggers into the power mode at that throttle position, and it wants the AFR to be 12.36, which is why the white trace heads straight down. And that takes .17 seconds.

But note that AFR doesn't change that quickly. It was running at 14.66 and then it goes to 12.64, but that takes .78 seconds. So apparently it takes ~.6 seconds for the change to be seen at the O2 sensor after the computer alters things.

Another observation is that the sawtooth of LAMBSE2 is typically from 14.5 to 15.2 when running at steady state of 65 MPH. At the same time the actual AFR is going from 14.7 to 14.8, which is pretty close. And from what I've read you are pretty well "tuned" when the actual matches commanded AFR.

 

[85lebaront2]

I reset the parameters that were preventing adaptive learning and wrote that tune to the ECU.

• Lowest ACT To Enable Adaptive Learning: From Ben's 246F to the stock -20F

• Lowest ECT To Enable Adaptive Learning: From Ben's 240F to the stock 150F

Then I started it and took it for a drive. The idle AFR seemed better but it'll take time to learn. In fact, the ECT hadn't gotten to 150F before I left so learning wasn't enabled. However, when I got back it appears to have learned something. I'm letting it cool down now so will start it up in a bit and see how it idles then.

But I've been looking at the log and it is interesting. The section below is while on cruise pulling a hill at 65 MPH. You can see the colors in the legend to the left, and the values for each parameter are from where the vertical white line is.

Note that each trace has its own scale, and the one we are on is TP - Throttle Position. IOW, you can't compare AFR, which is the blue line, to LAMBSE2 which is the white line as they are on different scales. However, AFR, which is from the wide-band gauge, is at 14.66:1 while LAMBSE2, which is commanded AFR for the left bank, is 15.14 at that instant.

But notice that Throttle Position is 2.02 and the commanded AFR is headed down almost immediately. Apparently the ECU triggers into the power mode at that throttle position, and it wants the AFR to be 12.36, which is why the white trace heads straight down. And that takes .17 seconds.

But note that AFR doesn't change that quickly. It was running at 14.66 and then it goes to 12.64, but that takes .78 seconds. So apparently it takes ~.6 seconds for the change to be seen at the O2 sensor after the computer alters things.

Another observation is that the sawtooth of LAMBSE2 is typically from 14.5 to 15.2 when running at steady state of 65 MPH. At the same time the actual AFR is going from 14.7 to 14.8, which is pretty close. And from what I've read you are pretty well "tuned" when the actual matches commanded AFR.

Looking good sir! Lag may be the time it takes for the exhaust stream to reach the O2 sensor and the older portion has cleared out. On Darth's WB I can definitely see the delay from "open throttle" to sensing the desired 12.5:1 for max power. Keep in mind Darth's WB is about where the middle of BB's transfer case sits as it is behind the E4OD mount crossmember.

[Gary Lewis]

I reset the parameters that were preventing adaptive learning and wrote that tune to the ECU.

• Lowest ACT To Enable Adaptive Learning: From Ben's 246F to the stock -20F

• Lowest ECT To Enable Adaptive Learning: From Ben's 240F to the stock 150F

Then I started it and took it for a drive. The idle AFR seemed better but it'll take time to learn. In fact, the ECT hadn't gotten to 150F before I left so learning wasn't enabled. However, when I got back it appears to have learned something. I'm letting it cool down now so will start it up in a bit and see how it idles then.

But I've been looking at the log and it is interesting. The section below is while on cruise pulling a hill at 65 MPH. You can see the colors in the legend to the left, and the values for each parameter are from where the vertical white line is.

Note that each trace has its own scale, and the one we are on is TP - Throttle Position. IOW, you can't compare AFR, which is the blue line, to LAMBSE2 which is the white line as they are on different scales. However, AFR, which is from the wide-band gauge, is at 14.66:1 while LAMBSE2, which is commanded AFR for the left bank, is 15.14 at that instant.

But notice that Throttle Position is 2.02 and the commanded AFR is headed down almost immediately. Apparently the ECU triggers into the power mode at that throttle position, and it wants the AFR to be 12.36, which is why the white trace heads straight down. And that takes .17 seconds.

But note that AFR doesn't change that quickly. It was running at 14.66 and then it goes to 12.64, but that takes .78 seconds. So apparently it takes ~.6 seconds for the change to be seen at the O2 sensor after the computer alters things.

Another observation is that the sawtooth of LAMBSE2 is typically from 14.5 to 15.2 when running at steady state of 65 MPH. At the same time the actual AFR is going from 14.7 to 14.8, which is pretty close. And from what I've read you are pretty well "tuned" when the actual matches commanded AFR.

Well, I think we may be "there" on the EFI. I'm still wanting to play with the idle speed a bit, but it is much better. And I think the AFR is going to 12.6 too soon on hills, and my next post will be speculation on how to resolve that.

But in this one I want to document the changes to this point. I've done a compare between Ben's changes awa a few of mine and the YER2 tune that is for a stock CA-spec 460 with a manual transmission, as shown below.

Note that the Tables with the "..." are clickable in Binary Editor and open up the table that was changed, but that will be for the next post.

 

[Gary Lewis]

Looking good sir! Lag may be the time it takes for the exhaust stream to reach the O2 sensor and the older portion has cleared out. On Darth's WB I can definitely see the delay from "open throttle" to sensing the desired 12.5:1 for max power. Keep in mind Darth's WB is about where the middle of BB's transfer case sits as it is behind the E4OD mount crossmember.

Thanks, Bill! But I have a question for you, albeit a long-winded one, and anyone else that has input.

As said, I'm seeing the AFR go to 12.5ish sooner than I would like, and am sure I've found why. The tables below are for Base O/L Fuel Table Stabilized, and they show the desired AFR vs load and RPM.

The first shows Ben's changes vs stock, and basically he is taking the AFR to 12.35 at 60% load and any RPM. And he takes it to 13.15 at 2800 RPM with 30 - 60% load.

The second chart shows Ben's changes vs Bill's, and things are quite different. On Bill's it doesn't go to 12.35 until 70% load, and then only from 1600 RPM up. But it does go to 13.15 at 60% and 3200 RPM.

I like Bill's approach. It seems to be a good compromise between stock, which never gets richer than 13.15 so is leaving power on the table, and Ben's which jumps to 12.35 very quickly. And one reason I like Bill's is that I've reviewed the logs and I'm not getting quite to 70% load on the hills around here, even on cruise. So I'm pretty sure this would give the same MPG as the stock table but still provide more power when wanted. However, I'm wondering about going to 13.15 at 65% load starting at 1600 R's if it seems to sag on the hills.

Thoughts? Better ideas?

[Gary Lewis]

Thanks, Bill! But I have a question for you, albeit a long-winded one, and anyone else that has input.

As said, I'm seeing the AFR go to 12.5ish sooner than I would like, and am sure I've found why. The tables below are for Base O/L Fuel Table Stabilized, and they show the desired AFR vs load and RPM.

The first shows Ben's changes vs stock, and basically he is taking the AFR to 12.35 at 60% load and any RPM. And he takes it to 13.15 at 2800 RPM with 30 - 60% load.

The second chart shows Ben's changes vs Bill's, and things are quite different. On Bill's it doesn't go to 12.35 until 70% load, and then only from 1600 RPM up. But it does go to 13.15 at 60% and 3200 RPM.

I like Bill's approach. It seems to be a good compromise between stock, which never gets richer than 13.15 so is leaving power on the table, and Ben's which jumps to 12.35 very quickly. And one reason I like Bill's is that I've reviewed the logs and I'm not getting quite to 70% load on the hills around here, even on cruise. So I'm pretty sure this would give the same MPG as the stock table but still provide more power when wanted. However, I'm wondering about going to 13.15 at 65% load starting at 1600 R's if it seems to sag on the hills.

Thoughts? Better ideas?

Well, this table says it is for "open loop". And I wouldn't have thought we were in open loop at this point. In fact, the data doesn't quite fit with being in O/L as the transitions to 12.35 are at about 65% rather than 60% as shown in the O/L table.

However, I'm wondering if there's a threshold on throttle position where we go into open loop? It is happening almost exactly at 2.0 volts on the throttle position, so might that trigger O/L?

I have a closed-loop indicator on my "dashboard" but it isn't working so I either have the wrong parameter or maybe I'm not logging it, and it won't work on the dashboard if you aren't logging it. So I need to do some work to figure out what to monitor so I can see in the logs when we go from O/L to C/L and vice versa.

But, it won't hurt to change the table to match Bill's and see if that changes the AFR. Then I'll know for sure.

[Gary Lewis]

Well, this table says it is for "open loop". And I wouldn't have thought we were in open loop at this point. In fact, the data doesn't quite fit with being in O/L as the transitions to 12.35 are at about 65% rather than 60% as shown in the O/L table.

However, I'm wondering if there's a threshold on throttle position where we go into open loop? It is happening almost exactly at 2.0 volts on the throttle position, so might that trigger O/L?

I have a closed-loop indicator on my "dashboard" but it isn't working so I either have the wrong parameter or maybe I'm not logging it, and it won't work on the dashboard if you aren't logging it. So I need to do some work to figure out what to monitor so I can see in the logs when we go from O/L to C/L and vice versa.

But, it won't hurt to change the table to match Bill's and see if that changes the AFR. Then I'll know for sure.

Ok, I set my table up to match Bill's, with one exception. I don't see a reason to wind the thing to 4400 with only 20% load, but if I do then I can't see why I'd want the AFR to be 16:01:1.

We have a picnic at noon tomorrow, but maybe I can get it into the truck and test it before that.

Anyway, here's mine on top and Bill's on the bottom:

[85lebaront2]

Ok, I set my table up to match Bill's, with one exception. I don't see a reason to wind the thing to 4400 with only 20% load, but if I do then I can't see why I'd want the AFR to be 16:01:1.

We have a picnic at noon tomorrow, but maybe I can get it into the truck and test it before that.

Anyway, here's mine on top and Bill's on the bottom:

The reason for 4400 rpm, is that is my WOT upshift rpm on the E4OD. If you go back and look at that part of the tune you will see that. New engine may be set to 5000 or even 5500.

[Gary Lewis]

The reason for 4400 rpm, is that is my WOT upshift rpm on the E4OD. If you go back and look at that part of the tune you will see that. New engine may be set to 5000 or even 5500.

This sounds like something my father would have said, but you might want to change your tune a bit if you expect to wind it to 5K. I'm limited to 70% torque at 4700, and you are at 4800.

[Gary Lewis]

This sounds like something my father would have said, but you might want to change your tune a bit if you expect to wind it to 5K. I'm limited to 70% torque at 4700, and you are at 4800.

I wrote the tune in with the "Base O/L Fuel Table Stabilized" in it that is shown above, which is basically Bill's, and then took the truck for a drive. And after studying the log, particularly the section shown below, I think I'm learning a few things.

First, the system is going from closed loop to open loop, and back, at some throttle position or load setting, or combination thereof. The reddish square wave indicates that, and when it is high it is in O/L and when it is low it is in C/L. And notice that the white LAMBSE2 trace, which is the AFR the system wants, goes flat when the system goes into open loop mode - there's no use of the O2 sensors to control the AFR.

Second, when it goes into open loop it does use the Base O/L Fuel Table Stabilized values, but not quite in the way I expected. As shown, when LOAD exceeded 65% the desired AFR went from 14.64 down eventually to 12.35. And as soon as LOAD drops below 65% the desired AFR comes back to 14.64. In other words, the percent row is the upper limit for that value, not the lower limit.

If I can figure out where that parameter or table is that is taking it into O/L I might change that point. However, the system really does a pretty good job of keeping the actual AFR at the desired level even in open loop. For instance, there are three places in this snippet where the desired AFR stays flat at 14.64 for some time. And actual AFRs are 14.46, 14.67, and 14.36 from left to right. So it isn't critical that I make it quit going into O/L, although as things like barometric pressure, air temp, etc change that may get more important.

Last, even when the actual AFR does come down to 12.4 to follow the commanded AFR I cannot feel a difference in power. In other words, keeping the AFR at 14.6 or a bit higher doesn't cause the power to drop off. And that tells me that I can change the table to run 14.64 at a higher load percentage, especially since I now know that the shown percentage is the upper limit rather than the lower limit.

Does this make sense?

[85lebaront2]

I wrote the tune in with the "Base O/L Fuel Table Stabilized" in it that is shown above, which is basically Bill's, and then took the truck for a drive. And after studying the log, particularly the section shown below, I think I'm learning a few things.

First, the system is going from closed loop to open loop, and back, at some throttle position or load setting, or combination thereof. The reddish square wave indicates that, and when it is high it is in O/L and when it is low it is in C/L. And notice that the white LAMBSE2 trace, which is the AFR the system wants, goes flat when the system goes into open loop mode - there's no use of the O2 sensors to control the AFR.

Second, when it goes into open loop it does use the Base O/L Fuel Table Stabilized values, but not quite in the way I expected. As shown, when LOAD exceeded 65% the desired AFR went from 14.64 down eventually to 12.35. And as soon as LOAD drops below 65% the desired AFR comes back to 14.64. In other words, the percent row is the upper limit for that value, not the lower limit.

If I can figure out where that parameter or table is that is taking it into O/L I might change that point. However, the system really does a pretty good job of keeping the actual AFR at the desired level even in open loop. For instance, there are three places in this snippet where the desired AFR stays flat at 14.64 for some time. And actual AFRs are 14.46, 14.67, and 14.36 from left to right. So it isn't critical that I make it quit going into O/L, although as things like barometric pressure, air temp, etc change that may get more important.

Last, even when the actual AFR does come down to 12.4 to follow the commanded AFR I cannot feel a difference in power. In other words, keeping the AFR at 14.6 or a bit higher doesn't cause the power to drop off. And that tells me that I can change the table to run 14.64 at a higher load percentage, especially since I now know that the shown percentage is the upper limit rather than the lower limit.

Does this make sense?

Gary, you might want to look at what the timing is doing at the same points. If the computer is retarding the timing to prevent spark knock as you come up on the throttle your end result may be that the actual torque/HP stays the same.

On Darth's new engine, there are a number of parameters that will need changing, higher displacement, bigger injectors, more cam etc. and I will be doing a lot of what you are right now. Fine tuning for the end result.

Do keep in mind that when using my tune, that at moderate throttle I essentially have a manual transmission as long as my converter clutch is engaged, rpm will jump up roughly 500 rpm when it unlocks.