MAF Scaling - Open/Closed Loop Fueling info

[DocWalt]

Well, unfortunately changing the primary open loop table just meant the AFR error was even higher since the commanded AFR changed but the actual AFR was the same.

Any tips? I'm sure I'm doing something wrong, heh.

edit: See this log from ~30 minutes ago: http://datazap.me/u/docwalt/04-09-15...mark=3530-3583

and this log from a few days ago: http://datazap.me/u/docwalt/04-06-15...mark=6761-6865

The commanded AFR changed, but the actual AFR was essentially the same shape and AFR error

[Kodename47]

Originally Posted by DocWalt

Well, unfortunately changing the primary open loop table just meant the AFR error was even higher since the commanded AFR changed but the actual AFR was the same.

Any tips? I'm sure I'm doing something wrong, heh.

edit: See this log from ~30 minutes ago: http://datazap.me/u/docwalt/04-09-15...mark=3530-3583

and this log from a few days ago: http://datazap.me/u/docwalt/04-06-15...mark=6761-6865

The commanded AFR changed, but the actual AFR was essentially the same shape and AFR error

Revert the open loop fuel map. It doesn't look like the MAF scale is that close to give the commanded AFR. Did you scale the open loop part of the scale as it looks like after 2.5V hasn't been done?

[DocWalt]

Originally Posted by Kodename47

Revert the open loop fuel map. It doesn't look like the MAF scale is that close to give the commanded AFR. Did you scale the open loop part of the scale as it looks like after 2.5V hasn't been done?

To make the open loop part of the maf scale match what it's supposed to do at WOT I have to pull ~12% fuel in that rich spot, which means I get tip in knock correction of -1* or so and STFTs of +10-15% from 2.7v+ while in CL. I get a VERY nice smooth curve everywhere else, about -3% off the rest of the way up the revs at WOT.

CL maf scaling is a piece of cake, and OL would be fine except for the weird rich dip I get when WOT.

[Kodename47]

Well the generic way is to get MAF Scale = Commanded AFR. If you don't want that you can just adjust the fuel map to get the AFRs you want, or even the MAF scale to the same end. I'm surprised you're even in CL that high MAFv range. TBH I'd rather the OL range was better scaled than CL in the cross over regions due to the potential issues on tip in as you come out of CL operation. I actually use a combination of OL and CL in one go to do my MAF as it seems to be more reliable.

Which ROM are you on? Is the open loop fueling additive defined? You could always use that to make it leaner.

[DocWalt]

Originally Posted by Kodename47

Well the generic way is to get MAF Scale = Commanded AFR. If you don't want that you can just adjust the fuel map to get the AFRs you want, or even the MAF scale to the same end. I'm surprised you're even in CL that high MAFv range. TBH I'd rather the OL range was better scaled than CL in the cross over regions due to the potential issues on tip in as you come out of CL operation. I actually use a combination of OL and CL in one go to do my MAF as it seems to be more reliable.

Which ROM are you on? Is the open loop fueling additive defined? You could always use that to make it leaner.

Yeah, that's what I figured with wanting the MAF scaling to be closest. I just didn't like the idea of a massive dip in the MAF scale. I did CL then OL in one go, which resulted in the big dip which I manually smoothed.

I'm working from OFT Stage 2 UEL 91 Octane v2.063. I remember seeing the open loop fueling additive, I think. I'm not at home so I'll have to try that tonight.

Thanks for the help so far!

[Kodename47]

Originally Posted by DocWalt

Yeah, that's what I figured with wanting the MAF scaling to be closest. I just didn't like the idea of a massive dip in the MAF scale. I did CL then OL in one go, which resulted in the big dip which I manually smoothed.

I'm working from OFT Stage 2 UEL 91 Octane v2.063. I remember seeing the open loop fueling additive, I think. I'm not at home so I'll have to try that tonight.

Thanks for the help so far!

How did you do both in one go? Is there any chance your ranges overlapped causing the dip? By that I mean scale one, copy that scale into the other part of the tool and then run that. If so, if you don't use the Mac/min voltage settings, then you'll get skewed results around the change over regions.

I do mine manually rather than the tool these days, find it more reliable and will take into account everything if you can sort the data.

[DocWalt]

Originally Posted by Kodename47

How did you do both in one go? Is there any chance your ranges overlapped causing the dip? By that I mean scale one, copy that scale into the other part of the tool and then run that. If so, if you don't use the Mac/min voltage settings, then you'll get skewed results around the change over regions.

I do mine manually rather than the tool these days, find it more reliable and will take into account everything if you can sort the data.

I did CL up to 3v and OL down to 3v. I used the max/min settings, am I not supposed to?

FWIW I manually smoothed out that region of the MAF scale so there wasn't a massive dip and that resulted in the above behavior. Note the same AFR dip existed before scaling and after.

edit: Oh, and is the OL fuel additive simply an AFR adder? I assume so, but wanted to double check. The description in romraider isn't defined, but the units are defined as "Lambda Additive"

[DocWalt]

This is how my MAF scaling ends up compared to the stock OFT Stage 2 scaling if I remove the 3v min/max limits. Kinda lumpy, to say the least. I did the math and it should fix my WOT issues, but I'll end up with tip-in knock again, I think.

[Kodename47]

If you look at a graph of a WOT pull on Datazap what are the MAFv you have at 2-2.5k. That will be the cross over region, it's usually ~2.5v which is why I think the MAF scale changes it's spacing. If you want to use the tool I would:
1) Paste the original MAF scale into both open and closed loop screens and into excel (call this A)
2) Set no upper or lower voltage range
3) Run the tool for both open and closed loop and then paste both into excel (B and C)
4) Remove data from both new scales that is the same as the original, using the formula =if(A=B,"",B) may make that quick and easy (let's call these new rows D and E)
5) Average the 2 scales that only have the changed data using =average(D,E)

When I get my laptop i'll show a bit better what I do.

Yes, the fuel additive appears to be AFR points, -0.1 is richer, 0.1 is leaner. But remember it is a % of where you are. If your fueling is way off the MAF then 0.1 won't necessarily make it 0.1 AFR points leaner, it may be more or less. Again I have a spreadsheet I was trialling. I'll post it up.

[steve99]

Originally Posted by DocWalt

This is how my MAF scaling ends up compared to the stock OFT Stage 2 scaling if I remove the 3v min/max limits. Kinda lumpy, to say the least. I did the math and i

Their is a table for tip in enrichment, maybe adjust that for the tip in knock or take a bit of timing out of the base timing B table in that area or add a bit of fuel in the Open loop fuel table.

[DocWalt]

Kodename, I get what you're saying. I can do that, which will smooth out the dip at 2.5v-3.0v. I only had a couple of samples of CL past 3v anyway (and not sure how I managed that anyway)

Steve, I think pulling a touch of timing and/or adding a touch of fuel is a good idea. I can get 93 octane, but it's a bit out of the way so I tend to run 91/92. Every time I've tried 94 locally my IAM drops so Sunoco's tanks of 94 must be pretty old.

[DocWalt]

So after going through that work suggested by Kodename and Steve I got a couple of nice WOT logs. Some minor knock correction, but no IAM drops and I'm still on mediocre 91 octane gas.

http://datazap.me/u/docwalt/04-11-15...zoom=1249-1338

http://datazap.me/u/docwalt/04-11-15...zoom=2370-2504

The MAF changes I came up with based on those logs gives me a curve like this:


I guess I'll still have the little bit of a rich spot with the slightly lean spike. Unsure what's going on there, but based on my virtual dyno results from earlier it's not causing a loss of power.

Thank you for the help guys!

[Kodename47]

If you have used the log stats screen on the tool before, you may be able to workout how to use this:
Open Loop Additive Tool

For best results:
Set the cell range to ~20% on the define axis screen and paste in your OL Additive table as the axes.
Filter dV/dt to 0.3 or less, you'll need to add a column for dV/dt to your logs. The formula can be found here.
Filter Fuel System Staus = 2 (Open Loop Only)
Filter error <17.5% which will remove anywhere where you're off throttle and AFR is showing ~20:1

You'll need to creata a column called Error % into the log too, which is where you'll get the data. The formula is also in my spreadsheet

[DocWalt]

Originally Posted by Kodename47

If you have used the log stats screen on the tool before, you may be able to workout how to use this:
Open Loop Additive Tool

For best results:
Set the cell range to ~20% on the define axis screen and paste in your OL Additive table as the axes.
Filter dV/dt to 0.3 or less, you'll need to add a column for dV/dt to your logs. The formula can be found here.
Filter Fuel System Staus = 2 (Open Loop Only)
Filter error <17.5% which will remove anywhere where you're off throttle and AFR is showing ~20:1

You'll need to creata a column called Error % into the log too, which is where you'll get the data. The formula is also in my spreadsheet

Neat, I just ran through the process on that. Thanks! That should let me kill the AFR dip and spike after I get some fresh logs with the tweaked scaling. Hopefully I see some good results from that. Not sure what it is in my setup causing the dip and spike, but ok.