NGK AFX wideband and E85

[jim2092]

Ah yes, that looks better. I shall make a log and post it and my config. Thanks

[vltrb-0]

Ah yes, that looks better. I shall make a log and post it and my config. Thanks

yep so check your lambda reads correct to your gauge

then you'll setup commanded lambda which is 1/EQ (eq is a standard parameter you can list)

then the error is plot with respect to that value as a percentage

[jim2092]

Always heard that the people in Melbourne were great. When I was in Viet Nam in the 60's, Melbourne was the #1 choice for R&R but we had to be in-country for 11 months to go there. Couldn't wait that long so I went to Bangkok instead. lol

[jim2092]

The config and a small log as requested. What impact does the stoich AFR in the tune have on these lambda results?

[LSxpwrdZ]

I'll have to take a look at the log tomorrow but looks like you got the custom wideband PID setup correctly now. My spreadsheet was setup to use AFR to figure the Lambda function not the other way around. Sometimes if the equation can't be figured out it's easier to just input the settings in the "List" instead of the "Linear" tab on the custom PID screen.

Stoich AFR has a global impact on fueling. It's very important that it is set correctly in the tune based on the fuel you are running on. In my area I've found that we don't have full E10 like alot of stations are now using so to get in the middle I use a stoich value of 14.42 so that it will run fine on E10 or E0.

[vltrb-0]

The config and a small log as requested. What impact does the stoich AFR in the tune have on these lambda results?

your cfg looks good now, two of the parameters for me show up as unsupported i'm not sure why but i can see the equations and they appear correct

with regards to AFR they will not have any effect on your calculations from now on so you can change fuels and keep this setup, all you'll have to do is change your programmed stoich value and adjust VE/spark tables

[jim2092]

It turns out that my commanded EQ (PID.68) is actually EQ and not lambda. It reads 1.0 at all times except at WOT it reads 1.172 which is my PE at WOT and lambda reads .85 - both correct. But my lambda error reads .72. Therefore my function for that is wrong. Using AUX.20122 as my lambda input (LM1 thru the EIO), what would my correct function equation be for lambda error? I will spare you my guesses.

I presently have [AUX.20122]/[PID.68]

[backchannel]

:

It turns out that my commanded EQ (PID.68) is actually EQ and not lambda. It reads 1.0 at all times except at WOT it reads 1.172 which is my PE at WOT and lambda reads .85 - both correct. But my lambda error reads .72. Therefore my function for that is wrong. Using AUX.20122 as my lambda input (LM1 thru the EIO), what would my correct function equation be for lambda error? I will spare you my guesses.

I presently have [AUX.20122]/[PID.68]

(([AUX.20122]-[PID.68])/[PID.68])*100 ..........maybe
or ((AUX.20122]-1/[PID.68])/1/[PID.68])*100
post if this doesn't work

[jim2092]

The second equation seems to work. When lambda was 1.4, lambda error was 4.30 and when lambda is 1.01, error was 1.33, etc. It appears to be an exact % of lambda error to 2 decimal points. Now that I have it, not sure of its value as lambda is already providing that info rounded to whole %? I guess in PE it would be useful.

[vltrb-0]

The second equation seems to work. When lambda was 1.4, lambda error was 4.30 and when lambda is 1.01, error was 1.33, etc. It appears to be an exact % of lambda error to 2 decimal points. Now that I have it, not sure of its value as lambda is already providing that info rounded to whole %? I guess in PE it would be useful.

your EQ ratio is just a factor the ecu uses to increase or reduce fuel flow otherwise it will always target lambda = 1 (stoich)

your commanded lambda is 1/EQ, that is 1 = lambda commanded, EQ is the modifier

your equivilance ratio should be a standard parameter in your PID list

you will make a user defined PID and display this on the scanner, this will give you for example something like [user9004]

now anywhere you want to use a calculation using your commanded lambda you'll use the user defined PID

so to work out your error you'll want to have the following equation

100*((wideband input - commanded lambda)/commanded lambda)

so wideband for EIO might be aux20122

commanded lambda = in my example [user9004]

and 100 = converting to a %

the display you make in your scanner should be set to units of percentage

also some error will be expected, this is why your car would require tuning the setup you have just created is telling you the error in what your ecu is trying to give your car and what your car is actually given

[jim2092]

So I should not use the equation below? I should make a user defined to replace the 1/(PID.68)? PID.68 being my commanded EQ ratio.

[vltrb-0]

So I should not use the equation below? I should make a user defined to replace the 1/(PID.68)? PID.68 being my commanded EQ ratio.

nothing wrong with writing it like that, but if you've got the commanded lambda in your scanner setup may as well use it to shorten the calc

[jim2092]

[vltrb-0]

that's it, now any errors you see are a result of your tune

you should plot this in a histogram

[LSxpwrdZ]

So I should not use the equation below? I should make a user defined to replace the 1/(PID.68)? PID.68 being my commanded EQ ratio.

Either of those setup's will work. One just incorporates the commanded EQ into the Lambda Error PID so that you don't have to have multiple PID's to do the same thing. I have mine setup like the later just because I like to visually have the commanded lambda displayed right above my actual lambda in my table display to watch while tuning.

[backchannel]

Either of those setup's will work. One just incorporates the commanded EQ into the Lambda Error PID so that you don't have to have multiple PID's to do the same thing. I have mine setup like the later just because I like to visually have the commanded lambda displayed right above my actual lambda in my table display to watch while tuning.

I use commanded lambda and lambda error in the table and histogram. What is the advantage of User defined vs listed pid other than a one time shorter equation entered?

[vltrb-0]

I use commanded lambda and lambda error in the table and histogram. What is the advantage of User defined vs listed pid other than a one time shorter equation entered?

there isn't one it's potato, potato

its like writing

1+1=2 or 1+(4-3)=2

[backchannel]

there isn't one it's potato, potato

its like writing

1+1=2 or 1+(4-3)=2

aha!

[LSxpwrdZ]

Exactly, same end result either way you go.

[jim2092]

Since the E40 (2005 C6) uses a different application of the EQ to obtain the correct lambda values then the config. used for the E38, I understand it would not work on an E38 (2009 C6)? If that is correct, does anyone have the neccessary config. and histograms to do open loop MAF tuning on this PCM (E38)? Using lambda and not AFR of course.