new to tuning help with afx

[bluehawk2]

I am new to tuning and I am trying to figure some stuff out as far as logging my afr and logging it in lambda so I can figure my maf and ve tables easier.

I purchased a AFX wideband and I am wondering about the transfer function

so far I have 5/(1.098-0.618) + 0.618

v/10.4166 + 0.618


Is this correct?

Aslo I have read about stoich in hpt being 14.68 and with the afx it is 14.56

Where does this come into play with my calculations?


Also If i'm using diff fuels such as e10(which I understand has a stoich value of around 14.2)........how does that affect my calculations? What do I need to change for that? Is that the part in my Maf error PID where I would divide actual afr by 14.xx

I have a lot of questions

thank you

[mr.prick]

Based on the Lambda range above and stoich AFR=14.68
(Volts/0.7096)+9.072=AFR

(3.98/0.7096)+9.072 =14.68
(3.98/10.4166)+0.618 =1.00

Based off the Lambda range from the AFX manual (0.62-1.1) and stoich AFR=14.68
(Volts/0.7096)+9.102=AFR
(Volts/10.4167)+0.62=Lambda

The AFR & Lambda ranges/expressions in the AFX manual do not match perfectly.
WBO2 Expressions.xls

If you change the stoich AFR value you need to change the PID expression to match.
The voltage equal to stoich AFR should be the same voltage equal to Lambda 1.00

[bluehawk2]

I see now. thanks for the info


How is everyone hooking the AFX up?.........It says don't use a cigarette lighter plug, and it looks like it wants all the grounds going to one central location...............so just use the input wire on the interface and run the ground to the chasis?


thanks

[carlrx7]

i think the afx reads stioch as 14.5 instead of 14.68

[DSteck]

AFX reads stoich as 14.57, which I ignore. I don't use the gauge at all.



I ended up adding a wire to mine. I put new ends on the power hookup wires , but when I did that, I ran another wire from the ground hookup to the signal ground. This way, you can plug the ground into the HP Tuners interface, but that ground will be shared with the power ground. I always run directly off of the battery.

I use V/0.709 + 9.068 for my transfer function, but that's for AFR. For actual Lambda, use V/10.42 + 0.62 and you'll be set.

[A_W_O_L]

Not to takeover the thread, but I've been doing alot of searching on the AFX setup and it seems like there WERE alot of different opinions on how to set it up, the pre-configured NGK Wideband setup, a user-defined setup that uses offsets, a user-defined setup using measurements with a DMM of the analog output ...

Is there a clear-cut answer? or where those other methods just being a little over analytical?

[DSteck]

Not to takeover the thread, but I've been doing alot of searching on the AFX setup and it seems like there WERE alot of different opinions on how to set it up, the pre-configured NGK Wideband setup, a user-defined setup that uses offsets, a user-defined setup using measurements with a DMM of the analog output ...

Is there a clear-cut answer? or where those other methods just being a little over analytical?

Probably over analytical. People having consistency problems are probably using shitty power sources. I ran an extra ground in the wiring itself as mentioned above, and have had no problems with mine. Once a car is tuned at any given point, if I tell it to run a different AFR during those operating conditions, the actual matches the commanded, so I'm fairly confident it works.

[eatgoat]

Not to takeover the thread, but I've been doing alot of searching on the AFX setup and it seems like there WERE alot of different opinions on how to set it up, the pre-configured NGK Wideband setup, a user-defined setup that uses offsets, a user-defined setup using measurements with a DMM of the analog output ...

Is there a clear-cut answer? or where those other methods just being a little over analytical?

I have used the pre-configured setup with offsets and it seems spot on for me. Assuming all your grounds and power source are correct.

[A_W_O_L]

I just got off the phone with Biran at AFX Tech support. According to what they see the stoich value of gasoline (no ethanol) out of the pump is actually closer to 14.57 AFR than it is 14.7 (or 14.68 if you dont want to round) which is why their wideband controller is setup to switch at 14.57. According to AFX if you use the transfer functions listed in thier manual you will be using the same scaling as thier wideband controller (lambda = 14.57 AFR). This info is on page 12 of their Tuning Manual Rev 06.

I spoke with him about the output voltages not ACTUALLY reacing 0v or 5v. I asked if my transfer function should be setup to reflect the ACTUAL output voltages from the controller or the published 0v - 5v range stated in the manual. His response was that he has heard that people were not actually seeing the 0v - 5v but it was usually VERY close 0.01 on the low end and 4.96 or so on the high end, but the published transfer function will report the AFR or LAMBDA value that the sensor sees (LAMBDA = 14.57 AFR). And that you could re-work the transfer function to see 14.7 = Lambda if you wanted to but the difference is only 0.1 AFR, and that the difference between different tanks of fuel ( especially between winter\summer blends) will probably be bigger than the difference between 14.57 and 14.68.

[eatgoat]

I just got off the phone with Biran at AFX Tech support. According to what they see the stoich value of gasoline (no ethanol) out of the pump is actually closer to 14.57 AFR than it is 14.7 (or 14.68 if you dont want to round) which is why their wideband controller is setup to switch at 14.57. According to AFX if you use the transfer functions listed in thier manual you will be using the same scaling as thier wideband controller (lambda = 14.57 AFR). This info is on page 12 of their Tuning Manual Rev 06.

I spoke with him about the output voltages not ACTUALLY reacing 0v or 5v. I asked if my transfer function should be setup to reflect the ACTUAL output voltages from the controller or the published 0v - 5v range stated in the manual. His response was that he has heard that people were not actually seeing the 0v - 5v but it was usually VERY close 0.01 on the low end and 4.96 or so on the high end, but the published transfer function will report the AFR or LAMBDA value that the sensor sees (LAMBDA = 14.57 AFR). And that you could re-work the transfer function to see 14.7 = Lambda if you wanted to but the difference is only 0.1 AFR, and that the difference between different tanks of fuel ( especially between winter\summer blends) will probably be bigger than the difference between 14.57 and 14.68.

What do you mean by switch at 14.57:1? Assuming you are tuning the vehicle to 14.68:1 for a gasoline based engine how is the lambda value that AFX uses going to effect the results? The wideband can still see 14.68:1 ? Thanks in advance for any further clarification on this.

[carlrx7]

once you go lambda, you wont go back.

i have my afx configed to display lambda in hptuners. i dont even pay attention to what the display is showing.

-carl

[A_W_O_L]

eatgoat:
As I understand it:

What do you mean by switch at 14.57:1?.

See below. The Display will read 14.57 when Lambda = 1. This is true if you are burning gasoline, methanol, propane.... the point is that the wideband measures lambda and the controller's display formula is setup (and fixed) to calculate the AFR of gasoline which they define as 14.57 AFR. So if you were to watch the display the readings would go above and below 14.57 as the engine runs (not in PE) due to the fueling corrections of the PCM based off of the narrow band O2 sensors. This oscilliation of the AFR is based off of the "switching point" of the narrowband O2 sensors. This would affect your tuning if you were tuning using AFR and trying to offset the transfer function in hptuners so that the value in the scanner matches what the display on the controller reads.

Pg 12
Compatable Fuels
The AFX is compatible with fuels other than gasoline. The AFX’s display is designed to show AFR values based on a gasoline scale with 14.57:1 AFR as the stoichiometric ratio. The display will simply show AFR on a gasoline scale even though the fuel used may have different properties (stoichiometry for methanol is approximately 6.5:1). As long as you understand this, interpreting the values should be easy.

Assuming you are tuning the vehicle to 14.68:1 for a gasoline based engine how is the lambda value that AFX uses going to effect the results?

If you were tuning in lambda, then the AFR that the controller's display is reading is irrelevant because lambda is lambda. The AFR that AFX choses to display when the wideband is at Lambda has no effect on what Lambda is. My discussion with Brian was based on the proper transfer function to use in order to get an accurate reading in the scanner. Some people are using AFR and trying to to make the scanner's AFR match what the controller's display is reading.

The wideband can still see 14.68:1 ?

Not sure I understand exactly what you are asking, but the display can still show 14.68:1, it's just that the display is defining 14.57 as lambda instead of 14.68 which is one of the reasons that tuning and speaking in terms of Lambda is better.


Guys feel free to correct me if I'm worng...this is how I am understanding it.
-Charles

[eatgoat]

A W O L:
So when the wideband detects a value of 1 lambda than it will display 14.57. So I am understanding from this that the values passed between the wideband and hptuners is done in lambda and the transfer function is what calculates the lambda reading into the proper AFR? So offsetting it in the scanner via the transfer function will allow you to read the same AFR on the scanner and the wideband and not effect tuning?

once you go lambda, you wont go back.

i have my afx configed to display lambda in hptuners. i dont even pay attention to what the display is showing.

-carl

Going lambda sounds great any pointers on properly setting this up?

[A_W_O_L]

DSteck:
Thanks for the confirmation, I feel more confident now about the transfer functions thanks to you and Brian.

carlrx7:
lol, I've got the transfer function down...now I have to get all of my histogram straight.... I think I'm over thinking it...lol but I think I've almost got it.... About the time I start getting somewhere the program crashes... beta still needs some work...

eatgoat:

So when the wideband detects a value of 1 lambda than it will display 14.57.

correct, the controllers display will show 14.57 when lambda = 1

So I am understanding from this that the values passed between the wideband and hptuners is done in lambda and the transfer function is what calculates the lambda reading into the proper AFR?

Close...It's actually output from the controller as a voltage. The range is published as 0v - 5v. The transfer function is what the computer uses to convert the voltage to a reading we can more easily understand. Some choose to convert it to an AFR reading, some to a lambda reading....that's where the different functions come in.

So offsetting it in the scanner via the transfer function will allow you to read the same AFR on the scanner and the wideband...

Yes but after talking with Brian with AFX, this is not necessary for the reasons stated in the above post (see the 2nd half of post #9)

...and not effect tuning?

Again, according to Brian there's no need for the offset using the published values for the transfer function. After the transfer function is setup there is no need to look at the display. By offsetting the values in the transfer function so that the scanner matches the controller display you are throwing the numbers off some, but like stated in the 2nd half of post #9, the difference in 14.57 and 14.68 is small. The idea is to get as close as you can... keeping in mind that you are never going to reach a true 100% perfect tune because there are too many variables that you can't control (slightly different fuel blends, atmosphereic conditions, not so perfect sensors)... hence the use of long and short term fuel trims to make small corrections.

I think that because the difference is so small, you see some that choose to offset, some choosing to use the published readings and both ways work... except it seems that the ones using the published readings are posting that they are getting very consistant results. In the end I'd suspect that unless your transfer function has a serious error, you'd have a hard time proving one way works better than the other. After all the searching and reading .... I'm going to use the published info and tune in lambda... AFR gets to be a pain in the ass.

Everyone:
Again, if I have any of this wrong, someone correct me I'm trying to make this easier for us inexperienced guys... I'm sure some of the more experienced guys will chime in soon, but this is how I understand it.

damn long posts suck!

[07MontRedcp]

I think you guys are having to much fun with this. Good stuff.

BJK

[eatgoat]

A W O L:
Thanks for the great posts. This definitely cleared up my understanding on the matter!

[DSteck]

I'd stick to the 0-5V scale, because even though you won't ever hit 0V or 5V output, the meat of the voltage output will still follow that linear relationship, and that's the part you care about.

[A_W_O_L]

07MontRedcp:
Fun...umm...sure ... I think setting up the transfer function, histogram, charts, ect., in the scanner was more of a pain in the ass than understanding the tuning tables... the thing that was kicking my ass the most was the primay table and the histograms, but that was my own fault...see below.

eatgoat:
I'm glad that I could help someone else while I was trying to understand this stuff myself.... I kept thinking that it couldn't be that damn hard to setup!

DSteck:
I got to thinking about that after I got off of here last night....and I remember that testing TPS sensors and such that you never get 0v -5v either... you get close, but never a true 0v - 5v range.... I wish I would have remembered that sooner and saved myself some trouble....oh well... Thanks for the help!

Everyone:
Just a heads up... while trying to get my primary table setup in the scanner... I kept running into problems trying to get my calculated pid's to populate. The problem was that when the PID didn't work I would right click and make changes to the PID, saved the changes, and started logging data again, and naturally I wasn't seeing the results I was looking for. I figured out that I was not deleting the PID and re-inserting it back into the table, if I had deleted the PID made my changes then re-inserted it... that the PID would actually work like it was supposed to.... I had it right the 1st or 2nd attempt, but it wasn't applying the changes. I spent several hours trying to get everything to populate correctly and after I deleted and re-inserted it only took me 20 mins to get everything working properly. The sad thing is that I know better, it's just been soo long since I've messed with setting them up that I forgot.

[SilverG]

Can someone see what I am doing wrong? I have my afx in a new bung, hooked up directly to the battery with the yellow analog going to pin 1 and the brown going to pin 5 on the mpvi pro. I can not get the voltage or Lambda to report in the scanner. Also my AFR error % is way off from the LTFT & STFT. I have my low MAF pretty close by using the trims but I do not trust my AFX enough to do any WOT tuning. I have attached both my config and a scan. Maybe if someone could post a working scanner config. I really would like to get this nailed down tomorrow night as we head to the dunes with the truck on Thursday. TIA.

[SilverG]

I figured out getting Lambda and the voltage to display.

I still have a problem with AFR error %. I have disabled SD. If my fuel trims are +-2% in closed loop, when I switch to open loop shouldn't my AFR error % also be +-2%?