Best way to find WB voltage offset

[Matt_lq4]

Please describe your ways of finding the voltage offsets between the wideband monitor display and the hpt scanner readout.

My way is to start the vehicle for a few seconds then shut it off and compare the scanner readouts with the wideband monitor. Adjust the offset accordingly until both read the same.

However that's with the motor off so hopefully that's a good enough method that will give accuracy at wot.

[07GTS]

what i done was loop the wideband going into the hpt module from say pin1 also into pin2 then set the pin2 input to 0-5 volts and then u can see on your gauge also what volts reading is what AFR so u can use that to help setup the hpt scanner, what i done was use the volts in CL so its closest to lambda 1.0 went for a drive at cruise so it was constant and used that volts to set the scale for the WB readout, my WB is now pretty bang on 1.0 and the smallest shift either side shows richer/leaner on the 02's (i dont have a WB display only straight to hpt)

[mowton]

Just log wb in closed loop (add idle/decel and throttle positio filters to isolate only good steady state data) and the difference from stoich is you offset.

Ed M

[Matt_lq4]

Unfortunately from memory I recall the cruise and idle stoich on the gauge to never steadily say "14.68," for example but always jump around its usual few numbers at stoich CL such as 14.12, 14.33, 14.68, 14.3, 14.79, 14.01... would my method of gathering data with a quickly turned on and off motor be good enough? I'm basing my offset comparison on the residual exhaust gases.
Last night I looked at a time consuming comparison after a quick shutoff and it went from 11.6 afr all the way to 14 and I ended up finding an offset with identical data in the scanner and monitor. I'll test all the methods later on in the day but am curious on the results I'll get at OL PE in a 1:1 gear.

[mowton]

Unfortunately from memory I recall the cruise and idle stoich on the gauge to never steadily say "14.68," for example but always jump around its usual few numbers at stoich CL such as 14.12, 14.33, 14.68, 14.3, 14.79, 14.01... would my method of gathering data with a quickly turned on and off motor be good enough? I'm basing my offset comparison on the residual exhaust gases.
Last night I looked at a time consuming comparison after a quick shutoff and it went from 11.6 afr all the way to 14 and I ended up finding an offset with identical data in the scanner and monitor. I'll test all the methods later on in the day but am curious on the results I'll get at OL PE in a 1:1 gear.

You are best, in my opinion, to use the logging method as it will provide an average value. You can't always use the gauge display because on most of the units operating in AFR will display 14.68 regardless of the fuel type (E0, E10, E85 etc) which is why I teach to log in Lambda. Use the stoich value you have set in your tune as the expected and set the offset so your logged wb averages around that value.

Ed M

[Matt_lq4]

You are best, in my opinion, to use the logging method as it will provide an average value. You can't always use the gauge display because on most of the units operating in AFR will display 14.68 regardless of the fuel type (E0, E10, E85 etc) which is why I teach to log in Lambda. Use the stoich value you have set in your tune as the expected and set the offset so your logged wb averages around that value.

Ed M

I appreciate the help Ed. I just came back from a test drive with my new 0.609 lambda (in place of the 0.62 default for afr500 v2) offset correction and at WOT PE the values are extremely close now. Before it would read a little leaner in the scanner than the gauge and now it's as good as it will get. Tried in 2nd and 3rd gear, I'm happy with the results as I didn't see your method until after the test drive but it looks like I'm good to go now. Thanks to Ed and 07gts as well.

[Matt_lq4]

I always have my logs in lambda. So far I'm more use to reacting quickly to an afr value rather than lambda so it helps reading the display that way. On my monitor, to display lambda you have to disassemble it somewhat.

[dr.mike]

There is a problem that pops up a lot.

The voltage offset is usually a result of the high current draw that the WB uses to drive the sensor's heater. BUT, the heater current changes based on EGT and exhaust gas velocity. So, getting the offset at idle does NOT mean that it is right at WOT. It is PROBABLY not the same.

[Matt_lq4]

There is a problem that pops up a lot.

The voltage offset is usually a result of the high current draw that the WB uses to drive the sensor's heater. BUT, the heater current changes based on EGT and exhaust gas velocity. So, getting the offset at idle does NOT mean that it is right at WOT. It is PROBABLY not the same.

That was my worry however an actual full throttle pe triggered highway run confirmed that the majority of the values are accurately reading the same at wot. The biggest deviations were related to the comparison of a lambda value to afr.
0.81 lambda in the scanner may be 11.83-12.03 on the gauge as in the scanner a value of 0.815 will only show itself as 0.81 but the 1 will blink half way with 2 and the display shows 0.81 and 0.82 simultaneously.
So it seems like it depends how the programs do the math with their decimal place accuracy and rounding threshold to determine what to display.

[dr.mike]

Here is a mild example.

heater-load.JPG

The bottom (YELLOW) trace is the load on the sensor's heater. You can see how it reacts to the RPM (BLUE) and Engine load (RED) traces.

It is not uncommon for the heater load to hit 100% during WOT. That can cause a few hundred millivolts of offset difference in the AFR signal compared to what happens at load loads/idle, etc.

It is not predictable. So, the only way to get rid of it is to have a fully differential input or output signal, with ground offset compensation.

[Matt_lq4]

Here is a mild example.

heater-load.JPG

The bottom (YELLOW) trace is the load on the sensor's heater. You can see how it reacts to the RPM (BLUE) and Engine load (RED) traces.

It is not uncommon for the heater load to hit 100% during WOT. That can cause a few hundred millivolts of offset difference in the AFR signal compared to what happens at load loads/idle, etc.

It is not predictable. So, the only way to get rid of it is to have a fully differential input or output signal, with ground offset compensation.

So in other words, if there is an offset present at high load high rpm PE that's different than at cruise and idle, there is nothing I can do to the hpt pro mpvi and hpt program to correct it unless I could command multiple offsets to be present at different rpms/loads?

[mowton]

Understand the difference between idle and WOT which is why I recommend like 30-40 mph log scenario...depending on how you have your PE set, you can run right up to just before it engages and log the WB while Closed loop is holding stoich for the offset variance from idle, light cruise to moderate (just before WOT) cruise and with 3.X will have to average out the error as we don't have the LIST feature in the user defined wideband system we had in 2.24..

Ed M

[Matt_lq4]

My vehicle is officially in storage until spring but I will definitely do a 30-40 mph cruise then and see how my offset must change. I'll average out the different offsets collected, unless it's dead on with my current -0.011 offset.

[mowton]

My vehicle is officially in storage until spring but I will definitely do a 30-40 mph cruise then and see how my offset must change. I'll average out the different offsets collected, unless it's dead on with my current -0.011 offset.

cool....now that we are in Florida there is no "hibernation" :-)...I was thinking and if you are running in closed loop, then the calibration/accuracy of you WB could be shifted towards the wot area's and just have the closed loop correct the small error below. That way you can concentrate the corrections where they are the most important. If we had the LIST option, then you coulkd have inserted the non-linear taper across the entire range.

Ed M

[Matt_lq4]

cool....now that we are in Florida there is no "hibernation" :-)...I was thinking and if you are running in closed loop, then the calibration/accuracy of you WB could be shifted towards the wot area's and just have the closed loop correct the small error below. That way you can concentrate the corrections where they are the most important. If we had the LIST option, then you coulkd have inserted the non-linear taper across the entire range.

Ed M

Oh Florida would be nice, maybe one day I'll live somewhere that's not an igloo for 6 months

Yes I truly only care about WOT accuracy of my wideband as I'm in closed loop the majority of the time due to my blended tune with maf/ve. I went for the upgraded Ntk sensor with my afr500 version 2 instead of the base Bosch sensor or whatever they come with and I calibrate my sensor/monitor once a week to make sure all is well. It's not the lab grade sensor but calibration grade I believe.

I analyed the heck out the cell phone recording of my wideband monitor (no fixed interior camera position yet) during a specific WOT run and compared it to my scan exactly when PE WOT was triggered and it is eerily close.
I did this from a 2nd gear roll and a 3rd gear roll from 3000 to 6000-6500 rpm.

I know there are doubts but I may be the lucky one as the numbers seem to match up as good as converting between units can get (depends on the threshold of what to display based on decimal places unless my monitor read in lambda like the scanner).

But I know what I have to do in spring, have a camera record my laptop scanner reading and also cover what my wb monitor reads simultaneously...then I can pause and compare as that will give me the 100% undeniable truth at WOT.

[Matt_lq4]

This all stemmed from the last track day when I was just passing the beams in the 1/4 mile in 3rd gear around 6000 rpm as my wideband monitor displayed 12.68 almost every run that day.
However each scan showed that my scanner was reading 0.88 up to 0.89 sometimes at the same point.

Yesterday on the highway, I confirmed that at that same 5800 to 6000 rpm point in 3rd gear, my scanner now reads 0.86 lambda (since offset correction) and sometimes blips to 0.87 as each lambda point covers a broad afr. So 12.69 afr on the monitor is 0.865 lambda so the 6 and 7 in 0.86 and 0.87 quickly interchange with each other. I added 3 decimal places now to get a less rounded lambda figure and all still checks out fine.