Thread: Calculating MPG from VSS and MAF from OBD2

Yeah, my Wideband sensor (LC1) is plenty responsive enough.
My software is actually capable of displaying both my Innovate LC-1 output (through the Serial port), as well as the output from the OEM ECU.

As for how I used the ECU value, the SSM protocol provides an actual Lambda value. From there, it's easy to convert to an Air/Fuel ratio, provided you know the type of fuel. I'm not using the Voltage output from the OEM one, only the ECU Provided Lambda value.

I think I've already posted this before, but here is an early, short video of ECUTracker in action. You can see the LC-1 wideband output on the left, and on the right, until I change pages, you can see the OEM Wideband (AF) value too.


[media]http://www.youtube.com/watch?v=YCN8sj_zMiM[/media]

Compare the time for the stoich reading (voltage) to the time to balance it back to stoich ideal.
Isn't it obvious - the narrow band sample or sensing time (or rather, reaction time) is the same as the wideband, but whereas the narrow then puts out a voltage, the wide has to measure the amount of O2 to reach stioch neutral - and that takes time.

Hence while widebands are still based on "narrow +", the narrow will always be faster.
The exception will be if the narrows "freeze" at current (no pun) technology because the wides are the way to go, hence the wides continually get faster.


But lets say for argument's sake that O2 sensors are fast enough and able to measure EVERY stroke (don't forget wasted strokes!), then you have pressure and temp to measure as well.

Do it by all means, I still reckon the flow else injection volume is the most accurate. Flow is best where as injection volume needs injector profile, but that can be a continuous function. O2 sensors also need to be continuous (which they had no hope of last I looked) plus the pressure/temp profile.

As you wrote, even 1/10th of a second is a mere 600RPM - assuming a 1/2 cycle for a one-cylinder engine is good enough for your economy readings, or one reading every 15 ignition strokes for a 6 cylinder doing 3,000RPM.

The only alternative I know of is ionic sensing as that gives the combustion profile for every ignition for most of the ignition burn. It still misses wasted injections though....

Originally Posted by OldSpark

But lets say for argument's sake that O2 sensors are fast enough and able to measure EVERY stroke (don't forget wasted strokes!), then you have pressure and temp to measure as well.

Do it by all means, I still reckon the flow else injection volume is the most accurate. Flow is best where as injection volume needs injector profile, but that can be a continuous function. O2 sensors also need to be continuous (which they had no hope of last I looked) plus the pressure/temp profile.

As you wrote, even 1/10th of a second is a mere 600RPM - assuming a 1/2 cycle for a one-cylinder engine is good enough for your economy readings, or one reading every 15 ignition strokes for a 6 cylinder doing 3,000RPM.

Alright, I misunderstood what you were saying. I agree that it is not nearly fast enough to measure fuel for 100% accuracy, fuel flow is really the only way to do that, but few sensors are sub-$100 for that. There is no perfect world of course, so you'll never get 100% accuracy, but the question is how inaccurate are you willing to accept?

Plus, I'll happily trade off a tiny amount of speed for resolution.
A narrow band is only really accurate right near Stoich.
Where as a Wideband can be easily accurate between 10:1->20:1.

So, if you want to adjust your MPG calculation with an AF reading.
Only a wideband will give you what you want.

Neither O2 sensors are accurate enough for my liking.

I'll stick with the more accurate injector pulse - even if temp and pressure are ignored.
Besides, it's much easier to calculate. (I prefer simpler solution - especially if better results.)

Originally Posted by Skidd

http://www.youtube.com/watch?v=YCN8sj_zMi

Skidd could you please check the link? I can't open it.. So can you connect the LC1 with your mobile phone?

Oop.. Sorry about that.
It's fixed now
[media]http://www.youtube.com/watch?v=YCN8sj_zMiM[/media]

Oh yeah, and in theory, yes.
However, right now, I'm connecting via an In-Car-PC.
But, with a DB9 Serial to Bluetooth Adapter, I "should" be able to connect to an LC1 with my phone. I just have not had a chance to grab an adapter and actually try it out.

Originally Posted by OldSpark

Neither O2 sensors are accurate enough for my liking.

I'll stick with the more accurate injector pulse - even if temp and pressure are ignored.
Besides, it's much easier to calculate. (I prefer simpler solution - especially if better results.)

What's your formulae using injector pulse width/duty cycle?
I'm looking for a l/100km value (not mpg).