MPGuino metric version bugs?

[njakts]

Hi!

After using MPGunio for a while I found a problem. VSS is calibrated and works well, but consumed litres and average consumption does not match. If I calibrate it using consumed litres, average consumption is off, if calibrated by average consumption, litres consumed are off...

also there is a difference between instant consumption, current and tank average. I hooked my mpguino up to a signal generator at home to have constant speed and injector input and after reseting tank and current and after `driving` for a while I collected the readings and this is what I got:

cruising at 100 kilometers per hour:
instant consumption - 12.66 l/100km
tank average and current - 10.27 l/100km

after 10 kilometers consumed fuel showed 1.27l, so that means instant cons. is correct.

why are current and tank averages incorrect?

Is that a calculation mistake in the code from converting mpg to l/100km?
Thanks

[Quezacotl]

I think the average is smaller because it drops bad signals. I remember reading that it does that. I don't remember how picky it is about that. And same could happen with your car. (Inj delay time needs tuning)
Or other could be because you might be using 555 as signal generator? it is not very accurate.

But are you using the old metric "canadian" unofficial version or the current version supporting metric?
Could you do the tests with MPG and see if it does same than metric?

[njakts]

Thanks for your reply.
At home I am using laboratory grade signal/function generator so signal should be good and clean enough.

I am using v0.86 with metric and imperial option. I did some calculations and the difference stayed +/- constant -1.25. I have not tried comparing MPG's, but I solved L/100km current and average difference by multiplying current.eocMiles() and tank.eocMiles() with 1.25...

Also adding ''digitalWrite(6, HIGH);'' line where backlight is turned off after timeout turns display off (not only backlight).

sbi(TCCR0A, COM0A1);//contrast pwm enable
OCR0A = parms[contrastIdx];

turns it back on.

[AndrzejM]

I think I know what's the issue. As you said

I am using v0.86 with metric and imperial option

and MPGuino is counting mpg US and not Imperial. And as you may check Imp to US factor is about 1.20 so that can be the reason for the "problem" you've found.

[njakts]

sorry, I meant US not Imperial...
but anyway, what difference does it make? If instant cons. and litres are correct, but averages are off?

now I cant figure out how to add custom chars. I want consumption instead of ''LK'' show like OBDuino L/100km :

I found where they are created, but how to call them in place where LK is ?

[AndrzejM]

Originally Posted by njakts

sorry, I meant US not Imperial...
but anyway, what difference does it make? If instant cons. and litres are correct, but averages are off?

Right, I have to take a look at the code...

[AndrzejM]

I've found interesting line in the code #953

// init64(tmp2,0,1250);

Maybe that's the reason for misreading? Try to uncomment it and check if the readings will be accurate.

EDIT: It won't help... I think it's the dispadj & dispadj2 parameters that we have to taka a closer look at.

[nickdigger]

Originally Posted by AndrzejM

I've found interesting line in the code #953

Maybe that's the reason for misreading? Try to uncomment it and check if the readings will be accurate.

EDIT: It won't help... I think it's the dispadj & dispadj2 parameters that we have to taka a closer look at.

I believe there was some confusion in the code, with those two parameters. When dcb switched to 20mhz, he only tested the code for that clock speed, but the parameters were untested for 16mhz users. From my recollection, one of the values was wrong. If you search the forum here for 1250 or 125, you should eventually find a post with the answer.

[AndrzejM]

Recently I've installed MPGuino in my Berta and it seems that there is actually the problem with instant vs current/tank FE.
I'was using 0.86v code and the problem was there, I've switched to "canadian" metric 0.81v code and everything is just fine.

I have to investigate 0.86v code further to find the bug, for now I'm using "canadian" version, but it needs calibration.

Soon I'll post pictures and description of my MPGuino installation process, it's slightly different, because Berta has a diesel engine