MPG accuracy / calibration vs pump MPG

[thewanderer]

The VSS seems to be perfect between the MPGuino and my 92 Civic VX, but the MPG reported via the MPGuino seems to be off considerably from what I measured at the pump...

I put 309.7 miles on this tank (MPGuino states 309.72, so that matches pretty well) According to the pump, I put 7.719 gallons in the car and on the MPGuino, it states 7.70 gallons.

But the issue is... 309.7 mi / 7.719 gal = ~40.12 MPG (pump) and 309.72 / 7.70 = ~40.22 MPG (what MPGuino should report) but the MPGuino is saying my tank MPG is 50.25 MPG, so it's off by a little over 10 MPG.

I had the uS/Gal set to 258562433 (I saw someone on the forum mentioned that number) so following the guide, I did (258562433*7.7)/7.719=257925992 and set the microsec/gallon to 257925992. But the small amount the gallons were off by (0.019 gallons) seems like the microsecond value was close... but I don't understand why the reported MPG is so much higher?

Am I missing something as to why there's such a discrepancy between MPG but not gallons or miles?

[AndrzejM]

I have the same issue, different pumps shows results that are way off. 10-15% off. Which difference is huge IMHO. Unfortunately pumps are set up with not the exact accuracy so you can put like 7 gallons of fuel and the pump will show you 7.2 I've read somewhere that here in Poland pumps accuracy have to be in +/- 10% range. There's no way to check the amount of fuel that you're pouring into the tank unless you'll use a canister and some other measuring device. That could be an interesting experience anyway

[thewanderer]

10% would be a considerable improvement for me. The MPG the mpguino reports is high by 25%. The mpguino reported my mileage traveled as 309.72 and fuel used as 7.7 gallons. Doing this math by hand would give you 40.22 MPG but the mpguino states 50.25 MPG. Since the mpguino has accurately reported the mileage and fuel, I don't understand why it's so off on the MPG.

I don't have the numbers on hand, but my second fill up was about the same as far as the MPG being way high but the mpguino miles matched my odometer and the fuel used matched my pump receipt.

Is there some setting I have incorrect or could there be something wrong in the code itself?

[AndrzejM]

I would say that there's nothing wrong with the code... To be honest i would trust MPGuino more than any pump... Especially here in Poland... Sad but true...

[thewanderer]

Originally Posted by AndrzejM

I would say that there's nothing wrong with the code... To be honest i would trust MPGuino more than any pump... Especially here in Poland... Sad but true...

The part I'm confused about is the mpguino doesn't seem to match its own values. If you completely ignore the receipt from the pump and just use the miles traveled and gallons used that the mpguino logged, dividing that out doesn't come close to the stated MPG.

Mpguino says 309.72 tank miles and 7.7 gallons. 309.72 / 7.7 = 40.22. Mpguino tank MPG calculation says 50.25.

Is there something else that needs factored in?

[njakts]

Originally Posted by thewanderer

The part I'm confused about is the mpguino doesn't seem to match its own values. If you completely ignore the receipt from the pump and just use the miles traveled and gallons used that the mpguino logged, dividing that out doesn't come close to the stated MPG.

Mpguino says 309.72 tank miles and 7.7 gallons. 309.72 / 7.7 = 40.22. Mpguino tank MPG calculation says 50.25.

Is there something else that needs factored in?

Hi! If I remember, I noticed the same problem some 3 years ago. If I calibrate mpguino by average readings, averages are precise since then, but tank litres (yes, I used metric version) are way off, and if mpguino is calibrated by consumed litres, litres will be displayed correctly, but averages wont...

I solved the problem my multiplying instant, current and tank readings by a constant (I think it was 1.2 or 1.25) and calibrated mpguino with consumed litres and averages were correct too. I dont remember why, but multiplying only worked in Instant/current, instant/tank, and big screens, but in current and tank displays, average readings were a lot lower than they should be, but litres were correct. since I always used mpguino with instant/current display, it did not matter.
Now I want to update my mpguino with improved code found here. I tried alin151 code, but it has some strange bugs and now I will try t_vago's code as recomended here...

good luck

[AndrzejM]

There was such issue, I remember that but that was with the "old" code... That's why I'm using t_vago's code, the best one on the market
The difference with the previous code was IIRC due to clock setting. 16MHz vs 20Mhz that gives you 1.25 factor.

[thewanderer]

Mine is 16MHz but I did comment out the "use with 20mhz" block and uncomment the "use with 16mhz" block. Do I need to do something else as well?

I also used the source at http://opengauge.googlecode.com/svn/...no/mpguino.cpp which is what the wiki said was the "official" code. Is there better code to use?

[firehawk618]

Originally Posted by thewanderer

Mine is 16MHz but I did comment out the "use with 20mhz" block and uncomment the "use with 16mhz" block. Do I need to do something else as well?

I also used the source at http://opengauge.googlecode.com/svn/...no/mpguino.cpp which is what the wiki said was the "official" code. Is there better code to use?

Subscribed.

What you say is strange, MPGuino miles / fuel used being way off from what it displays.

I haven't gotten a tank through mine yet but have a feeling MPGuino is going to be very optimistic in my tank MPG calculations.

I really doubt I'm averaging 50 right now.....

Curious to see what you figure out.

I ended up installing a pre-assembled unit that I had purchased. I can tell immediately the calculations are different and seem MUCH more realistic vs the code I had been running.

I was running the one that's on the wiki page, but as AndrzejM pointed out, that code is very old.

I have to wonder if that's what your issue is?

[stewie]

mpguino being off all the time with 0.86 code was my experience as well.
I needed to contestant adjust values cause every time they were off which made the instrument useless in my humble opinion.

Andrzej how is new code working for you? Andrzej has this improved with new code and how so and what made the change?