Lifetime MPG reported in Car is wrong! Anyone Else tracking their actual usage?

[tboult]

I was trying to figure out what seemed to be going with my data at voltstat.net (my MPG CS keeps dropping even though there is no CS usage). After seeing what Mike is doing I wanted to understand why/how its wrong.

So I pulled all my gas purchase logs (and checked with the dealer on their usage, it had ~100 miles when we bough it) and still could not make it work out.

Yesterday while reading an older edmunds post about the prius and others in their fule sipper tests I saw the following:

.... the onboard fuel economy gauges also differed from reality. While the others ranged from 1 mpg better than reality to 3 mpg worse, the Toyota Prius was woefully inaccurate. On the back-road route, the Prius' computer claimed it was getting 6.8 mpg (14 percent) better than it actually was, while the city yielded a 5.4-mpg (11 percent) over-report. This sort of shenanigan could mean a higher customer satisfaction score from owners who think they're getting better mileage than they actually are. .....

And that got me thinking.. maybe the problem is not voltstats or the algorithm or my analysis. Maybe the car is lying. (I track CS and EV miles separately so did not think to check the overall MPG. )

Sure enough the car is inflating the lifetime MPG; the data on the center council and via iphone and volstat is wrong. The logs show the car has used 25.2 gallons of gas, and the odo says it traveled 2037 miles. So the lifetime should be 2037/25.2= 80.83 But the readout in the car says my lifetime mpg 83.5. If the MPG was really 83.5, it would mean only 24.39 gallons, or a difference of .8 gallons which is well beyond any filling station erros, and Trip A shows more gas used that that! so parts of the car knows its used more gas than its lifetime MPG suggest. I have not found any way to back into the cars MPG computation. Note this computation is not even looking at actual distance vs reported distance (which tires or something might impact). I'm computing based on actual gas vs odo miles.

Maybe its just my volt or my errors. But given Mikes comments that on voltstat.net he is seeing what he though was an initialization error, I'm thinking this is much broader.

So the question is.. does anyone else have enough data to see if this is a broader issue.


Maybe the car is not counting gas used while the car is not moving? I make this hypothesis because I believe the dealer burned gas while it was parked

You might be thinking, why would a dealer run the gas engine for more than a min or two (it starts if you open the hood and its on). I realized I already had suggested this to another dealer. MM can be a faster way to charge the battery for a car that is being demoed, much faster than even a level 2 charger. 15-20 min of MM idle and you have enough CD range to show off the car. I learned about MM that on this list. I was very disappointed when I took my first volt test drive at a local dealer and the volt was in CS mode. When I asked why, the dealer said they had not plugged it in and where they have the charger in the service bay is not a good place to show off the car and somewhat in the way, so if they forget to move it and plug it in at night, they just demo in CS mode. I told them they can always just run in MM mode on the lot and then let people experience EV driving. He asked me to show him, so I switched modes and did about 5-10 min of MM mode during the rest of the demo drive an when we returned to the shop, turn it off then back on it showed 3 bars, enough to for miles of EV demo for the next guy person. I also suggested he setup the phone app to email if they forget to charge (but I guess that still does not work).

So do MM miles while not moving count to lifetime MPG? I someone knows the answers please let us know.

I'm guessing someone that has used it on a cold climate may know from heating pre-start which may be the same. Otherwise, I'll try an experiment when I get a chance on a trip that needs some MM anyhow. At least I can answer if the non-moving gas usage is counted.

dep_6269899-Fuel-Economy-Speedometer-Measures-MPG-Efficiency-in-Car-or-Vehic.jpg

[stephent]

Yeah, I don't know what's going on there. My Volt's pretty new, so there just might be weird things going on at the beginning. But anyway, after my one and only gas trip, I burned around 2.9 gallons, and haven't done anything but EV for a week. The gas reported in the tank appears to be stable, bouncing between 5.62383 & 5.3682 left in the tank (got tank from dealer @ 8.47226, with .15 gallons burned and ~2.3 non-EV miles).

But right after my trip, the Volt was reporting only 2.5 gallons consumed total, i.e. that my trip only took 2.35 gallons not 2.9. For 49+ MPGCS on the site which is a big error. But over the past week it's been creeping up, now it's up to 2.9 and has almost reached the real amount. I'm wondering if it will eventually go past the actual amount of gas consumed.

Could people with way more miles on voltstats but not currently doing any ICE driving chime in?

[rubley]

I have to say, I don't really care about the lifetime MPG reported by the car. Mainly because its a made up number, just because you can divide the mileage by the gallons used doesn't make it a "miles per gallon" figure in the way that most people understand. I'm at about 615 miles and 2.5 gallons used, you could divide those two numbers and say I'm getting 246 MPG, but that's a meaningless figure. If I add 1 gallon of gas to the tank my range doesn't go up by 246 miles.

I'm much more interested in gas mileage when the engine is actually running, and I calculate that myself at the end of a trip based on the mileage reported for that day - and I'm getting 42 MPG (I've only been in CS mode 3 times so far). If someone asks me what kind of mileage I get in the car I'm not going to say "246 MPG", I say "41 miles avg on a charge, and 42 MPG when I'm burning gas".

[WopOnTour]

The dealer may (if they followed proper procedures) have manually "reset" your lifetime average just prior to delivery (with whatever miles showing on the odometer).
This PID is also automatically reset after certain software updates. (in either case displays 250+mpg)
So in a sense "lifetime" average is somewhat of a misnomer as it wil only refelect the most current "lifetime" (since it was last reset)
HTH
WOT

[tboult]

I have to say, I don't really care about the lifetime MPG reported by the car. .
....

Agreed, which is why I had not bothered to see if its correct, (other may care though).
More important for me, I wanted to understand what is going on in the log data (from voltstat) because if that is good then I can simplify my tracking of data manually. The car apparently does not report MPGcs, rather tis computed from data that is making it wrong (why Stephen's data is probably not making sense to him.. that is where I started as a new Volt owner).

Rubley, I'm guessing you are tracking by hand. I'm trying to automate my processess so I can track well but with less effort. But because of the onstar reports data, (EVmiles and totals miles and Lifetime MPG, ) on can estimate MPGcs, but there is not direct MPGcs report. Also, if there are problem in the data it may also impact myvolt.com data (which at least or me has even bigger problems right now) and so fixing them before more volts roll out would make sense.

[tboult]

The dealer may (if they followed proper procedures) have manually "reset" your lifetime average just prior to delivery (with whatever miles showing on the odometer).

Thanks WOT.. a plausible explanation. I can see if that is the case by measuring it again after more CS mode and seeing if the error is a constant offset. If so then a recalibration process can fix the voltstat data :-)


I would not care at all, if the car actually reported MPGcs. Will the fancy meters you have report other data (like lifetime gas used...) Data is also useful to study driving effects. Instantaneous "leaf motion" is not very precise and even if there was instant MPG/KWh (which I do really want), can only see them when your not looking at the road and hard to "average" over enough of a trip to see the real impact. But the data as it is makes it nearly impossible to use.

[stephent]

Yeah I agree with rubley that lifetime MPG by itself is a meaningless number. I'm trying to get an accurate MPG-CS, without as much manual record keeping. BTW for all you Android users out there, there's a very nice app called FuelLog written by Simon Morgenthaler. It lets you keep track of multiple cars, and best of all it supports electric vehicles, in kw-hrs consumed and displaying as kw-hr/100 miles, or miles/kw-hr, or some other options. I have my Volt set up as 2 cars, one gas, one electric.

The voltstats downloads with OnStar data appear to be fairly accurate with regards to gallons in the tank, total miles, and EV miles. It's just that "lifetime MPG", which is supposedly (total miles)/(gallons consumed), seems to be weird. In theory, if you haven't gotten your car reset, (total miles)/(lifetime MPG) should equal gallons consumed, and if you are doing just EV driving this number should remain constant. But at least for my car, and tboult's, this is not constant despite doing no gas driving. And also in theory, if you do CS driving, the change in gallons consumed by dividing total vs. lifetime MPG should be equivalent to the drop in the level of the tank, at least to some level of accuracy. In my case it was off by 0.5 gallons out of 2.9, although the increase since then during just EV driving has almost equalized this. These theories ought to hold despite any lifetime MPG resets.

[rubley]

BTW for all you Android users out there, there's a very nice app called FuelLog written by Simon Morgenthaler.

Thanks for the tip!

[tony6777]

I have never reset my trip counter so it has all of the miles on it through the lifetime and calculates a MPG. It does not agree with what the lifetime is in the center console. I go by what the trip counter says. It is more accurate. Sorry, I'll have to look up the actual numbers later when i'm in my car.

[Jim I]

Just a quick thought:

Don't forget that the volume of gasoline in your tank can vary quite a bit based on temperature. When you pump gasoline into your car, it is cool, as it has been underground. As it warms up, the volume will increase in the tank. The oil industry defines a gallon of gasoline at 60 degree F. The volume will change up or down by 1% for each 15 degrees in temperature change.

This could also be why what you have logged at the pump is not in sync with what has passed by the sensor.

JMHO