Completely unexpected MPG results
 

In advance, sorry if this should go under "Instrumentation" but that forum looked completely dead.

So I just bought a ScanGauge for my Honda Accord and I had used it for the "Daily" MPG tracking features prior to first fillup, and it seemed consistent with results I had previously been getting without instrumentation. Today I filled up the tank completely and followed the instructions for calibrating the device to a full tank of gas. I also reset my trip meter since I've noticed the SG does not recognize coasting with the engine off.

One shift of food delivery with Uber Eats runs me 52.0 miles. I ecodrove the hell out of that car, especially considering that I was also working. SG reports 31.5 MPG for the tank which is slightly above average for me, but I figured it was pretty accurate now that I can monitor the MPG in real time.

So I go back to the same gas station, same exact pump, and somehow the car manages to take 3.43 gallons to fillup. Barely 15 MPG.

The numbers make zero sense to me, can someone explain what might have happened? I filled to the first "click" each time as SG instructions say plus I really don't want to screw up the EVAP system. I don't think its possible to have gotten 15 MPG.

Short fills are wildly inaccurate vs filling at 1/8 tank or so.

Over time, with many data points, a more accurate picture will be painted. But in the short term there will be wild variations in spite of efforts to use the same pump, same click, and all that which should still be done.

It does not really matter whether the gauge is off or not.
If it is off, it will underreport or overreport by the same margin every time. You'll still learn what works and what doesn't.

Variations in fill-ups can be huge. I've had differences of 7 liters (judging by the point where the fuel needle leaves the F), that's almost 2 gallons.

Fuel pump readings are useful as they give an accurate reading of the amount of fuel you tanked (irrespective of whether your tank is brimmed or not), and will over time become fairly accurate in aggregate.
After several fillups you can use it to recalibrate your gauge.

Tank-to-tank data is not very useful, and partial fill-ups increase the margin of error.

Short fill. Also try calibrating you SG using a 3 fill average.

About the only way to be reasonably accurate with short fills is to top off the tank. I always fill the tank until gas is visible in the filler neck and my mileage still varies by a few MPG sometimes but there's going to be a certain amount if variation from tank to tank simply because of weather, terrain, etc. In most cases with my Versa I can get between 1-2 gallons in after the first click. On the Escort it's probably at least 2 gallons. Things like pumping speed can effect when the pump will click off. At the station I use I've noticed that the pumps will pump at different speeds depending on how many people are using the pumps at the time. The faster the pump is pumping the earlier it will click off.

the slope is the problem maybe a couple of inches it might have vapor locked quicker or later causing it not to fill up correctly

you can fix this by lining up the edge of your mirror with a fixed object this is what i do I even do it when i park my SUV so i don't crash into the curb when parking.. as not to damage the aluminium suspension in the front(even if you hit it slowly it could damage it or pop a tire)


also wait a min for the sloshing of the fuel if it's close to being full..


I all ready tested this as long as i fill up at my "home station" even if i filled up and only drove 70 miles it has very little effect on the results..
(as sometimes I can get CHEAPER chevron gas or mobil elsewhere i can save an extra dollar or two doing that)




overtime it will have very little effect on average MPG. my averaged dropped to 22mpg as I have only been driving City only..



it sucks i have to drive slower since the ground is wet and my tires are pretty worn out I rather spend a few $ on gas driving slower then have to pay thousands for damage repairs if I was to spin out or something



$ is better then spending $$$$

filling
 

Automakers discourage owners from 'topping off' modern cars due to potential harm to emission control hardware,and without topping off it's impossible to know if the tank is 'full.' Any difference is ambient 'soak' temperature only complicates the issue,on account of the 'expansion bubble' over the fuel,designed into the tank.
Short of having a removable racing fuel cell,which you could weigh,and go by fuel mass,we're all kinda SOL as far as accuracy goes now.It's very frustrating.

For a single tank, sure. We don't know when exactly it "clicks". But over a few tanks it doesn't really matter. If the tank fills a little short, it'll fill it a bit more the next time and average out the same.

Of course this is only as good as the pump's meter. There are regulations on the accuracy and testing of that.

It may catch up with me tomorrow but I've been doing it to all my cars for years and hundreds of thousands of miles without any damage.

next time
 

Rather than make a sweeping,broad-brushed,generalized rule,I'd want to say 'it depends.'
I'm filling my truck 6-times a year.There are variables in flux during that time frame,which might not sort themselves out for an entire year.The fuel itself wouldn't even be the same for the duration.
Driving a single mod at exactly the same condition for an entire tank full would be one thing.Driving various mods for 50-miles would be another animal.Overnight temperatures alone would skew data.
The only way to know is to test.Other than that all we can claim is informed speculation.
Some station owners can tell you the current temperature of the fuel underground or above ground.A stainless steel 6" stem thermometer,placed in the filler neck will tell you about the fuel delivered.These thermometers can be calibrated in an ice bath.You'd need to know the temp of the fuel already in the tank.Stratification.The coefficient of thermal expansion for motor fuels is available online.Dry bulb temperatures should be monitored during testing.Along with coolant,transmission,differential,tires.Rain or snow.Wind,wind direction.Topography.Road curvature.Road surface.Road temperature.
According to the SAE,your vehicle needs to be warmed with 50-miles of driving at 50-mph,before testing ever begins.