First, if this is a repost, my bad. I don't read most of the stuff posted on this forum.

What I do know is that everyone knows the MPG Meter is off. I decided to further inspect it since I'm an engineer, and that's what I do.

Rather than looking at how far off the meter was after one tank, I looked at it across 3 tanks. I think I can draw conclusion that that's enough and that it's not necessary to average across more tanks or gallons. Here is the data in a per tank format:

Miles | Gallons | MPG Average | Meter | Delta (meter-actual)
300.3 | 9.926 | 30.2538787023977 | 33.9 | 3.64612129760225
317.2 | 9.992 | 31.7453963170536 | 35.0 | 3.25460368294636
339.5 | 10.17 | 33.3824975417896 | 36.1 | 2.71750245821042

Here it is again in cumulative terms:

300.3 | 9.926
617.0 | 19.918
957.0 | 30.088 | 31.8067003456528 | 36.1 | 4.29329965434725

You can see I averaged 31.8mpg across the 3 tanks. If I average the Meter's numbers at each fill up, it says I averaged 35MPG although the final meter showed 36.1MPG. So we are not seeing a running average from the meter as shown by the delta of 4.29 across the 3 tanks and delta of 2.72 from the last tank. I drove much more conservatively and on the highway on the last tank. That being said, already having 600 miles on two tanks, I'd have to had averaged nearly 40mpg on the last tank to move my meter up to 36mpg. You can see I only averaged 33.4mpg on that last tank.

It appears to me that the meter is showing something closer to how you are currently driving, say over the last few days, as apposed to how you drove across an entire tank, or how you drove across several tanks.

Does anyone else draw a different conclusion from this?

Basically what I learned is that the meter is actually more worthless (less worthwhile?) than we had though.

Feedback welcome.

 

Maybe you should try doing the MPG meter software upgrade at your local dealer, then collect more data.

 

1. Honda issued a TSB for a corrective software upgrade.

2. After having the software corrected, I have to say that the mileage meter is usually more accurate than tank-by-tank calculation.

Unless you drive until the tank is almost dry, then fill until you can see gasoline in the filler neck (neither practice is recommended), with a 10-gallon tank and a long fill path from cap to tank, it's extremely easy to have +- 5% tank to tank variations in actual fill. If on one tank I get a calculated MPG that exceeds what the meter shows, at the next fill the calculated mileage is almost always less than what the meter shows.

Averaging meter MPG is mathematically meaningless, unless you drive exactly the same number of miles between fills. For example, if you make a 5-gallon fill at 40 MPG (according to the meter) and a 10-gallon fill at 30 mpg, the meter average will be 35 mpg, while calculated will be 33.33 mpg (15 gallons for 500 miles).

 

There are three or four threads on this. I think this is the definitive one. Link

 

Yes- please merge this with the other threads. Merge them ALL. It is very frustrating reading multiple threads about this- and they all seem to bump up at the same time.

Dan

 

My meter is accurate on the 2010. The difference is less than 1 mpg (meter vs manual calculation).

 

there is a software upgrade ummmm what do i do to get it lol is it free?

 

If there was a need for a sticky I think it would be for TSBs and recalls (no recalls to my knowledge). Of course explaining how to program your door locks is much more fun.

Crash001: take a look here <-click me!

 

Your conclusion makes sense to me, at least for the pre-software update mpg meter. I haven't had the update, and you never know if they change the way the average is calculated (running total vs. lopping off old measurments periodically).

I've a little over 6000 miles on my car now, and I noticed a steady increase in my meter's reported average mpg over the last 2000 miles or so. Considering I've never reset the meter (not even sure how) I wondered how it was possible for the running average to increase by .7 - .8 mpg over merely 2000 miles. The only logical conclusions are that 1) my actual mpg went up like a rocket or 2) the time (mileage) over which the mpg is being calculated is being chopped off the left end periodically (discarding old measurements either periodically or continuously).

I'll read the manual to see how to reset the meter. If there is no change in the meter's reported mpg, I'll conclude that it discards old data. If the reported average goes up I can only conclude that my actual current average mpg has gone up, but I won't be able to conclude if/when the meter discards old data.

BTW, this thread is about how the mpg meter calculates its inputs, not where the inputs originate nor "OMGZ my MPG thingy is wayyy off!!1!"
I think to merge it into one of the other threads would not be a good idea.

 

The average mpg display resets when you reset the trip mileage to zero (select and hold for a few seconds). Afterward when you view avg mpg it will be "dashes" until you drive a bit.

I obsessively reset it every tank (with the new software flash) so I don't have a feeling for what it does over extended miles. I would think it stores two values: distance and fuel consumed and displays the mpg. The distance or volume of fuel shouldn't matter (unless it's to Mars and back). I think the problem with the old software was the calculation of fuel consumed. The TSB mentions an incorrect 'data set'. I don't think they could get "miles" wrong; while it was correctly metering fuel to the engine, its interpretation of volume was wrong.

 

 

Since the update, my meter has still been consistently high, just less so.

Date | Miles | Gallons | MPG | Gauge | Diff
12/25 | 291.6 | 8.132 | 35.9 | 35.9 | 0
12/29 | 314.3 | 9.225 | 34.1 | 35.2 | +1.1
1/3 | 308.7 | 9.804 | 31.5 | 32.2 | +0.7
1/9 | 275.7 | 8.825 | 31.2 | 32.8 | +1.6
1/16 | 308.3 | 9.336 | 33.0 | 33.9 | +0.9
1/24 | 332.5 | 9.551 | 34.8 | 35.9 |+1.1

Every tank except for the first one has been high. Clearly, it's not as high as what it used to be (the normal 3-5 mpg high that everyone else got), but it's still not right on my car. After 6 tanks, this couldn't just be random error or filling inconsistencies. Those would have to even out at some point, and they aren't.

Basically, it seems that instead of the old 10-15% error, now mine works at about a 2-5% error. But I certainly wouldn't call what I've seen a small +/- some percent error that other people seem to be getting, because mine never evens out with a low readout on the meter.

 

yeah, consistently 2-3% high on mine four or five tanks after the flash. Any inconsistencies on fills would be apparent in this length of time. At least it's not off in la-la land.

 

And all this time I thought it was just me. Mine is still about 1 MPG optimistic tank after tank since the upgrade, but, that's better than 5 MPG off before.

 

same here since the software update:
monitor is always higher than hand calculated:
+.6
+.5
+.9
+.1
+.7
+.4
+.2
+.7

vast improvement... not perfect.

 

If that's true then it's the definitive answer to the question originally posed in this thread. The meter only resets the input data when you reset your trip meter, meaning the meter is not worthless. You just have to understand the period of time being measured, which is under your control.

fmcfad01, did you test before or after the update? The update likely changed a coefficient that could have been artificially inflating normal fluctuations in calculation accuracy (e.g. you can't always perfectly measure the amount of fuel you put in the car).

 

This is pre-update.