How much diesel left when the display says 0 miles left? And accuracy of MPG calcs.

[kgall]

On an Audi A-3 TDI,
how much fuel is left when the display says you have 0 miles left before refueling?

On my Prius, the answer is generally agreed to be 1.9 gallons or so.


Also, how accurate is the MPG calculator generally thought to be on Audi A-3s?
Is there any way to keep the MPG calculator from resetting if the car has been off for two or more hours--I would like to be able to keep it going for full tanks, so I measure the error for myself.

Again, on my Prius, the computer reads about 5%-7% too high, based both on my experience and the consensus on a Prius fan website like this one.

 

[Birdman]

the Audi works like the VW TDI the MPG display has 2 settings, 1 that resets if the car has been off for 2 hours and one that will hold the settings till you reset it. On my 03 Jetta there is a button on the bottom of control stalk that switches between the 2 settings. You might have to look in your manal if you have one or check Google.

My car on the miles to empty, the low fuel light will come on at 80 miles till empty. I recalibrated my tank so that the light comes on with 2 gallons of fuel left. So it is fairly accurate. when my MTE show 40 miles I will put 15.5 to 16 gallons in the tank.

 

[kgall]

Birdman,
Thanks, I will look for the "second computer".
My low fuel light comes on at 50 miles till "empty". The fuel gauge shows about 1/8 of a tank left.
I guess my question is, how much is left when the gauge and/or the "low fuel" warning says "empty"?
Every car I've ever had still has fuel at that point, and I'm just wondering how much.
I'm the sort of idiot who likes to know that stuff, but not the sort of idiot who deliberately runs my car out of fuel just to find out.

 

[tditom]

I would not chance running a CR out of fuel.

This question about calculated mpg accuracy comes up from time to time and I always wonder why people doubt the accuracy on modern vehicles. The fuel injection system is so precise that the fuel use can be very accurately tracked. Why would they be off?

 

[Birdman]

Your manual will have all those answers, so you really need to find it. I just guess that your car is mostly the same as our VW's if that is the case when the light comes on the book IIRR says 1.9 gallons left. I like to know also, I opened the tank pumped all the fuel out , then added 2 gallons back. Then I used the VCDS and made the fuel gage read right on the red mark where the light come on.

 

[kgall]

I would not chance running a CR out of fuel.

This question about calculated mpg accuracy comes up from time to time and I always wonder why people doubt the accuracy on modern vehicles. The fuel injection system is so precise that the fuel use can be very accurately tracked. Why would they be off?

Who knows why?--my guess is to make folks feel good. I know that in the Prius the computer shows MPG about 5 to 7% higher than you actually get--one of a few things that annoy me about the car--because you are right that there's no reason they shouldn't be accurate.

 

[Roshermoore]

A3 Discussion Area

On an Audi A-3 TDI,
how much fuel is left when the display says you have 0 miles left before refueling?

Also, how accurate is the MPG calculator generally thought to be on Audi A-3s?

You know there is an A3 Discussion Area down the page in the listing of Discussion Areas.

 

[tditom]

Who knows why?--my guess is to make folks feel good. I know that in the Prius the computer shows MPG about 5 to 7% higher than you actually get--one of a few things that annoy me about the car--because you are right that there's no reason they shouldn't be accurate.

but why are you so confident the computer calculated numbers in your prius are wrong?

 

[03Springer]

Remember that the fuel in the tank keeps your in-tank fuel pump cool and lubricated. So when your sucking in the last of the fuel at the bottom of the tank your pump is not being cooled by the surrounding fuel and your ingesting all of of the swill at the lowest point of the fuel tank.

 

[kgall]

Errors, Running out of Fuel, the A-3 Forum

but why are you so confident the computer calculated numbers in your prius are wrong?

Thanks to all

tditom, for every single tank of fuel I've put in my Prius and measured the computer against my calculation of miles driven and fuel used, the computer has read about 1.9-2.7 mpg too high. For the dozens of tanks, it averages out to about 2.4 mpg too high. The variation in error for each tank could be due to errors in how much I fill it up each time. The consistency of the errors, I don't think so.

For 03Springer and others who think I would be an idiot to deliberately run out my tank, I agree for all the reasons you give. I like to know how far I can go before running into those dangers.

Roshermoore, I've only been on this list a few weeks. But I've noticed that there are a lot more people reading this forum than the A-3 forum, and I've gotten a lot more useful answers when I've posted here, even if sometimes things are not exactly the same for VW and Audi TDI's.

 

[tditom]

So you think that the odds that the computer is wrong are better than your filling technique being the big variable?

If you can't be assured of a top off to the same level each time, then you really don't know how full your tank is when the pump switches off.

 

[No More Buffalo]

So you think that the odds that the computer is wrong are better than your filling technique being the big variable?

Yes?

Now, if it was off 1.5-2 mpg, sometimes high, sometimes low, then yea I'd say variable. But when you get the same offset tank after tank after tank, it is much more likely to be the computer misreporting than operator error.

Read the Mk6 Golf forums some time. Our cars do the EXACT same thing, with most repeating the the computer MPG is 1.5-2MPG high consistently.

 

[tditom]

OK- so this would be intentional mis-programming if true.

I will keep a closer eye on my Camry Hybrid consumption (hand calc vs computer)

 

[No More Buffalo]

My pet theory goes like this - Our speedo is set to read a bit high, and it's THIS number that is used to determine instantaneous MPG, and that tank MPG is a continuously averaged derivative of that. So basically the built in speedo error is what causes it to read high.

 

[tditom]

so they back calculate the distance by speedo reading, instead of using the odo?"

 

[No More Buffalo]

Yea, something like that. It's about the only I can come up with that would explain it, unless they really are just intentionally cheating.

 

[IndigoBlueWagon]

FWIW, my '12 Golf said 5 miles to empty today and I added 14.5 gallons of diesel. I've put more in, 14.9 being the most, I think. So I'm thinking it's pretty close to accurate. That last 50 miles before empty goes by awfully fast.

 

[Roshermoore]

Search is your friend

Roshermoore, I've only been on this list a few weeks. But I've noticed that there are a lot more people reading this forum than the A-3 forum, and I've gotten a lot more useful answers when I've posted here, even if sometimes things are not exactly the same for VW and Audi TDI's.

kgall, since you have been on the forum for a few weeks please take this post in a constructive manner. There are many subjects discussed on this forum, some over and over again. Yours is one of those. The subject of the accuracy of the MFD and variations thereof has come up numerous times. Accordingly, use of the search function, especially the use of a Google search of the TDI Forum, can be very useful. Here's one of many posts almost identical to yours which you may find helpful.

http://forums.tdiclub.com/showthread.php?t=326126


Frequently the start of a new thread is not necessary when the subject has been discussed thoroughly.

 

[No More Buffalo]

FWIW, my '12 Golf said 5 miles to empty today and I added 14.5 gallons of diesel. I've put more in, 14.9 being the most, I think. So I'm thinking it's pretty close to accurate. That last 50 miles before empty goes by awfully fast.

Yep, ain't that the truth. Was driving home late last night (~3AM), the fuel warning came on ~40 miles out, reporting 55 miles to empty. 10 miles later, it was already down to 35 (and I still doing the same 65mph highway cruise I'd been doing for the last 3 hours). Luckily I found a station open there.

 

[Chemboy]

I'll pass on a piece of information that Chris from Langhorne VW told me when I bought the Sportwagen:

• The average MPG reported by the computer is only over the last 100 miles driven since reset

I haven't found any documentation to collaborate this, but it would explain why calculate MPG (based only volume added to tank) is different from that reported by the computer.

--Andy

 

[No More Buffalo]

Maybe he meant for the range calculation? Because MFD2 defiantly calculates for over 100 miles.

 

[Plus 3 Golfer]

My pet theory goes like this - Our speedo is set to read a bit high, and it's THIS number that is used to determine instantaneous MPG, and that tank MPG is a continuously averaged derivative of that. So basically the built in speedo error is what causes it to read high.

Yea, something like that. It's about the only I can come up with that would explain it, unless they really are just intentionally cheating.

The speedometer is not used. The three values used to compute speed, mpg, and miles to empty in the MFD are odometer for distance, fuel consumed from the ECM, and time. Speed is computed by distance / time. Take the MFD values - miles and time and you will see that the MFD speed is miles / time.

I have yet to find the time interval used to compute the instantaneous values but it seems to be seconds and the algorithm used that gives more weight to near term data than older data in computing miles to empty.

 

[No More Buffalo]

Interesting. Could it maybe be that the ECM fuel counter is NOT counting fuel used for exhaust treatment (e.g. regens)?

 

[Plus 3 Golfer]

It counts all injections. The fuel consumed is only an estimate. You would need a very precise flow meter to accurately measure the fuel.

At least on the A5s, the MFD calculation for mpg can be adjusted IIRC +- 15% in 1% increments via VCDS. IIRC, I tracked the difference in MFD vs actual for about 40k miles and found the MFD was optimistic by about 0.6 mpg on average per fillup with a standard deviation of about the same. So, at my 42 mpg average, my MFD was about 1.5% higher than my actual.

 

[tditom]

...You would need a very precise flow meter to accurately measure the fuel...

Not really. The pressure is either measured or well characterized, can be compensated for fuel temp (measured). The geometry and inner volume of the injector can be quite well characterized. If you know the dimensions of the orifice (I.D. and length), and the pressure, you can quite accurately estimate the volume of fuel per injection sequence.

 

[DPM]

but the cumulative error will be enormous, adding all the tiny injections together.

 

[tditom]

but the cumulative error will be enormous, adding all the tiny injections together.

I don't think so.
I was fortunate to work with a very smart group of people one time. We devised a way of metering extremely precise fluid volumes through an orifice by montioring pressure and temperature of the fluid and having precise controls on the dimensions of the orifice. Of course we verified the operation through hours and hours of lab testing (that was my contribution )

I can assure you that the internal volume of the injector and the temp and pressure are both well controlled and characterized. With this info you can be extremely precise in your delivery volume. With this info you can easily determine how much fuel has been injected.

 

[DPM]

I don't dispute that but...
(say)7 part-injections per cycle, 2000 rpm cruising, that's 7000 part-injections per minute. How many DPs of a cubic mm are we working to?

I say it's a miracle the system gets as accurate as it does.

 

[NoSmoke]

I don't dispute that but...
(say)7 part-injections per cycle, 2000 rpm cruising, that's 7000 part-injections per minute. How many DPs of a cubic mm are we working to?
I say it's a miracle the system gets as accurate as it does.

But the percentage error is not cumulative. If the measurement error for example is 1.5% high for a single injection quantity, it will be the same (1.5% high) for any number of accumulated injections. If the measurement error for each injection varies, the error over a number of injections will be the average of the individual errors.
IOWs, the volume measurement error is cumulative (in this example getting larger by 1.5% each injection) but the total fuel volume is also cumulative so we are left with the same percentage error in the end.