the fuel gauge read full after 110 miles after it had 2 marks down

sounds like something inside the tank, but really, i have no clue, sorry!

 

Sounds normal to me. I've seen it quite a few times.

Perhaps we need to write a 'sticky' to describe the many vagaries of fuel gauges? Many posters are expecting far more accuracy and stability than these things can provide.

 

Does not sound normal to me. Something is not right. Maybe you both have the same problem.

 

This same 'problem' also existed in my old Subaru and Honda with analog needle fuel gauges.

 

i don't know, even with the blatter, in 7 years and 2 pri, i have never seen it go up a pip.

 

I have, but I was parked on a hill. I've had it go down 1 bar, then back to full again. Maybe you parked on a hill? Although, it would have to be a pretty steep hill to make up for essentially 2 gals of gas being used to show as being back.

 

It has always happened to me, if we can only get the DTE to work like that too.

 

That's why some people here propose changing it's name to a "guess guage."

 

Yep. I've got ideas. And facts to back them up, too!

Had a 2000 Honda Odyssey. It started doing things like that right after my wife and I got it. It would read "normal" until it got a little bit below a half tank - and then it would read full. Then empty. Then "normal". And so on. And, with it being intermittent and all, the dealer couldn't make it happen on his watch.

Finally, they made it happen. Now, I'm a EE, and, unlike Toyota, Honda does sell their repair manuals for a reasonable penny. I had looked over the schematic: All that's in the tank is a sender: That is, a float that goes up and down and a potentiometer that goes from short to open, depending upon the tank volume. On the other end is, well, the dash, which had what I figured to be an A/D converter or some such that drove the actual gauge and made the low gas light come on. (The low gas light worked, the gauge didn't.. Which is why I didn't think it was the sender.)

Easy to troubleshoot: Pull the sender, put an ohmmeter on it, and run it back and forth. If it goes weirdly, replace the sender. If it doesn't, replace the dash electronics.

You can guess where this is going. The dealer replaced the sender, said, "Fixed!" and sent us on our way. One trip to Boston later and we smelled a strong odor of gas in the car. Then a puddle of gas under the car. To make a long story short, turns out that the gasket around the sender port was easily damaged, and the dealer did had done it. Another dealer in Boston took the car in, said, "You're not driving this!" and spent three days trying to locate that gasket, somewhere in the continental U.S.. Luckily, they found one in Vermont.

Got the car back, no smell - and the gas gauge started acting up again. This time the original dealer replaced the whole instrument cluster, including the odometer. Seems that gas gauge itself was just a non-repairable part of the whole thing. Bummer, and we then had a sticker on the car saying that the odometer reading was off by 1000 miles because of all this.

Therefore: If your gas gauge is jumping back and forth, go forth and lean on the dealer. Don't wait. And try and get them to do a little troubleshooting before they start replacing the easy stuff.

KBeck.

 

For a gauge that swings a full tank, as yours did, or even half a tank, I would agree.

But any measuring device will display at least one tick of uncertainty. That is an unavoidable consequence of noise in measurement theory. When displaying repeated measurements, without memory or hysteresis, it will jump back and forth at least that one tick when the quantity being measured is close to the tick boundaries.

Add in nonlinearities of low cost systems -- on mine, the second bar down represents much less than the first and third bars -- plus substantial sloshing around in the tank caused by up and down hills, acceleration and deceleration, and corners, and two bars of back and forth is not unreasonable. I see it frequently at the top of the gauge, especially on a hilly city commute route, but it has also been observed close to the bottom of the gauge.

I have not seen three bars of this noise.

 

All right. For what it's worth, In the year and a bit that I've had my Prius, I've seen the gas gauge go down, but I sure haven't seen it go back up. I would have noticed. I have noticed that it tends to go from dead full to two bars down pretty rapidly, but I've been attributing that to the general non-linearity of gas tank gauges in general.

What I have noticed with all the gas gauges I've had the fun to use has been that they don't change their readings quickly. With the older cars I attributed this to lots of inertia and stiction in the needle: In the new ones, I figure that there's a long integration constant.

I haven't particularly noticed the gas gauge on the Prius moving around much in turns or under acceleration. I suppose if I was parking on hills then I'd expect some movement. But, going from two bars down to full? No way.

I still think he's got something wrong with the sender in the tank, something in the dash, or a loose wire somewhere.

KBeck.

 

The large fuel level slosh I see is on Adrian's FLV Xgauge on a ScanGaugeII, not on the Prius dash display. Comparing the two shows that the later is very heavily filtered. The majority of my tanks march down the gauge monotonically, but not all of them do so.

I agree on this, it doesn't move around much. My old analog Honda moved around far more.

Way! at least on mine. I haven't kept track, but would guess that about 25% of my tanks on the hilly / curvy / frequent stop commute route do this. But it is less common on my other commute route choices, and I can't say that it has ever happened on the open highway.

 

Between 8 and 10 pips it happens often
I also have seen it go back up from 5 to 6 and from 6 to 7. I believe I saw 1 to 2 a few times as well.

 

Mine does this too. May be a sticky sender. May have something to do with the inclinometer is the display (that adjusts the signal).