I thought your answer to the backwards reading gas gauge question was pretty lame. How about maybe everybody was telling the truth, and the float assembly had been mounted upside down? Perhaps it was of rather symmetrical construction and looked fine either way.
It seems to me I remember a caller from some years back that had a Ford Tempo which had an electrrical fire in the dashboard. The caller had substuted the cluster from a Mercury Topaz. All worked correctly except the gas gauge which read backward. Tom and Ray suggested that the owner live with it.
If the sending unit had been replaced in the tank of the recent caller’s 1996 Mazda pickup truck, it may be set up so that it gives maximum resistance where the original unit gave minimum resistance. Perhaps there was some change in the sending units since 1996 and while the parts interchange, I would just live with it.
I owned a 1948 Dodge and the orginal owner had replaced the battery. Everything except the radio worked, but the ammeter gauge read backward. The person who replaced the battery didn’t realize that these old Chrysler products were positive ground. I completely discharged the battery and brought it back up on a trickle charger. I then repolarized the generator and everything then worked including the radio. I am certain that the problem in the caller’s truck, however, isn’t a battery where the poles have been reversed.
I rented a brand new Chevy HHR a month ago that did exactly the same thing. The guy checking me out pointed out it had a half tank and I needed to return the car with the same level. I drove it a total of about 450 miles. The gauge advanced to nearly full in about 220 miles and and I filled it. The gauge went to empty, and then crept back up to a little over half full before I returned the car. Fortunately, the computer’s “miles-to-go” readout worked just as it should have. Of course I turned in the car after the rental shop closed and I never did get to talk to them about it.
I like PurplePhlogiston’s & Triedaq’s diagnosis— the float arm on the sending unit is essentially mounted backwards, giving max resistance when full rather than empty. It’s easy to switch the max and min on some senders; I converted my Mercedes 300sd to run on WVO, with a two tank system. I wanted to use the dash gauge for both tanks. The MB gauge was 0-90, the kit supplied a second gauge/sender that was 33/240. I bought a 0-90 aftermarket GM sender, but it was reverse of the MB. (I forget which was which, but one was 0 full 90 empty, probably the MB sender with tube style sender, the other was 90 full, 0 empty, and the gauge reads only one way.) I was able to install the float arm on the GM sender the opposite way so that it’s resistance read properly on the MB gauge.
I was thinking that IF the chafed wire excuse was correct, AND the chafed wire was around the gas tank - the weight of the gasoline might shift the tank enough to un-short the short.
I am a computer controls engineer. While I know nothing about fuel level sending units, I felt it only proper to add my worthless two cents, since “I wasted a perfectly good hour listening to Car Talk”.
An interesting question to ask would have been “does the gage go to full instead of empty when you turn the engine off”. If it goes to full, then I would have stronger support lords the gage itself being bad.
I suspect that the fuel gage is no more than a voltmeter, and the fuel level sending unit uses a rotary potentiometer wired as a voltage divider. To do this, you place 12 volts on one side of the potentiometer, and ground (also known as 0 volts) on the other side. These 2 terminals connect to opposite sides of a resistive coating. There is then a 3rd terminal, called a wiper that taps into the resistive coating at different spots. When the wiper is at the ground end, we see 0 volts on it. When the wiper is on the 12-volt side, we see 12 volts on it. If the wiper is 1/2 way in between, we would see 6 volts on the wiper.
The catch is, if you were to swap the 0 volt and 12 volt connections on the potentiometer, it would read 12 volts on empty, and 0 volts on full. Therefore, I think that an improperly wired fuel level-sending unit caused the problem.
In reality, they probably don’t feed the gage with 0 to 12 volts. They probably use a precision voltage reference, which reduces the supply voltage down to a constant 5 volts regardless of the system voltage. This would prevent the gage from moving due to system voltage variations.
There is probably a bunch of filtering taking place in the gage, so it responds to changes slowly. This would make the meter not respond every tome we go over a bump or stop sharply.
I could also go along with it being an improperly installed fuel level-sending unit, if there is a mechanical way to install it wrong so the potentiometer moves in the reverse direction.
I guess I got even, now that you wasted some perfectly good minutes reading my stupid comments. Seriously though, I enjoy your show on WFYI Indianapolis, and this story was of high entertainment value.
I’d like to +1 keithbeidelman’s comment. A reverse wired potentiometer in the sending unit sounds exactly right.
Reverse pot equals the electrical equivalent of upside down! Could be easy to do if the connector wasn’t polarized.
I have a possible explanation for the question about the fuel gauge operating backwards. The gauge normally has two magnets; the one on the left (empty end) operates at full voltage all of the time. The one on the right is in series with a variable resistor on the sender; when the tank is full the resistance is zero and when the tank is empty the resistance is maximum. The needle is connected to a vane between the two magnets. Thus it would seem when the tank is full the needle would be at midpoint but it also has a spring attached which pulls it to the right, full position. As the tank goes down the resistance increases and the magnet on the right gets weaker and the needle moves to the left. Now suppose the magnet on the right—the one in in series with the sender resistor–fails or becomes disconnected. And also suppose someone in an attempt to make a repair connects the magnet on the left in series with the sender resistor. We now have one magnet—the one on the left-- and the spring controlling the needle. When the tank is full the magnet is at maximum pull and moves the needle to empty. As the level in the tank goes down the resistance increases and the pull of the magnet is less allowing the spring to move the needle to full.
Whala, the gauge reads bass ackwards!
I bought a 1970 VW bus that had this problem. The guy who sold it to me demonstrated his technique for checking the tank. He would stand in the middle of the cargo area and rock the van back and forth until you could hear the sound of sloshing gasoline. Replacing the sending unit on that model would have required pulling the engine, so he learned to live with it. I guess he got tired of this nonsense and sold the thing to me. I got it home and figured out what the problem was in about an hour. The battery was hooked up backwards.
Since we are, sort of, on the subject, My 2002.5 New Beetle VW has started giving a low fuel light when the fuel is full or any time it is not low. Then when it is low, the light goes out. The low fuel chime is also opposite. Other than the opposite indicators, it appears accurate.
My first guess is that someone working on the car previously has installed a connector backwards somewhere. Assuming though that the gas guage read correctly at one time, and then started this extraordinary behavior of reading backwards for no apparent reason, and that no one worked on the car at that time, the first thing I’d visually check that all of the connections are good, from the sending unit connector, on the outside of the tank, right up to the fuel guage. You’ll probably need a schematic to do this correctly. You won’t have to remove the sending unit (the float & variable resistor) or the gas tank at this point. If the wiring and connectors look ok, then I’d use an electrical meter to check all the wires involved in the gas guage operation for broken wires, or wires shorted to ground or shorted to +12 v. Doing this requires some experience with electrical system troubleshooting. Get some help if you haven’t done it before. If that checks out ok, next I’d remove the harness connector at the sending unit and check the resistance the sending unit is providing. Doing all that will very likely show you where the problem is. If the repair involves removing the gas tank, or even draining the gas tank, or if it involves any leaking gasoline, have it done professionally.