Fuel Gauge Malfunction

Thanks to the recent good weather I decided to take the GTA out of hibernation and have been using it whenever I've had an opportunity to disappear off the radar for a couple of hours. One trip however ended prematurely with me having to call out the recovery only to discover rather embarrassingly that I was out of petrol despite the gauge reading 3.8 gallons.
The RAC patrolman sold me 10 litres of unleaded and off I went watching the gauge go down as I headed for home.
Today I took the car for a run and put £35.00 in with the gauge reading 2.9 gals. Since the refill the gauge has not gone up with the added 30 litres but continues to go down and is now reading 1.9 gals so I'm puzzled by this.
Any suggestions as to what's going on would be welcome.

Murf.

Can't speak for the GTA, but all the problems I've had on other cars fuel gauge were down to a poor earth at the tank sender.For example I had one case where the sender unit fitted into a circular aperture in the tank, and corrosion between the sender mounting plate and the tank resulted in a high resistance. To make matters more confusing, not all senders work in the same sense, ie some use zero volts to indicate a full tank, others use zero to show an empty tank. Assuming you can access the sender on the tank, it is straighforward to check. Simply attach a wire from a good earth point to a sharp object (eg a multimeter probe or a screwdriver) and use this to penetrate any rust on the sender. If the existing earth is OK, the fuel gauge should not move, but if you have a dodgy earth, this new earth path you have created will cause the fuel gauge reading to change.

Hi Murf,
Before you do anything ...... do a simple computer reset.
To do this, just disconnect the battery ..... find the Computer reset button at the bottom of the Rev Counter .... Press, and hold it down for at least 20 seconds ..... release button, and reconnect the battery. Switch on the ignition, and momentarily press the reset button once more....... it should now read correctly, if there is no other fault.
It could be something else, but try this first.
Let us know how you get on,

PS ......Suggest you don't try earthing anything on the sender as it is an electronic fuel gauge and does not behave in the way most other fuel gauges work. ........ Sorry Griff500 .... but thanks for your interest.

That's interesting johnc. I had understood that digital gauges used the same float/resistor sender principle as analogue gauges, but carried out a/d conversion in the computer. If that is not the case, how do they work?

[stephendell]

It depends if you have an early or late car. On ealry cars the fuel sender is wired directly to the instruments. On late cars there is a separate VDO interface box about the size of a packet of cigarettes tucked away near the brake fluid reservoir. The senders are not interchangeable.

That's interesting johnc. I had understood that digital gauges used the same float/resistor sender principle as analogue gauges, but carried out a/d conversion in the computer. If that is not the case, how do they work?

As you say in your original post..... "Can't speak for the GTA" ...... so you would not know that the fuel tank is plastic in the GTA, and there is no local earth at the tank on either models.
As Stephen says, there are two different types ... old and newer versions. Both these versions have a two wire configuration with neither wire on each sensor being connected to a visual earth. On the older unit, the two wires go directly into the computer, and on the later model the two wires go directly into the TSU.. (tank sender unit) ... as Stephen noted ... So as we do not know exactly what potential there is on any of the wires, and what they are connected too inside either TSU or Computer, I am sure you will agree it is very unwise to go poking around with an earthed wire without having satisfied yourself that you will not cause any damage to the internal circuitry of either model without some information gathering testing beforehand.

Many thanks for the responses. I will have a go at resetting the computer this afternoon as John has suggested and let you know the outcome.

Murf.

I was surprised to find when I got in the car this afternoon to do the reset that the gauge had re-registered overnight at 7.0 gals!
I have now followed John's instructions, disconnecting the battery and resetting the computer and it is now reading 5.1 gals. Will take it out later to see if the reading decreases with driving.

Murf.

Yours is a turbo isn't it Murf?

As far as I understand it, the gauge system also uses a bit of strange logic in how it calculates and shows the remaining fuel.

Essentially the computer only reads and stores a value when it sees the fuel level go up. All remaining values are then calculated from fuel consumption (so X start - Y used = Z remaining). The fuel consumption is calculated via the ECU (injector duty and duration) on the Turbo or via a flow meter on the atmo.

Now obviously if you use more fuel than it thinks you've used, then the displayed value will over-read - at least until you've put more petrol in. If you put bigger injectors in then the mpg and fuel remaining values are going to read way out. Likewise any other variables that the ECU (which is pretty basic) can't quite compensate for.

I don't think any other car would use such a system, or at the very least it would re-calibrate each time you restarted the car.

There is a way to switch to diagnostic mode to make it re-read the value from the actual float level.

As far as I understand it that's all correct on both early and late models, but Steve or John may be able to correct if I've got anything wrong.

Yes mine is an 89 turbo and thanks Darren for the explanation. When I got back in the car about an hour ago the gauge reading had increased to 6.9 gals so I went for a very pleasant sunset run around the block and monitored the steady reduction in contents, which by the time I got back home was down to 6.1 gals with the range reading 168 miles.
Will be interesting to see what figures are showing in the morning.

Murf.

Yes mine is an 89 turbo

Just as a matter of interest ..... your Alp is a 1989 model. This is change over year from the older gauge unit to the later one. If you look at the pic below, it shows the position of the TSU used in the later Alp. The TSU is item 327 which is attached to the wheel arch.
It might be worth checking if it is there ..... this can be done by removing the cover which protects the hydraulic fluid reservoir under the bonnet LHS front, and firstly with a torch having a look to see if you can see it .... a mirror might help, or having a feel as you can get a hand on it. The alternative is removing the boot liner which you don't want to do at the moment.
If it does not appear to be there ..... with your torch have a good look around just behind the headlights to see if it exists just lying there as I came across a 1989 Alp which seemed to be in transition between the old and the new. The TSU was there but not where it was eventually fitted on the wheel arch.

TSU p0sition.jpg

With regard to your gauge reading which looks as if it is a bit random at the moment. Don't hesitate to do more full resets even when you are on a run. Pull over .... do a full reset, and see when you press the reset button finally, if the reading is the same as when you stopped.
But the acid test is to fill the tank to the brim, ....... do a full reset at a convenient spot, allowing time for the fuel gauge tube to fill up, and the gauge should read very close to 15 gallons ...... mine always reads 15.2 galls when I fill up.
But it does appear that the system is operating correctly as the reading does go down as you use up fuel. The only problem is that it is not consistent with a fuel quantity reading.
Let us know how you get on.

John

As expected I've been getting completely random gauge readings during the course of the day without the car moving from the drive. 6.5 gals/173 miles, 5.9 gals/157 miles and just now 3.9 gals/103 miles. Have got the boot-liner out and taken pics of where the sender unit should be (see attached). Is it the inner unit with the metal locking bar?
My next step is to do another re-set as John has suggested in an attempt to calm down the bizarre behaviour.

Murf.

No that is just an electrical multi-plug connector.
The sender is in the tank, you can just see it in first pic top left. Has a two pin plug connected.
Looks like you have just the sender version.

Hi Murf,
It does look as if Lee's right about it being the earlier type ...... but just to confirm ..... the black unit with the white cable tie, ..... is it just a connector or do the wires enter it from one end only ?. The reason I ask is there is what appears to be a small box with wires only going in at just one end in the 1989 electrical manual .... but it is not numbered, and not shown in the circuit diagram.
Secondly ....... can you take a pic of the fuel filler cap as that should confirm which sensor is fitted.
And if possible .... can you follow the two wires back from the top of the fuel sensor and see where they go or attach too.

I have been doing some experiments with my fuel level system today, and like you, I have had some weird readings. So I decided to fill the tank to the brim to see if it would come up to 15 or so gallons. I had done a couple of resets and got different readings, but the level was quite low .... around 5 gallons. So converting the gallons required to fill the tank, that came to 49.95 litres .... say 50 litres. I set off to the fuel station which is only about a mile away and filled up. I was only able to put around 40 litres in, and the reading came to 14.1 gallons on the gauge....... not the 15.2 galls as expected.
The forecourt was busy so decided not to hold everyone up so drove about 2 miles to a level car park to do a reset.
As I switched off I remembered that I had read or been told possibly 30 years ago, not to be in too much of a hurry to do ...... Something ....... although it was to do with fuelling up. I sat there for a minute then heard the click of the timer relay dropping out ..... this relay is only fitted to the later models ..... but it did get me thinking. Was I supposed to wait till the relay clicked off before refuelling .... or starting up after refuelling. Anyway, after a few minutes thinking, I did the reset, and to my utter surprise up came 15.1 gallons, which was about right as I had used a bit of fuel from the garage.
But it struck me that switching on the ignition quickly after a refuel has a problem, which might set the gauge reading incorrectly.
My tank was quite low when I started to refuel, and of course the fuel went in quickly and came to the brim within about a minute or so. But of course the level of fuel is measured in side the sensors float tube, and that has a very small hole at the bottom to allow the fuel to fill up the tube and raise the float to the top. In the pic below you can see the little hole towards the right hand end of the tube. Add to this the air has to be pushed up and out through another hole at the top. Filling this tube with fuel will take much longer than filling the tank, so my thought is that if you start the car .... as I did today .... quickly before the tube is filled, the gauge will read the level that the fuel in the tube has reached.
the hole in the tube is there to damp the level in the tube as the rest of the fuel in the tank could well be sloshing around as you start/stop/turn left or right.
If I am right, that is why my tank read 14.1 galls when I set off from the garage. By the time I reached the car park, the fuel would have slowly filled the tube .... hence having done a reset, the reading now correctly read 15.1 galls.
I will leave you with that thought, and just post the pic below of my (later) sensor with its small hole.

Fuel Sensor.jpg

Thank you Lee and John for the feedback. Yes the connector block with the metal locking bar does have wires on both sides. I've taken pics of the top of the sender, the label reads Jaeger 337649 01, 12v, Made in France.
Tracing the two orange wires back in their plastic conduit they are joined by an orange and a black wire from the brake fluid reservoir which all then meet up with a cluster of taped wires heading back towards the scuttle.
Also attached is a pic of the fuel filler cap.

Murf.