Analog Circuit Design for Fuel Tank Sending Units: Help Needed

[johnerbes]

Typical fuel tank sending units have a resistance of 0 to 10 ohms in the full position and 90 to 240 ohms in the empty position. I have an old Weston 50uA meter I would like to use as a fuel gauge. I need help with designing an analog circuit that will drive the 50uA Weston meter from the tank sending unit (and not blow up the gas tank). V+ will be from the car's 13.5 V battery (running). Thanks for the help.

 

[uart]

Hi John. Is that a resistance to ground or do you have access to both resistor terminals.

Also do you know how much voltage/current that it is safe to use on the sender without risk of blowing up the gas tank?

 

[Type_R]

Typical fuel tank sending units have a resistance of 0 to 10 ohms in the full position and 90 to 240 ohms in the empty position. I have an old Weston 50uA meter I would like to use as a fuel gauge. I need help with designing an analog circuit that will drive the 50uA Weston meter from the tank sending unit (and not blow up the gas tank). V+ will be from the car's 13.5 V battery (running). Thanks for the help.

Doesn't you fuel tank sending unit(in your car) already have a gauge?
are you trying to see how much fuel is being sent by looking at amps output?

 

[johnerbes]

I'm about 95% certain that the tank sending unit resistor is isolated from ground. Many gas tanks are plastic these days. Also, a reliable ground through the tank attachment straps on even a steel tank over time is iffy.

In the recent past (30?) years, fuel gauge circuits appear to consist of; the tank sending unit variable resistor, the fuel gauge, some current limiting device and the car battery. The fuel gauge these days consists of a bi-metallic strip with a handful of windings. One end of the bi-metallic strip is anchored and the other end attaches to a short arm which attaches to the fuel gauge needle. When the tank is full the heat generated from the coil around the bi-metallic strip distorts the strip and moves the needle via the attachment arm to the full position.

I'm guessing there is a couple of watts generated and that the bi-metallic coil is an ohm or so. If the voltage drop across the coil is a couple of volts, my first order guess for current is around an amp or so (plus or minus 500%) with a full tank.

I was considering a small un polarized cap on each end of the sending unit to ground.

 

[johnerbes]

This isn't for my car. It's for a different application.

 

[Type_R]

This isn't for my car. It's for a different application.

So you're replacing the fuel gauge in the tank sending unit with your amp meter?

 

[Type_R]

http://img339.imageshack.us/img339/8146/fueltankit4.jpg

I found a simple little schematic for the unit.
And it appears the variable resistor is attached to ground.

So basically you're going to be disconnecting the part of the circuit which attaches to the fuel gauge...and make it goto your amp meter?

 

[johnerbes]

I have a friend that is trying to create a collection of authentic instrumentation for an old car replica he is building. I happen to have a bunch of old Weston meters. Weston actually made voltage and amperage meters for automotive applications in the old days that looked just like their other meters. It's not a problem to make an authentic volt or amp meter, just make the right shunt etc. I take the face out of the meter, create a new face with AutoCAD or something similar, and paint on a new face using one of the Epson inkjet photoprinter that uses a CD/DVD tray. Works quite well. I didn't want to just cobble together a fuel gauge by using the meter housing and stuffing in a modern bimetallic meter. I wanted something that could use the Weston 50uA meter movement. So, yes, I'm basically going to be disconnecting the part of the circuit which attaches to the fuel gauge and make it go to the amp meter. The schematic is accurate. One end of the variable resistor is connected to ground - but not at the fuel tank. Both ends of the resistor should be available.

 

[Averagesupernova]

Fuel tank sending units are anything but typical. Different manufacturers have had several different resistance ranges over the years. GM commonly used 0 ohms empty, 30 ohms full back when 6 volt electrical systems were common. This bled over into early 12 volt systems until GM switched from 0 ohms empty to 90 ohms full. Ford used for many years 73 ohms empty 10 ohms full. Stewart-Warner and Sun gauges used 240 ohms empty and 33 ohms full. VDO gauges used 10 ohms empty and 180 ohms full. Now with trip/mileage computers the fuel sending units have changed to something more precise. Very early sending units had a heater in the gas tank next to a bimetalic strip and set of contacts. The float level 'weighted' how much the heater had to heat before the contacts opened. So the unit would cycle on and off with the duty cycle dependent on the level of fuel in the tank. The gauge measured this similar to the diagram posted above. The bimetalic strip in the gauge does not move fast enough display the on-off nature of the signal coming from the sending unit.

 

[johnerbes]

Now that we have this fully analyzed, does anyone have any thoughts on an appropriate circuit? For the prototype circuit, let's assume that the resistance of the sending unit is 5 ohms full and 100 ohms empty and that both ends of the variable resistor are accessible. The Weston meter movement is 50uA full scale. Would an op amp stuck in between the resistor and meter be a good place to begin?

 

[Pumblechook]

A bridge with the meter + series resistor on the ouput of the bridge. Bridge balanced with the resistance as per the empty condition.

One preset resistor as one leg of the bridge (maybe with a parallel fixed R) to allow zeroing. One preset (maybe in series with a fixed R) in series with the meter to set the full scale.

 

[hummingbird]

Did anyone ever come up with a neat solution to this. I'm currently faced with the same design issue. I thought of having 8 resistors that can be grounded using low resistance MOSFETs. Resistor values 1R, 2R...128R would then allow resistance of between 1R and 255R (or open circuit of course).

The issue is that the gauge may require a floating resistance rather than a gounded one, which makes the design more complex.

There is the option of using a programmable current source, measuring the voltage across the terminals and setting the current to create a resistance. Seems like a lot of work if one is to make it "float".

Or there is the option of using a FET output optocoupler to create an isolated resistance, but how reliable would that be over temperature and between devices - not at all, I imagine.

Any ideas?

 

[Proton Soup]

various basic op-amp circuits can be used to transform a varying resistance/impedance into a voltage signal.