Seymour Snerd
Lifetime Supporter
There are some other threads documenting various woes associated with fuel level senders in GT40s, including some Superformance cars. Generally the complaints are that one end or the other (or both) of the range of fuel levels is not represented by the gauge. Several people have switched to solid-state, usually capacitative senders; others to similar looking columnar sensors that contain an internal float. Some have simply calibrated the supplied sender by bending the float arm.
Having read all that I decided to "calibrate" my sender ahead of time rather than waiting until there is gasoline to contend with. I put "calibrate" in quotes because my intent was to fully capture the entire range of fuel level within the range of the gauge, regardless of how accurate those readings would be, although later on I might fiddle some more to get "E" to correspond with some known value relative to where the engine stops.
Since the nominal resistance range for Smiths instruments is 20-240 ohms I set as my objective to have the sender read ~240 with the float resting on the bottom and ~20 with the float as high as it can mechanically move within the tank. This can be done with the supplied sender but, due to the fact that the sender is set down into the tank about 1 inch it requires the float to be able to rise above the plane of the sender mounting. As delivered, my float did neither that nor did it fall all the way to the bottom.
Thus my main problem was to increase the range over which the float operates. To do this requires both straightening and lengthening the float arm. Oddly enough, the last inch or so of arm is actually two pieces of wire overlapping and soldered together. So I "lengthened" mine by simply melting the solder and sliding the two piece so the overlap about a half inch less than before. Then "calibration" is a matter of bending the arm at its one remaining angle until the bottom of the float is 6" below the float mounting surface when it is fully dropped.
Note that at the ends of arm travel the resistance may go to an "open" condition. This does not matter as long as the tank is constraining the float's travel. However, when you are working with it and an ohmmeter the open condition can be a distraction so I adjusted the sender's arm limiters to prevent that. It's also true that at one extreme of motion the internal wiper of the sender can over-travel and drop off a step in the sender, after which it will refuse to move. This can be fixed either by poking carefully through an inspection hole provided, or by bending up three tabs that allow access to the innards of the sender.
Also, the arm engages the variable resistor through a hole in the nylon wiper piece, and the end of that piece pokes out the side of the case and is slotted such that you can hold the arm still and turn the slotted shaft with a small screwdriver in order to make very fine adjustments. This is how I would tweak the units value at a single point, since the arm rod is quite stiff and thus hard to adjust finely by bending.
For those considering replacing the stock sending unit with a columnar float design be forewarned that you are inherently giving up the ability to measure the top inch or so of the tank's capacity, at least 15%, possibly more since the tank is wider at the top than the bottom.
Having read all that I decided to "calibrate" my sender ahead of time rather than waiting until there is gasoline to contend with. I put "calibrate" in quotes because my intent was to fully capture the entire range of fuel level within the range of the gauge, regardless of how accurate those readings would be, although later on I might fiddle some more to get "E" to correspond with some known value relative to where the engine stops.
Since the nominal resistance range for Smiths instruments is 20-240 ohms I set as my objective to have the sender read ~240 with the float resting on the bottom and ~20 with the float as high as it can mechanically move within the tank. This can be done with the supplied sender but, due to the fact that the sender is set down into the tank about 1 inch it requires the float to be able to rise above the plane of the sender mounting. As delivered, my float did neither that nor did it fall all the way to the bottom.
Thus my main problem was to increase the range over which the float operates. To do this requires both straightening and lengthening the float arm. Oddly enough, the last inch or so of arm is actually two pieces of wire overlapping and soldered together. So I "lengthened" mine by simply melting the solder and sliding the two piece so the overlap about a half inch less than before. Then "calibration" is a matter of bending the arm at its one remaining angle until the bottom of the float is 6" below the float mounting surface when it is fully dropped.
Note that at the ends of arm travel the resistance may go to an "open" condition. This does not matter as long as the tank is constraining the float's travel. However, when you are working with it and an ohmmeter the open condition can be a distraction so I adjusted the sender's arm limiters to prevent that. It's also true that at one extreme of motion the internal wiper of the sender can over-travel and drop off a step in the sender, after which it will refuse to move. This can be fixed either by poking carefully through an inspection hole provided, or by bending up three tabs that allow access to the innards of the sender.
Also, the arm engages the variable resistor through a hole in the nylon wiper piece, and the end of that piece pokes out the side of the case and is slotted such that you can hold the arm still and turn the slotted shaft with a small screwdriver in order to make very fine adjustments. This is how I would tweak the units value at a single point, since the arm rod is quite stiff and thus hard to adjust finely by bending.
For those considering replacing the stock sending unit with a columnar float design be forewarned that you are inherently giving up the ability to measure the top inch or so of the tank's capacity, at least 15%, possibly more since the tank is wider at the top than the bottom.