Despite what you may have experienced, the fuel gauge can actually provide useful information about the contents of the tank. The circuit is very prone to problems from marginal connections, especially on the ground side. While certain connections may have been adequate when originally produced, they will degrade over time and cause baffling problems. The system can be easily upgraded for the utmost in accuracy and reliability. While a meter is shown in some of the following pictures, it is for illustration only. It is a simple process to upgrade and calibrate the fuel gauge system and a meter will not be needed. While some of the following details apply only to the early civilian Jeeps, much of the information is applicable to any fuel gauge system, new or old.
In all honesty, you don't need to know what occurs inside the sender or gauge as long as everything is properly connected. The early CJ series uses the balanced-coil design, with two electromagnet coils inside the gauge as shown. The two coils counteract each other and pull the needle one way or the other depending on their relative strength. Fluctuations in the voltage supply have little effect on this design, as the relative difference between the coils doesn't change. The Jeep's original 6 volt electrical system could range from approximately 5.0 to 8.5 volts depending on load, RPM and state of battery charge. A balanced coil gauge is quite happy anywhere in that range.
The variable resistance of the sender changes the current flow through the coils. When the sender provides more conductivity (less resistance), current takes the easy path and bypasses the left-hand coil on its way back to ground. Greater relative current flow through the right-hand coil pulls the needle in that direction. When the sender provides less conductivity (more resistance), those two-timing electrons follow the easier path through the left-hand coil, pulling the needle that way. Since current flow would now be nearly equal through both coils, there is apparently some additional means to pull the needle to the left. I've never had this gauge apart, but most likely the left-hand coil has more windings and/or there is a light spring acting on the needle.
In this particular type of gauge, a higher fuel level means greater conductivity (less resistance) through the sender. Many other vehicles used a circuit that was just the opposite. Imagine the internal working of the gauge as a mirror image. The matching sender would need to work in the opposite manner, too. Heaven knows why there was a difference, as neither offered any advantage over the other. Near the bottom of this page, I'll show a very simple method to convert one type to work with the other as long as the resistance range is about the same. This might come in handy if you are swapping in a non-stock tank and need to use that sender, or are using a different gauge with the stock tank and sender.
Remember, this is all gee-whiz info so far. An understanding of the internal circuitry is not needed to calibrate the system and make it more reliable. Forget anything else you'd like, but do remember there are three critical connections that are easy to overlook. A teensy smidgen of extra resistance at any of these location will cause all sorts of mystifying problems.
Potential Problem #1: The wiper arm inside the sender is grounded through the tank itself. A lack of proper conductivity (extra resistance), will cause the gauge to read low.
Potential Problem #2: Although shown above with a separate wire for illustration purposes, the left-hand coil is actually grounded to the dashboard through the body of the gauge. Inadequate conductivity (extra resistance) will cause the gauge to read high.
Potential Problem #3: The gauge's mounting bracket is secured by nuts tightened on the terminal studs. Marginal connections will result if the terminals are secured between the bracket and mounting nuts.
Notice I keep using the term "Conductivity"? That's a $10 word describing a circuit's ability to flow its designed current. Throw in some extra resistance, whether wanted or unwanted, and conductivity drops. Remove or bypass resistance, whether wanted or unwanted, and conductivity increases. I find it a lot easier to describe current flow in terms of conductivity. If you absolutely, positively have to think in terms of resistance, the Luddite terminology is included in parentheses.
Where were we? We were talking about some pesky connections that will eat your lunch if you don't pay attention to them. We'll come back to the dashboard ground and gauge terminals later, but first let's examine the sender and learn how to make it more reliable. From my Pile-O-Parts, here are three senders with three types of floats. Note how the case of the top sender is different, although electrically all three are the same. Cork floats eventually become less buoyant and sink. They can also disintegrate and clog the fuel line and/or filter. Brass floats eventually sink, too, after developing pinholes from corrosion. This leaves plastic as the most durable choice:
Here's part of what's inside the sender case. (This is the mounting flange from the sender at the bottom of the previous picture.) Normally there'd be no need to open the case, but this scrap unit was cut open for illustration. A single length of fine, bare wire is wrapped around a phenolic insulator. At the left, the small rivet electrically connects one end of the wire to the screw terminal via the L-shaped support bracket:
With the side of the lower case cut away, the inner workings are visible. The float arm moves the grounded wiper arm across the resistance windings. Note the two adjustable stops which limit the mechanical travel of the float arm and wiper. Look closely, and the arc worn by the wiper arm is visible on the face of the resistance windings:
Here's the wiper arm to the far left, making the most conductive (least resistance) connection to the screw terminal on the on top. The meter is showing 0.3 ohms, a nearly perfect connection as far as tank senders go. Note how the float is up and the arm is against the high-level stop. Later on, this stop will be adjusted to calibrate the sender to the gauge. Remember the meter is for illustration only and won't be needed for calibration:
With the tank empty, the float drops and the arm is against the low-level stop. The wiper arm is to the far right. The sender is now adding about 133 ohms of resistance (Yep, I used the word) into the circuit, making it less conductive. Only about 120 ohms is needed to drive the gauge to the "E" mark, so the low-level stop will adjusted to match the gauge in a later step. Please note that the 0-120 ohm range may not necessarily apply to all model Jeeps:
The float and wiper are in the middle in this next image. Note that 50 ohms is not exactly halfway between the full and empty readings. A plot of resistance (Doh!) values versus the tank level would show a curve, not a straight line. I don't know why the circuit was designed this way but it will affect the accuracy. Later in the calibration process, the importance of this will be explained:
In the three pictures above, note how the negative meter lead was connected to the float arm. The float arm is connected directly to the wiper arm, making a good connection. However, the ground circuit is completed through the pivot point to the sender case and hopefully the tank itself. This is one of the areas that can lead to trouble. Here's a close-up of a new sender, with a flexible jumper added between the base of the wiper arm and the case itself. The wiper is bronze and the case has a tin coating, making soldering easy:
The jumper must be very flexible, of course. Desoldering Braid, available at any electronics store, works well. This spool came from Radio Shack, with the Archer part #64-2090B. This is a fine copper braid designed to wick away excess solder from circuit boards. Clamp each end when soldering, or the braid will fill up with solder:
This jumper will make a big improvement, but the sender has another troublesome area. The upper and lower parts of the case are riveted together. This connection can slowly degrade as corrosion sets in. Without any need for disassembly, a small drop of solder can be added between the two parts. There is no need to solder the entire perimeter. Between the braided jumper and this solder bridge, the wiper arm will have a perfect connection to the mounting flange for years to come:
The rest of this circuit leg still has a long ways to reach ground. In an example of starry-eyed optimism, the mounting flange was grounded via the tank itself. With a gasket between the flange and tank, the mounting screws provided the current path. This was assuming (Danger! Danger!) that the tank itself was grounded. Let's see, a painted tank resting on strips of padding, secured by a padded strap... With the sole ground being the fuel line secured to the frame with spring clips, it is a wonder the fuel gauge ever worked in the first place. This shortcoming can be easily rectified with a ground wire attached to one of the mounting screws. We're getting a bit ahead of ourselves here, but this installed view shows the ground wire in addition to the sender wire on the center terminal:
With a reliable ground squared away for the sender, let's turn our attention to the ground connection between the gauge and the dashboard. The original connection left much to be desired. The mounting bracket had pointed ends which allegedly cut through any paint to contact the back of the dashboard. A spring pressed against the back of the gauge to hopefully complete the circuit:
Those are two connections waiting to fail. One method around this is to secure a separate ground wire to the case. A hose clamp can hold a common ring terminal against a clean spot on the case. One disadvantage with this method is that the clamp and wire must be installed after the gauge is inserted in the dash:
An easier solution is to solder a small, L-shaped brass tab to the back of the case. The ground wire has a spade terminal for easy installation. With enough slack in the wire, the terminal can be attached to the brass tab before the gauge is installed. The case has a tin coating for easy soldering. Use caution to avoid sanding off the tin coating when preparing the surface for solder. While these old gauges are pretty robust, work quickly with the soldering iron to avoid the possibility of internal heat damage:
Even though it's not time to install the gauge yet, let's improve the terminal connections. Each terminal stud has a nut to secure it to the case against insulated washers. Do not loosen these nuts, or the studs will spin and break the internal connections. A second nut on each stud holds the mounting bracket in place. The bracket will bend before these nuts can hold the wire terminals securely. It is very important that each wire terminal is sandwiched in place by a third nut. Clean the threads with a wire brush and use new brass nuts to ensure the best connection for the wire terminals. (Don't ask) Note the labels, highlighted in red, stamped in the gauge case. On the driver's side, "GA" is the connection to the sender. On the passenger side, "IGN SW" is connected to power controlled by the ignition switch:
Click here to read more about internal repairs of the fuel gauge at The CJ3B Page.
With reliable connections squared away, let's begin the calibration process. While possible to do this on the vehicle, it is much easier on the workbench with a power supply and test leads. A lantern battery makes a great power source for a test of a 6v system. Don't use a power source like a model train transformer, which will bleed an AC signal into the variable DC output. This will throw the calibration way off. Don't ask how I know.
If performing the calibration on the vehicle, there is an important precaution to observe with points-style ignition. If your ignition switch doesn't have an accessory position, the key will have to be left on to power the gauge circuit. To prevent overheating the ignition coil or contact points, disable the ignition. If there isn't a separate fuse to pull, insert a strip of thin cardboard between the contact points:
The calibration is done in two distinct steps. First, by bending the two stops, the sender will be electrically calibrated to the gauge. This will stop the needle precisely on the E and F marks, or a little further beyond if desired. After that, the sender will be mechanically calibrated to the tank by adjusting the length and angle of the float arm. When adjusting the arm to fit the tank, care must be taken not to disturb the electrical settings controlled by the two stops.
With the circuit powered, start by bending each stop to adjust either end of the needle's travel. I prefer to have the needle directly over the E when empty, but you can leave a little extra margin if desired. The choice is up to you. While watching the gauge in the background, the low-level stop is being adjusted in this image:
Here's a close-up of the adjustment to set the needle exactly on the E mark. If you choose to set the needle slightly beyond the E mark, don't go too far or the tip of the wiper arm will contact the side of the sender case. This direct connection will cause an erroneous Full reading:
Repeat the adjustment for the high-level stop. Even though this type of gauge responds very quickly, let it warm up for a few minutes before making the final adjustments. As the gauge reaches normal operating temperature, the reading changes slightly.
Some senders have a different design with rigid, nonadjustable stops. In that situation, the main method of adjustment is to bend the arm between the pivot point and the stops. This will change both the high and low readings, so aim for the best possible split between the two. You might be able to file the stops if a bit more travel is needed.
With the sender electrically calibrated to the gauge, it is now time to make the sender match the tank. Without disturbing the float arm stops, the arm will be adjusted. The arm can be bent anywhere your little heart desires, except between the stops and pivot point. Any changes there would throw off the previous calibration. The arm is stiff piano wire, so pliers will be needed for all but the gentlest bend:
Measure the tank's depth and draw a simple profile on a sheet of paper. Secure the sender's mounting flange to a piece of small angle iron or bar stock. If you try to just hold the sender in your hand, a barely observable tilt could throw off the float by an inch or two out at the end of the arm. (Don't ask) The bar stock under the flange serves as a handy guide to hold the sender level with the top of the tank mockup. Here's the test setup in its full glory, shown with the float resting at the bottom of the make-believe tank. Since the float will ride about halfway submerged, there would be about 1/2" of fuel left when the float first drops to the bottom. Should you wish to have more fuel left in the tank with the needle on E, bend the arm to place the float the desired distance from the bottom. The S-bend in the arm will be explained shortly:
The electrical characteristics of the sender don't match the tank perfectly, as previously noted. If the arm is adjusted for a perfect reading at full and empty, the needle will show approximately 3/8 when the tank is half full. Despite our best intentions, you can only calibrate two points and hope the rest is close. Since accuracy is more critical as fuel is depleted, I decided to pick empty and 1/2 full as my calibration points for the float arm. When the tank is about 7/8 full or higher, the needle will indicate full. Above this level, the float will be submerged with the arm against the high-level stop. This slight tradeoff in accuracy is worthwhile in exchange for greater accuracy at lower levels. Here's the setup with the gauge reading 1/2 full. Note how the float arm is not centered between the stops:
The length of the float arm is very important. The electrical stops control the angle through which the float arm pivots. For a given arc of movement, the float will travel up and down a corresponding amount. Since the angle is fixed, the only way to change the float's vertical travel is by varying the length of the arm. Some universal (Ha!) replacement senders have a telescoping arm that can be adjusted and locked with a set screw.
Determining the proper length of the arm is easy. Set the float at the upper calibration point (Remember I used 1/2 full) and bend the arm for an accurate needle reading. Now swing the arm down to the previously calibrated low-level stop. When the arm length is correct, the float arm will hit the low-level stop with the float in its desired position, be it an inch from the bottom or whatever level you'd like indicated. If the float overshoots at the bottom of the tank, the arm is too long. If the float doesn't travel down far enough before the low-level stop is reached, the arm is too short. It is doubtful you will encounter an arm that needs to be lengthened, and I'm not real sure how best to do that. Shortening the arm is easy enough with an S-bend. Double-check the upper calibration point after forming the S-bend, with a little bit of trial and error to be expected.
Pretty exciting stuff, huh? With only the slightest bit of electrical knowledge, you now have a properly calibrated fuel gauge system. When installing the components, don't forget the gasket between the sender and tank. If the threaded holes extend into the tank, sealer will be required on the mounting screws. Connect the sender's ground wire to one of these mounting screws. With this ground wire and the one for the gauge itself, don't assume (Danger! Danger!) that the body tub is well grounded. The body rests on insulated pads and the ground path must hopefully be completed through one of the mounting bolts. For a trouble free system, run a ground wire directly from the (-) battery terminal to the body tub.
Nearly all of the above details apply to practically any vehicle, although it was written for early civilian Jeeps with separate, round gauges. Military Jeeps were very similar but worked in the opposite fashion, with maximum conductivity (zero resistance) at the sender with an empty tank. Military gauges and senders are not interchangeable with their civilian counterparts due to this fact. Some other gauge/sender applications are also like this, especially after market brands. If a sender has the correct resistance range (0-120 ohms in this example) a simple bend in the float arm will let the sender work with the opposite type of gauge. Once again, here's the unmodified sender from above, showing how the wiper moves to the left as the float rises:
By bending the arm to the opposite side, the wiper now moves in the opposite direction as the float rises:
Most senders are keyed to only fit a certain way in a tank. If there is not enough clearance for the relocated float arm, the keying arrangement will require modification so the arm can be placed in the same relative location in the tank.
Now that Fuel Gauge 101 is online, I'm going to be bombarded with letters something like the following:
"Wow, what a great description of fuel gauges. You are definitely a genius and I'm not worthy to lick the dust from your boots. I'm a high-school drug dealer who just finished putting 44" mudders on my Jeep with a spring-over conversion and a blown 454. Can you help me with the 12 volt conversion?"
Or:
"Greetings, what an incredibly well-written article about fuel gauges. I just inherited my grandfather's Jeep, which was hermetically sealed the day he drove it home from the dealer in 1946. It has been maintained religiously by a visiting Jeep mechanic and is in absolutely pristine, original conversion. Can you help me with the 12 volt conversion?"
Everybody wants in on the lemming-like rush to convert to 12 volts. If you have an early Jeep with the individual round gauges like shown above, that type can run on 12 volts but it is not recommended. A balanced-coil gauge is tolerant of voltage changes but will have a shorter life at higher voltages due to heat buildup. For best results, use a gauge regulator as seen at Antique Automobile Radio, Inc. A gauge regulator will supply a steady 6 volt source for the gauge. This is different than a dropping resistor, which is only adequate for a fixed load. When the sender changes the total current flow through the gauge, a dropping resistor cannot maintain a steady voltage.
There is another type of gauge that is VERY sensitive to input voltage. This type, the thermal gauge, came equipped with its own regulator. More than one gauge, such as temperature and oil pressure, can share the same regulator. The thermal element inside the gauge expands when heated due to higher current flow. Expansion or contraction of the element moves the needle. Since the movement is fairly slow-acting, this helps dampen any oscillations if the sender is bouncing up and down. The Instrument Voltage Regulator (IVR) maintains the 5 volt average for which the thermal element is designed. Just like the dashboard ground on a balanced-coil design, the dash ground for the thermal gauge system is very important. However, the gauge itself doesn't need the ground but the IVR does. This ground provides a reference point for the IVR output. Although the IVR is shown separately, in most applications it is attached directly to the back of the gauge. This ground path is via the gauge case attached to the dashboard, so a separate ground wire as shown is suggested to eliminate potential problems. Here's a diagram of a single thermal gauge installation:
Please note the IVR is different than the Gauge Regulator used in 12 volt conversions, although both perform similar functions. A Gauge Regulator is an electronic device that provides a smooth, 6VDC output for use in 12V conversions. An IVR has mechanical contacts that pulse open and closed, with an output that averages 5VDC. (The slow response of a thermal gauge masks these pulses.) An IVR can safely operate with a 12 volt input, but even better would be a Gauge Regulator feeding a steady 6 volts to the IVR.
Thermal gauges are calibrated at the sender, with the same dual-step process shown for the balanced-coil type. Don't forget to allow extra time for the slow-moving needle to catch up when making any adjustments.
Regardless of what type of gauge is used, there is a bit of misleading terminology to consider when choosing a replacement sending unit. You may hear a reference to a 6, 12 or 24 volt sending unit. The sending unit is nothing but a variable resistor and is not rated for any specific voltage range. The gauges themselves are designed for a certain input voltage, but not necessarily so with a sending unit. As long as the sending unit has the correct resistance range (0-120 ohms in the example above) for the fuel gauge, it will be just fine as far as the electrical portion goes. Expect some confusion, however, when buying a replacement sending unit. Let's say Manufacturer A used a 6 volt gauge for a few years and it was designed for a sender with 20-150 ohms. Meanwhile, Manufacturer B produced a 12 volt gauge that was also designed for a 20-150 ohm sender. The mounting flanges and float arms might be different, but these two senders are the same electrically. A vendor for Manufacturer A might refer to this part as a 6-volt sender, while the electrically-identical part for Manufacturer B might be called a 12 volt sender. Keep this distinction in mind if dealing with a mongrel vehicle. Perhaps you have a vehicle that was originally 6 volts, has been converted to 12, and just to make our example interesting, is using a military fuel tank that only takes a "24-volt" sender. The key to making the system work is not the potentially-misleading "voltage" nomenclature applied to the sender. The important part is picking a sender that has the correct resistance range (and direction of travel) to work with the fuel gauge. Vendors may not have this data on hand, typically selling parts only by model and year. Expect to do a little sleuthing on your own for non-stock applications.
