Automotive fuel gauge and sending unit issue

[axlmastr]

This is a common issue for some hotrodders needing to replace their old tired fuel tank sending units with a replacement and matching it to a fuel gauge they wish to use in their cluster.

Item #1 Fuel gauge - mechanical type with sweeping needle. The resistance range of the gauge is 30-240 ohms. At E (empty) the gauge expects to see 30 ohms and at F (full) it expects to see 240 ohms.

Item #2 Sending unit in tank - mechanical type sweeping arm with a dimmer-style wire coil and wiper. The resistance range of the sending unit is 2-36 ohms. Some are more like 0-30 if that makes much difference. When the sending unit is at empty it reads 2 ohms and at full 36 ohms.

Is there a way to match the two items so that sending unit's position will relatively match the needle position on the gauge value for value?

I have posted on this site's forum for several years and have performed hundreds of board level repairs. Thanks in advance

[stj]

Re: Automotive fuel gauge and sending unit issue

first you need to recognise that your not dealing with resistance, but current.
then you can compensate with an amplifier.

[redwire]

Re: Automotive fuel gauge and sending unit issue

It sounds like a fun project.

You would need a circuit to amplify the signal, and possibly invert it because some senders are low resistance empty, others low ohms when full. You could do it with a couple op-amps.

The hardest part is the gauge driver, these old gauges need several watts of drive, so a transistor there will make a little heat.

At least three different types of old car fuel gauges:
1. "balancing coil fuel gauge" with two coils, an Full and a Empty coil.
2. "Thermal type", one heating coil and a bi-metallic strip, which is probably matched to the really low resistance senders. Usually a 'voltage stabilizer' used. 1958 Ford, VW Beetle.
3. Two coils, one is a field coil.

1,3 are current-operated. With a car's voltage constant i.e. 13.8V, then the tank rheostat's changing resistance gives a changing gauge current.
2 heating is I^2*R so a parabola for needle deflection and current.

That challenge is linearity, where 1/2 tank is accurate. If you really want that, then you have to have the sender's curve known and probably use a little MCU to correct the curve.

I could work on a circuit but don't have a gauge cluster to test it on, unless I hit up a wrecking yard.

A modern approach is to use stepper-motor gauge motors in the cluster.

[petehall347]

Re: Automotive fuel gauge and sending unit issue

can you not try resistor connected in the right place to get the ohms matching ?

[redwire]

Re: Automotive fuel gauge and sending unit issue

Old GM 1930-1964 senders are 0-30 ohms; 30 ohms=Full. So the gauge wants to see Full=30 ohms, Empty=0 ohms.
New aftermarket senders are 240-33 ohms; 33 ohms Full, 240 ohms Empty.

So you need to invert the signal and scale it, if I'm hearing OP correctly.
Hot rodders want to keep the old (6V or 12V) instrument cluster but use a modern tank sensor.
Some Ford and Chrysler gauges do not have the backwards reading problem.

You could use an Arduino, or stay analog and use a bunch of op-amps.
I have to do research on controlling old gauges.

[Curious.George]

Re: Automotive fuel gauge and sending unit issue

This is a common issue for some hotrodders needing to replace their old tired fuel tank sending units with a replacement and matching it to a fuel gauge they wish to use in their cluster.

Item #1 Fuel gauge - mechanical type with sweeping needle. The resistance range of the gauge is 30-240 ohms. At E (empty) the gauge expects to see 30 ohms and at F (full) it expects to see 240 ohms.

Item #2 Sending unit in tank - mechanical type sweeping arm with a dimmer-style wire coil and wiper. The resistance range of the sending unit is 2-36 ohms. Some are more like 0-30 if that makes much difference. When the sending unit is at empty it reads 2 ohms and at full 36 ohms.

Is there a way to match the two items so that sending unit's position will relatively match the needle position on the gauge value for value?

All you've tied down are the extremes of sender/gauge. I assume that's not all that you care about in the response of the system (?).

So, without knowing (reproducible) the actual characteristics of the sender and gauge over there full operating range, the easiest way to KNOW you have a solution that can work is to implement the transfer function digitally.

"read" the sender value for a given amount of fuel in the tank (start with an empty tank and systematically add some known amount each iteration) and record this value.

Similarly, drive the gauge with a particular "output" and observe the position of the needle (and how the driver would INTERPRET that... is that "half a tank"? etc.).

Then, just build a piecewise linear approximation to the transfer function using these input and output points.

[An 8 pin MCU would suffice at << $1]

The hard part will be getting enough drive capability to handle a low impedance gauge. But, you could probably hack together a dog slow PWM current sink and count on the mechanics of the gauge to filter out the switching frequency.

[redwire]

Re: Automotive fuel gauge and sending unit issue

I've done car fuel gauge control for propane and natural gas conversions.
I use a microcontroller to read tank pressure, convert and drive the car's original (gasoline) gauge.

Covering 80 years of different cars, US and England, there are about 10 different fuel senders out there.
They are known and documented, just a rheostat. Getting this into a MCU is straightforward.

I add correction factors for non-linear fuel tanks and gauges, matching car maker psychology.
3/4 is usually 70% fuel, 1/4 tank 30%. Empty is around 13% fuel remaining, a few gallons.

The hard part is that old gauges (using 30 ohm senders) are bi-metallic and need several watts to heat up and move the needle. Some cars have voltage regulators like 8V supplying the gauge.

PWM did not work well as the gauge would whistle or buzz and make sound.
Higher frequency made switching noise and some voltage regs did not like it.
So I use linear control.

If I could get more information about resistance of old gauges, it would be OK to design something that could work with 6V or 12V gauges and any sending unit out there. It would end the hot rodder's troubles.