So my frame of reference her is a 1971 Camaro, with a standard GM 0-90 ohm rheostat style fuel level sender, but I suspect the conceptual circuit I'm about to describe is probably universally applicable.

My aftermarket fuel gauge has a built in buffer/dampener circuit such that the needle doesn't respond to quick transient type of float arm movement... i.e., going around a turn, short term acceleration/deceleration, etc.

But given enough time, like steadily going up or down a hill, a sustained period of acceleration/deceleration, the needle will react to it.

It works "ok", but I'd like to see if I could get the more steady behavior of newer cars.

I searched around for something commercially available, but everything I see is just another glorified add-on dampener circuit.

What I"m envisioning is the following, but I'm a mechanical engineer, not electrical.

A device that only samples the fuel level when on straight/level ground and not in the presence of acceleration/deceleration. So my mind immediately goes to a mercury switch. It would interrupt the signal wire to the gauge when not in a straight/level steady-state condition.

The only problem is, when I disconnect the signal wire by hand, the gauge needle goes to empty, so that would be annoying.

So the other aspect of this device is that when it reads the fuel level, it saves that reading. Then when the device interrupts the signal wire, it sends the last known good reading to the gauge (that way the gauge stays on the current needle position), until such time that the car is back in straight/level steady state... at which point it takes another sample reading of the fuel level and communicates to the gauge.

Any thoughts on how this could be done? Or if there is such a device commercially available or could be adapted to this application?

Thanks!

 

Welcome to AAC!

Certainly doable electronically. There are a lot of bright folks here on AAC. Give us some time and we will come up with a solution.

 

One solution is a 2-axis tilt switch that disables sampling when the switch detects off-level condition. The last reading is held.

 

Thanks! Yes a mercury tilt switch is what I was thinking.

But the question is how is the last reading held?

 

In this age, the answer is a microprocessor taking readings from both the fuel gauge and an accelerometer. Only the readings taken when acceleration is zero are used to calculate the value sent to the dash.

 

Analog version where IC1 and C1 are the holding circuit.
C1 samples the voltage from the Sender and connects to buffer chip IC1.
The tilt or mercury switches detect off- level and activate relay RL1.
Contact RL1A switches the Gauge from the Sender to the buffer chip and contact RL1B disconnects C1 from the Sender in order for C1 to maintain the holding charge.

 

Wow, that's awesome! What part is IC1?

 

Wow, that's awesome! What part is IC1?

Haven't decided yet, depends on how much current the gauge draws at full scale.

 

If you can measure the current through the gauge at full scale will help.

 

Analog version where IC1 and C1 are the holding circuit.

You likely want to add a spike suppression diode across the relay coil to minimize arcing in the mercury switches.

 

What part is IC1?

Likely an op amp with possibly a transistor output buffer to provide the gauge current, as determined by your measurement of that.

 

Inspired by sghioto, here is my contribution.



R1 is the same resistance as the meter. S1 is the tilt switch.
It depends on if the tilt switch is normally open or normally closed. The circuit can be modified for either case.

 

I drew this in a hurry. Tilt switch is two mercury switches or ball switches aligned at right angles, connected in series. Adjust the switches so that they are closed when level.

 

If you can measure the current through the gauge at full scale will help.

This can be done by getting voltage readings across the Sender when the tank is full and 1/2 full.

 

Any idea what will happen to the displayed value when going uphill/downhill for multiple miles?

The "perfect" display will never exist (modern cars might use a mix of fuel gauge signal plus consumption calculation and heavy (digital) low-pass filtering to achieve the perceived "better" display), a close-to-perfect display might require a microcontroller, 3- or 6-axis acceleration sensor and some fancy software.

 

Any idea what will happen to the displayed value when going uphill/downhill for multiple miles?

It would expect the transition from uphill to downhill where the vehicle is basically level, would typically be long enough for the gauge to acquire a new reading.

 

It would expect the transition from uphill to downhill where the vehicle is basically level, would typically be long enough for the gauge to acquire a new reading.

The major problem ist that the display will freeze for the whole way up. You could well run out of fuel while the display still shows >0.
(I know some routes through the Alps where you constantly drive uphill for quite a number of km/miles. Downhill is less critical due to way lower fuel consumption.)

 

You could well run out of fuel while the display still shows >0

Well I would expect the driver to know about how far he can drive based on the fuel level before he starts a long assent.

 

You will have fuel slosh in the tank from accelerating, braking and turning that will continue when the car is finally on level ground. You will still have to average your readings out, and the accuracy is probably no better than taking readings when you are driving.

 

Well I would expect the driver to know about how far he can drive based on the fuel level before he starts a long assent.

If you do not do such ascents regularly, you might be negatively surprised about how much more your vehicle is consuming uphill compared to the usual consumption.