Hi all

I am following Deve's design on a low fuel light for my classic mini utilising the existing fuel sender and gauge. I have attached his design however his design is based around a fuel sender that registers 30 ohms when full and 0 ohms when empty.

On the classic mini the sender registers approx 270 ohms when empty and 30 ohms when full. Also there is a voltage stabiliser between the 12V source and the gauge which regulates the voltage at a constant 10 volts.

What changes would I need to make to Deve's design for my application.

 

Basically you need to switch the + and - leads to the functioning op amp ... The lower triangle in the diagram. Besides that, you need to get a voltmeter and adjust the movable wiper arm of the 50k potentiometer to a reference voltage that you want the LED to turn on at ... A value of 9 volts on the + pin should give about 10% warning ... as the signal from the tank approaches 10 volts. You can change that whenever you want.
... Make sure that the circuit has a solid, efficient chassis ground connection.
If any of this isn't clear, or if there is something unusual showing up ... just saying that it should work, but there might be something else going on.

 

... something doesn't look right with that TIL431 regulator. The anode and cathode shouldn't be connected together. .... Have to check that out.

 

The TL431 is drawn wrong, you need the Ref pin connected to the wiper of the 50k, and set the preset to read10V at pin 5.

 

Basically you need to switch the + and - leads to the functioning op amp ...

There is a bit more to it than that. the 3.3M resistor adds about 0.3% hysteresis to the input signal; not enough to do much (in my opinion), but something to deal with when swapping the opamp inputs. With the inputs swapped, the reference is driving the non-inv input so hysteresis has to be added there. Since the reference is adjustable, that complicates things a bit.

Two solutions ...

a) leave the inputs alone, drive the 2nd opamp as an inverting LED driver.

b) swap the inputs, and use the 2nd opamp as a reference voltage buffer to drive a hysteresis resistor.

ak

 

There is a bit more to it than that. the 3.3M resistor adds about 0.3% hysteresis to the input signal; not enough to do much (in my opinion), but something to deal with when swapping the opamp inputs. With the inputs swapped, the reference is driving the non-inv input so hysteresis has to be added there. Since the reference is adjustable, that complicates things a bit.

Two solutions ...

a) leave the inputs alone, drive the 2nd opamp as an inverting LED driver.

b) swap the inputs, and use the 2nd opamp as a reference voltage buffer to drive a hysteresis resistor.

ak

Hi AnalogKid

Thanks for the reply, really appreciated.

Can I also move the 3.3M resistor with the swapping of the inputs or is this not possible. If not are you able to explain why (I've been furiously reading about op amps and hysteresis to see if I can work this out).

 

The TL431 is drawn wrong, you need the Ref pin connected to the wiper of the 50k, and set the preset to read10V at pin 5.

Thanks Dodgydave and drc_567 - I am looking into this more in how TL431 is wired to work out what is going on here (theoretically this is a working solution so need to understand a bit more here)

 

Thanks Dodgydave and drc_567 - I am looking into this more in how TL431 is wired to work out what is going on here (theoretically this is a working solution so need to understand a bit more here)

Here is the full article...

http://devestechnet.com/home/warningsystem

They are using the TL431 as a 2.5V zener, and lowering the comparator reference voltage to set the low fuel switch on level.

 

Here is the full article...

http://devestechnet.com/home/warningsystem

They are using the TL431 as a 2.5V zener, and lowering the comparator reference voltage to set the low fuel switch on level.

Thanks Dodgydave - I have read all this and the diagrams match the PCB layout which is a working solution. However it is mentioned by yourself and drc_567 that the wiring looks wrong