Voltage Sensing Circuit for Low Voltage Values

Thread Starter

Deve

Joined Dec 28, 2015
95
The vintage automotive fuel gauge circuit uses a 0-30 ohm sending unit. The gauge is then just a glorified ohm meter that tells the driver 0 Ohms is Empty, 30 Ohms if full, or anything in between depending on where the float is in the tank. [this is my understanding].

The 0-30 ohms of course wouldn't be possible without some voltage. The voltages are 100 millivolts for empty and 2.1 volts for full. The objective is to make a LOW FUEL WARNING CIRCUIT. I want to be able to sense the voltage that is present without effecting the operation of the existing circuit. When the voltage reaches [ an adjustable range ] of between 0 and 1 volt, I want an LED to come on.

Has this been discussed before? I am looking for all the clues I can get as to what this circuit would look like. Thanks!
 

Thread Starter

Deve

Joined Dec 28, 2015
95
I understand the idea of using a comparator, using the non-inverting terminal as the sense voltage, but what I am not real sure of is reducing the 12 volts down to the 100 to 400 millivolts which is when the LED needs to light up. I will play with a series of voltage dividers to see if I can get there. The important thing is to not have any load on the sending unit in the process.
 

MikeML

Joined Oct 2, 2009
5,444
Voltage across Sender is a function of time, namely it sweeps from 0.1V to 2.1V and back again. Current through the LED as a function of the Sender Voltage at different pot settings. Trip Point is adjustable from 0V to ~1V, as requested. Circuit has a bit of hysteresis, and is powered from the car's electrical system. The TL431 is used as a (pretty good) voltage reference.

170.gif

You may want to use the other half of the LM358 as a slosh filter per posting #8 in this thread.
 
Last edited:

Thread Starter

Deve

Joined Dec 28, 2015
95
Mike this is awesome! Thanks. You mentioned the voltage across the sender is a function of time. Did you mean how it is handled in this circuit or the sender itself? The sender itself doesn't behave that way. 2.1v and 30 ohms are the same, and 100 milllivolts and 0 ohms are the same. To complicate matters, I have to use a reducer because the voltage of the gauge is 6 volts, so I just use a 6.8v, 5 watt Zener to reduce the voltage. Anything in what I just said that changes anything?
 

MikeML

Joined Oct 2, 2009
5,444
The circuit I posted responds to the voltage across the sender. The upper plot pane is a plot of the simulated sender voltage; think of it as filling the tank, and then emptying the tank. The lower pane shows when the LED is on (current flows through it), and off. Here is a 6V version of the circuit. Dont worry about anything inside the dashed box. That is just the stuff it takes to simulate the circuit...

170a.gif

Note that with the pot set to 20% from the bottom, the trip point is ~0.2V (follow the red trace).
Note that with the pot set to 100% from the bottom, the trip point is ~1V (follow the blue trace).
 
Last edited:

MikeML

Joined Oct 2, 2009
5,444
...You mentioned the voltage across the sender is a function of time. Did you mean how it is handled in this circuit or the sender itself? The sender itself doesn't behave that way. ...
Sure it does. When you drive, the sender voltage reading decreases (as a function of time); When you fill the tank, the sender voltage increases (as a fucntion of time). The time scale in reality and in the simulation I posted is totally arbitrary...
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Mike, a few questions.. I will only need the 12v version because all of my stuff is aimed at 12 volts. So I will be using your first drawing to prototype it. When a capacitor has the value but no 'farad' listed, is it customary that it's uF? Is it ok to use one side of a LM393 for the comparator? Someday I want to learn to use LTSPICE but ran across an immediate problem of it having no potentiometer in its library. I use them alot. I noticed that you have one that works. Thanks for your time!
 
Last edited:

MikeML

Joined Oct 2, 2009
5,444
I used the LM358 op-amp as a comparator. I know it has sufficient common-mode range for the circuit to work. I would have to look at the data sheet for a LM393 to see if does...

2.2 without a suffix is understood to be 2.2F, so I screwed that up. I meant 2.2uF. Since it is used as a bypass, LTSpice didn't care...
 

Thread Starter

Deve

Joined Dec 28, 2015
95
Got the parts in today. Mike can you tell me what I would have to do to get the full range to be adjustable? The LED only comes on at 2 ohms or less. That is too near an empty tank. Can we do it so the entire 2.1volts clear down to 0 be adjusted via the potentiometer? That way there is no doubt I can adjust it to a custom tank level. A person who lives 30 miles from the nearest gas station may want his adjusted differently than someone who has a station down the street.

Also, not sure if I am doing something wrong, but the lower ohm values light the LED, but the adjustment doesnt seem to move the ohm value (to light lighting) upwards as it should. I want to be able to move the senders arm up and down, and as I do that, adjust the pot for when I want the LED to come on. That isn't what is happening.

Maybe the above will fix that. Thanks for everything!
 
Last edited:

MikeML

Joined Oct 2, 2009
5,444
...Mike can you tell me what I would have to do to get the full range to be adjustable? The LED only comes on at 2 ohms or less. That is too near an empty tank. Can we do it so the entire 2.1volts clear down to 0 be adjusted via the potentiometer? That way there is no doubt I can adjust it to a custom tank level. A person who lives 30 miles from the nearest gas station may want his adjusted differently than someone who has a station down the street....
What is happening to the Voltage at the Sender terminal as you move the arm up and down.? Is it changing from several volts down to near zero volts? The voltage at the non0inverting input of the opamp should follow this within a few mV (you did use the specified opamp and not a comparator, right?) .

If you run the wiper of the pot to the top, the circuit should trip at 2.1V; If you run the wiper to the bottom, the circuit will not trip at all. It starts tripping when the wiper is up about 20% of the way...

Measure the voltage at the inverting input of the opamp. It should follow the pot adjustment from 0V to ~1.2V. Does it?

The voltage at the cathode of the TL431 should be 2.495V. Is it?
 

Thread Starter

Deve

Joined Dec 28, 2015
95
I am using the correct parts and with the schematic connected correctly, there is less than 900 millivolts at the junction you have listed as 2.49v. Since I am unfamiliar with this TL431, but even if I had it in backwards, still less than 900 millivolts. There is only about 1/10 of the pot where any adjustments make any difference. Since that first sentence is always the one that gets you, I will redouble my efforts to figure out if I did anything wrong. I checked it like 4 times yesterday, but nothing wrong with another look. I have never seen it where you tie the cathode to the base before, but I am not very good at this, so defer to you.
 

Thread Starter

Deve

Joined Dec 28, 2015
95
I tried it both ways. 700mV one way, 900 the other. I'm still looking it over. Nothing so far.
 

Thread Starter

Deve

Joined Dec 28, 2015
95
No, it's wired correctly. When I said I swapped the 741 the other way, I meant I used a new 741 and put it in the other way. Pin 3 in pos, Pin 2 is neg, not sure why its behaving this way.
 

Thread Starter

Deve

Joined Dec 28, 2015
95
TL431. Its a long way across the room to the computer. All the parts are exactly as you stated, all inserted correctly. I even disconnected the sender to see if that would matter. Rechecked resistor values, everything I know to do.
 

MikeML

Joined Oct 2, 2009
5,444
TL431. Its a long way across the room to the computer. All the parts are exactly as you stated, all inserted correctly. I even disconnected the sender to see if that would matter. Rechecked resistor values, everything I know to do.
Then post a photo of the circuit, top and bottom if on vector board, re-sized so that it is less than 2meg...
 
Top