Need help determining which potentimeter to get

Thread Starter

pray2crom

Joined Aug 20, 2021
1
Hello!
I am trying to design a new way to accurately measure resistance on a automotive fuel level sender. I have a fuel level sender that uses a variable resistance strip that sends a resistance range between ~0 to 200 ohms to the fuel gauge. The float arm on the fuel sender moves the contact point from 0 to 90 degrees. 0 degrees maximum resistance at 200 ohms, 16 degrees at 150 ohms, 25 degrees at 100 ohms, 46 degrees at 50 ohms, 69 degrees at 25 ohms, and 90 degrees at 0-0.5 ohms. My design replaces the variable resistance strip with a potentiometer. Does such a potentiometer exist? thanks
 

Irving

Joined Jan 30, 2016
3,845
Why? Existing fuel senders work well given their relative inaccuracy which is mainly down to tank geometry. A potentiometer gains you nothing useful.

Most off-the-shelf potentiometers are 300deg rotation so it will give 90/300 or 30% of it's range. A standard 500 ohm unit will give a 150ohm swing, a 1k a 300ohm swing +/- typically 20%, which may include zero depending on how you mount it.
 

zophas

Joined Jul 16, 2021
165
Also most potentiometers are either logarithmic or linear. I doubt the TS would get the resistances he needs unless he uses some clever electronics along with it.
 

Irving

Joined Jan 30, 2016
3,845
Also most potentiometers are either logarithmic or linear. I doubt the TS would get the resistances he needs unless he uses some clever electronics along with it.
90%+ are linear these days, log pots are relatively rare since most audio gear has gone over to digital & remote control.
 

Tonyr1084

Joined Sep 24, 2015
7,853
Hello!
I am trying to design a new way to accurately measure resistance on a automotive fuel level sender. I have a fuel level sender that uses a variable resistance strip that sends a resistance range between ~0 to 200 ohms to the fuel gauge. The float arm on the fuel sender moves the contact point from 0 to 90 degrees. 0 degrees maximum resistance at 200 ohms, 16 degrees at 150 ohms, 25 degrees at 100 ohms, 46 degrees at 50 ohms, 69 degrees at 25 ohms, and 90 degrees at 0-0.5 ohms. My design replaces the variable resistance strip with a potentiometer. Does such a potentiometer exist? thanks
Keep in mind that the fuel sensor is often bathed in fuel; meaning it CAN affect the reading. IF you're going to use it as a water tank sensor - you didn't tell us how you're planning on using this, other than mentioning fuel system level monitoring as an example - water on the pot will affect how it performs. Dirty water will have an even greater affect.

Why are you re-inventing the wheel (fuel level sensor)? Or are you planning on building something else?

Older cars used the resistor in the fuel tank to direct current through a bimetallic strip in the fuel gauge. Change the current and you change how the fuel gauge works. Modern cars "Probably" use information gathered from the fuel level sensor and the computer sends the information to the gauge cluster on the dash. Now, I'm no expert on automotive electronics, but I think (just a guess here) the computer sends a 5 volt signal to the fuel sensor and then gets a reading back based on the level of the fuel in the tank. It may range from 0.6 volts to 4.4 volts, the difference from a full tank to no gas in the tank. Changing the resistor could hurt the computer. If you don't get the values right you could draw too much current through the circuitry and burn something out - or you don't draw enough current through and the computer thinks there's an error in the sensor and will throw a code. I think! I don't know that for a fact. But I suspect this may be the case.
 

Tonyr1084

Joined Sep 24, 2015
7,853
The float arm on the fuel sender moves the contact point from 0 to 90 degrees.
That's quite a lot of movement. Again, since I don't know I can't speak with any kind of authority, but if I remember when I've changed fuel pumps, the fuel level sender was more like 70˚ movement.

I would think having the lesser movement would mean greater accuracy since at near full and near empty slight changes in fuel levels would result in greater movement of the arm than would be the case at half a tank. I would expect the greatest accuracy to occur when the tank is half full (or half empty - depends on how you see it). It's possible the computer has coding to correct for this inaccuracy.
 

Irving

Joined Jan 30, 2016
3,845
Older cars used the resistor in the fuel tank to direct current through a bimetallic strip in the fuel gauge. Change the current and you change how the fuel gauge works. Modern cars "Probably" use information gathered from the fuel level sensor and the computer sends the information to the gauge cluster on the dash.
Bimetallic was used, but vast majority used dual magnetic coil one coil carries current through the sender, the other from the battery, the differential magnetic field sets the indicator position. This arrangement compensates for battery voltage changes.
Modern clusters just measure absolute resistance using, typically, a 5v reference voltage and are calibrated to better match sender position to fuel level, because modern tanks are more closely integrated into the vehicle structure so sender position is less directly related to fuel capacity.
 

Tonyr1084

Joined Sep 24, 2015
7,853
Here's an illustration of what I mean:
1629553236752.png
Of course, the gauge sense isn't set at 90˚, the mid point is typically approximately at 45˚, thus offering greater accuracy. But notice how at 45˚ the gauge should suggest 29% of the fuel has been burned, leaving 61% in the tank. One would expect that at 45% the tank should be half full (or half empty). It's not until you reach 60% deflection that you get 50% angular change in distance from the center point of the radius.
 

Tonyr1084

Joined Sep 24, 2015
7,853
With a sensor mounted at 45˚ as pictured below, using a 90˚ swing you can have as much as 29.26% error. Reducing the swing to 60˚ (15˚ to 75˚) you reduce the error to less than 14%. Even that is pretty high. Since I don't know the geometry of fuel level senders and their gauges, I can only guess. But my whole point was that having a 90% sweep on a potentiometer can introduce quite a bit of error. Now, if you're using an automotive computer to calculate fuel levels then I'd suspect they're pretty accurate. But then again, a fuel gauge isn't meant to accurately represent the amount of fuel, but more so it's to indicate when you're going to need to fill up again. That's the way it has been for many years. I don't think they care if there's a 10% error. After all, it's not meant to tell you how many gallons are in the tank. If you want to know that then you need to calculate the amount of fuel that has been used. My Toyota Tacoma tells me how many miles to empty if I set the gauges to tell me that. And it has been useful. One afternoon coming from South Dakota to Wyoming the gauge told me based on fuel usage that I had over 200 miles to empty. More than enough to get through the stretch of wilderness I was passing through. But during the trip I encountered a strong head wind and was watching the miles to empty versus the miles traveled and I knew I wasn't going to make it across. Had to turn around and find gas, fill up, then head south again. I think my MPG was below 13 against that headwind.

So again, I ask, what are you trying to accomplish?
1629554521933.png
 

Tonyr1084

Joined Sep 24, 2015
7,853
@Irving I remember starting the car and watching the fuel gauge slowly rise to the level of fuel in the tank. Shut the car off and the gauge would slowly fall to empty. I don't think I ever encountered a magnetic fuel gauge. I would imagine they would respond much more rapidly than the bimetallic strip I recall seeing. I think it was from a British car; but that was a very long time ago. I think I must have been around 10 years old. Gosh! Was I ever that young?
 

zophas

Joined Jul 16, 2021
165
Years ago I had a car that used to show me a full tank on fast right hand turns and an almost empty tank on fast left hand turns. Can't remember what car that was but the fuel sensor was really badly designed.
 

Irving

Joined Jan 30, 2016
3,845
Years ago I had a car that used to show me a full tank on fast right hand turns and an almost empty tank on fast left hand turns. Can't remember what car that was but the fuel sensor was really badly designed.
No, that was a badly designed tank without baffles to reduce how much the fuel sloshed around under lateral acceleration - in the worst cases you could get fuel starvation on long bends depending on where the fuel pick-up was.in the tank.
 

Irving

Joined Jan 30, 2016
3,845
@Irving I remember starting the car and watching the fuel gauge slowly rise to the level of fuel in the tank. Shut the car off and the gauge would slowly fall to empty. I don't think I ever encountered a magnetic fuel gauge. I would imagine they would respond much more rapidly than the bimetallic strip I recall seeing. I think it was from a British car; but that was a very long time ago. I think I must have been around 10 years old. Gosh! Was I ever that young?
Many German cars used bimetallic gauges, most UK cars used magnetic. The way to tell is whether the fuel gauge was a 2-wire or 3-wire connection. And yes the magnetic ones respond quicker but were damped to limit fluctuations due to car motion.

1629567541816.png
And below is a partial wiring diagram from my 1965 Triumph TR4A. Item 32 is the gauge and 33 the sender, Items 30 & 31 are the water temperature circuit which works in exactly the same way. Note the ground connection on both gauges. On my instruments the ground is the casing and the connection is made through the retaining clip rather than an explicit wire - corrosion leads to odd readings. Despite the 'balanced coils' approach item 27 is a crude bimetallic voltage regulator just for those two instruments; from memory it regulates at approximately 9 - 10v across the 10 - 14v range of the battery.

1629567641929.png
 
Top