Calibrating fuel gauge in antique truck

Discussion in 'Automotive Electronics' started by Circuits123, May 25, 2018.

  1. Circuits123

    Thread Starter Active Member

    Dec 7, 2012
    64
    0
    I'm trying to calibrate the fuel gauge in my 1970 Land Rover. When there's gas in the tank, it correctly reads full, but the needle goes down to empty too quickly while there's still plenty of gas in the tank. The gauge is made with a bimetallic strip connected to the needle. It takes 10v through the voltage stabilizer. The sender in the fuel tank is a variable resistor with a float. All the parts are new(ish) but from different vendors. When I tested the sender I get about 12 ohms when the float would be at its highest in the tank and 311 at the lowest. But when connected to the gauge, I get these results:

    12 ohms = F (the needle can go higher than full but doesn't with this sender) and 9.8 volts at the back of the gauge
    50 ohms = 3/4 tank
    73 ohms = 1/2 tank
    97 ohms = 1/4 tank
    135 ohms = E
    311 = < E (needle at the bottom) and 0 volts

    (I couldn't test the voltage between 3/4 to E because I had to hold the sender in one hand to move the float.)

    Does anyone have a suggestion of how to fix this? I suppose what I need to do is extend the range of resistance coming from the sender so it's from 12/Full to 311/Empty rather than 12/Full to 135/Empty. Or maybe do something to the gauge? My electrical knowledge is pretty limited so any basic suggestions would be welcome.
     
  2. dl324

    AAC Fanatic!

    Mar 30, 2015
    8,505
    2,047
    Are you sure there isn't a hole in the float and it's filling with gas?

    One of the fuel gauges in my truck has that problem. When I run that tank to just about empty, the gauge reads correctly for awhile after being filled, then it eventually reads empty; even if I haven't even started using that tank. In my case, it's not worth fixing; have to remove the bed to get to the fuel tank.
     
    shortbus likes this.
  3. Circuits123

    Thread Starter Active Member

    Dec 7, 2012
    64
    0
    I replaced the float too. The old one did have a little gas in it but probably not enough to make much difference. I made all the measurements above by pulling the sender out and just moving the arm up and down to see what it did to the gauge.
     
  4. dl324

    AAC Fanatic!

    Mar 30, 2015
    8,505
    2,047
    If you have everything out of the fuel tank, why not just replace the sender?
     
  5. Circuits123

    Thread Starter Active Member

    Dec 7, 2012
    64
    0
    That's the problem - I did replace the sender, but it doesn't seem to match the gauge, even though it's made specifically for this model truck.
     
  6. dl324

    AAC Fanatic!

    Mar 30, 2015
    8,505
    2,047
    Trying to alter the calibration of the sender would likely be an exercise in frustration. Unless the shape of the gas tank is very uniform, the sender resistance won't be linear. It may very well be piecewise to account for the volume of gas at any position.
     
  7. autotuner

    New Member

    Oct 2, 2012
    8
    2
    Perhaps I'm reading this wrong and over simplifying things, but perhaps try a resistor in parallel with the sender, around 300 ohms?
     
    Circuits123 likes this.
  8. dl324

    AAC Fanatic!

    Mar 30, 2015
    8,505
    2,047
    Doing that would affect all readings.
     
  9. crutschow

    Expert

    Mar 14, 2008
    22,456
    6,569
    That would help.
    240Ω in parallel with 311Ω gives 135Ω, so E on the gauge should then occur when he sender is at 311Ω.
    That still would give a non-linear indication, as the 50% float position would give ≈97Ω for about a 1/4 full gauge reading.
    But that is a big improvement compared to the sender without the resistor, which gives 150Ω at the 50% float position, for a gauge reading of below E.
     
    Last edited: May 26, 2018
    Circuits123 likes this.
  10. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,926
    2,397
    Here's a suggested adapter circuit configured to map a 12-311 Ohm sender to a 0-135 Ohm one pretty accurately :-
    GaugeAdapter.PNG
    The simulation plots are the gauge voltage using your actual sender and a reference voltage which would be obtained with a simulated 0-135 Ohm sender.
    R3/C1 damp the sender voltage as an anti-slosh measure.
    From the measurements you have given it seems that the effective gauge resistance is around 3 Ohms, but the sim works with various values.
     
    Last edited: May 26, 2018
    Circuits123 likes this.
  11. Circuits123

    Thread Starter Active Member

    Dec 7, 2012
    64
    0
    Cool. I'll give it a try. But I don't think the anti-slosh measure is needed because a bimetallic gauge is slow to respond to changes.
     
  12. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,926
    2,397
    I was probably in the middle of editing the image when you tried downloading. Now edited, but here's a larger version of just the schematic for clarity :
    AdpterCircuit.PNG
     
  13. ian field

    AAC Fanatic!

    Oct 27, 2012
    6,543
    1,197
    Gauges of that type I've seen were rated 9V - the regulator itself was a bi metal strip "simmerstat". Constantly switching on & off, relying on the thermal inertia of the bi metal strip in the gauge. There isn't actually a stable regulated voltage you can check is right. See if any other bi metal strip gauges are misbehaving - it could be the regulator needs attention.
     
  14. KeepItSimpleStupid

    AAC Fanatic!

    Mar 4, 2014
    3,436
    631
    I worked on a gauge from a 1968 chrysler, it turned out to be the gauge in the dash after dropping the tank. No fun. Me kid, unable to drive helping dad or should I say dad helping me. I did the dx.

    In any event there was a pulsing regulator, but the gauge wasn't damped. So, the system had 3 basic parts: 1) A pulsing regulator; 3) An ammeter and a sender.

    In a 1982 Toyota that I did not have troubles except I could not use the gauge to help tell when the tank was full. That gauge was damped with viscous oil. It also had a low fuel light, but that used the self-heating of a thermister.

    A relative had a Thunderbird and access to the sender was from the trunk which was cool. Aside; Gas fill doors behind the license plate, flip a tail light and two tanks on a Jaguar were other cars I was aware of.
     
  15. crutschow

    Expert

    Mar 14, 2008
    22,456
    6,569
    For extra info, here's the simulation of Alec_t's circuit with the circuit equivalent resistance at the gauge plotted versus the sender resistance.
    As shown, at the empty resistance of 311Ω, the equivalent resistance is ≈140Ω

    upload_2018-5-26_19-29-53.png
     
  16. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,926
    2,397
    Tweaking R2 in my circuit will vary the simulated resistance. 0.75Ω was chosen as the nearest standard value to the theoretical 0.77Ω.
     
  17. crutschow

    Expert

    Mar 14, 2008
    22,456
    6,569
    An easy way to tweak that value, of course, is to start with 1Ω and add a larger resistor in parallel to get the desired value, e.g., 1Ω in parallel with 3.3Ω gives 0.767Ω.
     
  18. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    9,926
    2,397
    Agreed. It's all somewhat academic, of course, given the non-uniform shape of a typical fuel-tank and the non-linear gauge scale :) .
     
  19. gerty

    AAC Fanatic!

    Aug 30, 2007
    1,239
    344
    If it's a one wire sending unit I would also check the tank ground...
     
    shortbus likes this.
  20. LDC1615

    New Member

    Nov 25, 2018
    1
    0
    The issue with the fuel gauge reading has a simple solution. Many older fuel gauges are highly damped, meaning that the needle movement is very slow to respond. This was done so that the users would not see wildly different fuel levels every time they looked at the gauge especially when the fuel was sloshing around in the tank. Most gauges show full when 10 volts was applied to the meter.

    Install a small IC voltage regulator to limit the fuel gauge's maximum voltage. The 7810 voltage regulator will output 10 volts, however, this voltage regulator is not commonly available. The solution is to install a 7809 voltage regulator and lift the center ground with 2 rectifiers in series. The rectifiers will lift the ground circuit by 1.2 volts, so the 7809 will output about 10.2 volts. The rectifier cathodes should point to ground. You can also use an LM317 or LM337 with additional circuitry, AND you must electrically to isolate the IC from chassis ground. The 7809 will also work without modification, however, the gauge will never read 100% full.

    Many old cars and farm tractors used a bimetallic strip that slowly opened and closed as the "voltage regulator". These were also called "voltage stabilizers" and were made by Smiths Instruments. This essentially provided 12 vdc for a few seconds, then no voltage for several seconds, with the on/off cycle repeating continuously.. Since the fuel gauge's needle was slow to respond, the needle did not move much when the voltage was removed.

    Attached is an article that I found on the internet prepared by Doug Lawson that provides an excellent explanation of the issue.
     
Loading...