adapting a gas tank sending unit (potentiometer)

Discussion in 'General Electronics Chat' started by shortmort37, Jul 20, 2016.

  1. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    A firm I know manufactured a number of gas tank sending units with the incorrect resistance range. The range should be 20-225 ohms - instead, it's 10-125 ohms.

    I thought some combination of a parallel resistor in combination with a series resistor might do it, but to achieve the high end, the series resistor would have to be >= 100 ohms, and in the low end, it would have to be <= 20. So, it's no realizable. Looks like I need an active circuit of some sort.

    Any ideas for something cheap and easy?

    adTHANKSvance,
    Dan
     
  2. benta

    Member

    Dec 7, 2015
    101
    24
    Not a lot to do, except telling the "firm" to work better next time. No resistor combination will get you out of this one.
     
  3. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    Yeah. Think I said that, when I said it wasn't realizable.

    Agreed, no passive circuit will do it. But, an active one?
     
  4. Aleph(0)

    Member

    Mar 14, 2015
    343
    325
    Sry if this is OT but those rheostatic sending units scare the bejesus out of me:eek:! I know O2 concentration SHOULD always be too low for explosive limits but $#!t happens:eek:!
     
  5. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    The circuit doesn't need to be anywhere near the gas tank, it could be under the dash, and powered by 12 volts. It just has to present 20 ohms to the gauge when the rheostat is at 10, and 225 ohms when the rheostat is at 125. I have no plans to touch the tank, or the sending unit.

    So, can I get a reply from someone who can answer the question?
     
  6. Aleph(0)

    Member

    Mar 14, 2015
    343
    325
    Shortmort37 Sry I wasn't scolding you:oops:! It was just general comment on commercial sending units:)

    You can come close with opamp translator setup to simulate 2X effective resistance. So the range to gauge is 20Ω to 250Ω. But I say easiest way to have it exact is with uC with onboard ADC then you could compensate for nonlinearity with formula or just lookup table:)

    However you do it for safety plz make sure that current through the sender is limited to the low uAs!
     
    Last edited: Jul 20, 2016
    Hypatia's Protege likes this.
  7. BReeves

    Member

    Nov 24, 2012
    412
    64
    Do a Google search for "voltage controlled variable resistor". Should give you a few ideas..
     
  8. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    Can we assume the the voltage on the sender is regulated?

    And do you have a print of the input circuit? And what accuracy did you expect with the in spec sender?

    This is a super question. There is always more than one way to skin a cat. But, we need all the information.

    The solution will depend on existing circuitry. One might start with changing sensor voltage.
     
  9. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,010
    1,530
    @shortmort37 , as a mechanic this would be a nightmare down the line and over time. When replacement time comes it would mean having two different part numbers for one car, one with the fix and one without. I retired from Delphi Packard Electric/GM and they would never stand for doing this. I saw over the years many defective parts scraped because of manufacturing faults, it's part of business. A"quick fix" now will end up costing your company a lot more down the line.
     
    Aleph(0) likes this.
  10. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    Hi @shortbus, I don't disagree - but, this is for a 1959 Plymouth! For me, it's more of an academic question: Is there a remedy that can be provided for folks who have bought this incorrectly engineered part, without having to go to the trouble of draining their gas tank and dropping it to replace it (and yes, someone is now making the correctly ranged part). In principle, it seems like it ought to be doable. The idea of using a voltage (or, current) controlled variable resistor is intriguing, as @BReeves suggested.

    @BR-549, to the extent that the impedance of the gauge is approximately fixed, the voltage on the sending unit will vary with current. I have a wiring diagram, but it's not really a schematic. The circuit is really simple, though: A grounded potentiometer, center tap wire from tank to dash gauge, gauge to 12 volts. I don't see +/- 3% being significant.
     
  11. SLK001

    Well-Known Member

    Nov 29, 2011
    817
    227
    So what's happening is that you're getting twice the current delivered to your dashboard meter (gauge) than the meter was designed for. Maybe a shunt resistor on the meter would work.
     
  12. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    Hmmm... Particularly, if the impedence of the meter is low relative to the sending unit (should be). This is an interesting proposal.

    A user could adapt to, say, the gauge indicating 3/4 full when topping off. It's important, however, to have a good sense when you are approaching empty!

    I'll explore this idea. Thanks for the suggestion!

    Dan
     
  13. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    I never met a gas gauge that I didn't have to get used to. And rarely have I seen a linear one.

    Let's see the print. We want to verify that the far side of gas gauge is grounded.

    I would try a six volt (couple of amps) source on the sensor. Using common car ground.

    My strategy would be to adjust sensor voltage for an accurate 0 to 1/4 or so, span on gauge, with a full scale reading close to full or a little over.

    This will take some time, because you need to siphon some gas. And if I remember right, these have a large tank. And be real careful with fumes. A six or twelve volt battery can spark when making a connection.

    And please don't try to pump gas with a non certified fuel pump. Or simply drop one tank for this prototype work.

    If a sweat spot voltage can be found, then a simple voltage regulator to the sensor will do.
     
    Tonyr1084 likes this.
  14. crutschow

    Expert

    Mar 14, 2008
    13,009
    3,233
    I also think a shunt resistor across the gauge may do the trick, perhaps somewhere in value to the gauge resistance.
     
    BR-549 likes this.
  15. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    A shunt would be the easiest solution, but it will increase the operating current.

    What is the current rating on the out of spec sensor?
     
  16. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    I'm not sure it's going to be helpful, but the wiring diagram can be found here...
     
  17. shortmort37

    Thread Starter New Member

    Jul 20, 2016
    7
    0
    I don't know the current rating on the sending unit, nor the impedance of the gauge coil (which could be a factor, if it's a fair fraction of the low resistance of the sender, and I intend to halve it). But the wire is 18 gauge, if that's any indication (of course, you don't want to come close to 9.5A).
     
  18. BR-549

    Well-Known Member

    Sep 22, 2013
    1,987
    388
    The power comes from ignition switch, to the gauge, not the sensor. Disconnect the gauge from the ignition switch and vary the voltage there.
     
  19. Tonyr1084

    Active Member

    Sep 24, 2015
    535
    86
    Hi. I'm jumping in from having read as far as post #10. I see this is a 1959 car. I believe the gas gauge was nothing more than a slow reacting amperage gauge. In other words, the more current that flowed through the circuit the more the gauge changed in readout. Not knowing if 20 Ω versus 225 Ω ran the gauge from full to empty or vice versa. So I have to make a few assumptions.

    First, assuming the 20 ohm range represents a full tank. Assuming a base voltage of 12 volts (actual is likely more like 13.8 volts) that's going to run 600 mA through the gas gauge. I believe in some cars there was a bi-metalic strip wrapped in a coil of wire. With the change in amperage the gauge would change its readout.

    Now, assuming 225 Ω, on a 12 volt system, your gauge should see 53 mA when the tank is empty. With the lower resistance (out of tolerance sending unit) the current would be higher. It would seem to me that a bypass resistor across the gas gauge might bring the "Full" / "Empty" swing of the gauge back into a descent range. But that would depend on the resistance presented to the circuit from the gas gauge. For instance, if it has a resistance of 5 Ω then another 5 Ω resistor rated for 15 watts might be the key to fixing the problem without having to drain the tank and installing a new sending unit.

    Sure you can use A/D - D/A converters or an op amp to modulate the current, but I think the simplest solution MIGHT be a parallel resistor across the gas gauge. Easy enough to remove if ever you install a proper sending unit.

    You're on the right track thinking about a parallel resistor, but this isn't going to be on the sending unit because a parallel resistance will only reduce the already low resistance from the sending unit. And a series resistance will limit you from full to half tank at best (or vice versa - half full to empty, whichever is the case).

    Check the resistance of your gas gauge. An old car like that won't be running any kind of smart circuitry. It's basic sense and response. The meter moves based on how much current is passing through the system. And it reacts slowly so it doesn't read from a quarter tank to empty just because you took that turn a bit too fast.

    [edit]

    Also keep in mind the cars wiring was designed to handle a certain amount of amps. The faulty sending unit will send higher current through the wiring. AND adding an additional resistance in parallel to the gas gauge - you could end up burning up your wiring harness. Since I don't know what gauge wire is present, adding a parallel 15 watt (from the example) resistor could over-tax the wiring system. In short, - you may end up with a short. And a fire. And there goes your nice vintage 59 Pontiac.

    [edit #2] - guess I should have kept reading. SLK already proposed the idea I so happily jumped in with. Still, consider the cars wiring. That's equally important here.
     
    Last edited: Jul 21, 2016
  20. Tonyr1084

    Active Member

    Sep 24, 2015
    535
    86
    BR549's proposal, drop the voltage to the sending unit sounds good. Reduced voltage and reduced resistance should swing you back into range without risking any of the wiring. Worth exploring.
     
Loading...