How to modulate resistance value

Discussion in 'The Projects Forum' started by rwp289, Sep 30, 2010.

  1. rwp289

    Thread Starter New Member

    Sep 30, 2010
    11
    0
    How could I "slow down" the change in resistance value of a sending unit?

    I have a fuel level sending unit that works on a 10 ohm (full) to 50 ohm (empty) scale. The maximum ohm value is 73.

    When the tank is not full, and I go around a corner, the gauge moves alot as the fuel moves to the side and the float is lowered. I would like to design a circuit that updated constantly but took the average of the last XX seconds - where XX is either adjustable or something like every 45-60 seconds. If it was updated only once a minute or so with the average reading of that period, it would take out most of the swings.

    Can I do this with stuff I could order from Radio Shack, Mouser, etc.?

    thanks,

    Randy
     
  2. wayneh

    Expert

    Sep 9, 2010
    12,140
    3,054
    I think what you need is called an integrator, and the simple parts you need are readily available even at the Shack.
     
    rwp289 likes this.
  3. someonesdad

    Senior Member

    Jul 7, 2009
    1,585
    141
    Basically, you want a low pass filter. A simple RC (resistor-capacitor) circuit could probably do this. Suppose you wanted an RC time constant of 100 seconds. Then RC = 100. Pick a capacitor, say 100 μF. Then solving for R in your head gives 1 MΩ. You can easily get such parts from Radio Shack.

    You'd put the capacitor in parallel with the load and charge this capacitor through the 1 MΩ resistor. The output goes to the gauge, which I'm assuming is a voltmeter. Hopefully, this will be adequate to stabilize the reading. You might have to experiment with different values. Make sure the capacitor is rated to two or three times the circuit's maximum voltage.

    Some of the EEs here will speak up if I've made some bonehead mistake... :p
     
    rwp289 likes this.
  4. rwp289

    Thread Starter New Member

    Sep 30, 2010
    11
    0

    Thank you for the detailed explanation. I am more mechanical than electronic so I will need to read up on the circuit schema.

    The cirucit is 12v DC.

    When you state "in parallel with the load", I assume that to mean that the existing lead (wire) that transmits the sending unit level (ohms) to the gauge is left intact and that the RC circuit is added along side (in parallel) to it? So, when the resistance changes (float in tank moves) the capacitor "cushions" the change.

    Forgive my ignorance.
     
  5. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    I have a fuel level sending unit that works on a 10 ohm (full) to 50 ohm (empty) scale. The maximum ohm value is 73

    This tells me that you've got one of the older bimetal fuel gauges and they draw a fair amount of current, a simple resistor and cap aren't going to hack it.

    The ideal method of course would be to get to the sending unit and add some mechanical damping but it would have to be reliable such that it wouldn't stick on you, also most cars require you to drop the tank to get at the sending unit.

    The first thing to do is get some actual voltage readings at the sensor at various levels. You later make a separate voltage divider circuit with the sensor and another resistor that approximates that gauge being in there, then the voltage level could be processed and converted back to drive something that would be capable of driving the gauge.
     
  6. rwp289

    Thread Starter New Member

    Sep 30, 2010
    11
    0

    Thank you for the reply. The original gauge was a bimetal unit but the one I replaced it with is electronic. That is the source of the issue as the bimetal unit had built in damping and the new one does not. Does that change your mind about the RC option?

    Your idea about taking voltage readings is a good one.

    Do you know of a place to view a schematic of a voltage divider cirucit? With an electronic gauge, which method (RC vs Voltage Divider) is preferred. Since I am a novice, simple is better for me but not so simple that it does not work.
     
  7. Bernard

    AAC Fanatic!

    Aug 7, 2008
    4,173
    397
    It would be helpfull to know meter input resistance. With fulltank, add resistance [pot] between sender and meter; add resistance untill meter reads 1/2 tank- thats about meter resistance. Increase C to 1000 uF, drop R to 100k & a darlington transistor emitter follower might do.
     
  8. rwp289

    Thread Starter New Member

    Sep 30, 2010
    11
    0

    I had to look that up but it looks like a really good idea. Thank you for the post!
     
  9. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    It's going to be difficult because of having that type of sensor, note that it barely changes resistance over its full range.

    The bi-metal gauges also didn't go from 0 - 12V or anything like that, they worked more as a current meter in that the amount of amperage that could flow through the gauge + the sensor's resistance would heat the bimetal strip to a certain temperature thus a certain displacement position on the scale.

    Was this gauge designed as a replacement for a bimetal unit? If so, it's got to have some sort of internal resistor to simulate the old gauge - that or it's working more as an ohmmeter.

    Measure the resistance across the new gauge itself, that should tell us something we can go on.
     
Loading...