Voltage Controlled Current Sink

Discussion in 'General Electronics Chat' started by Federov, Sep 20, 2010.

  1. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    I know there is a lot of information on what im looking for but I believe my situation is a tad bit different than most, feel free to correct me if im wrong. In this design i can not use an opamp, though I can build my own basic one which is what i have attempted to do. I also can't use any type of packaged regulator. I have supply rails of 20 volts avialable and am trying to drive LEDS with a total load resistance of 25 omhs. The LEDS can be driven with up to 700mA but i only care to sink 300mA - 400mA. The problem with my circuit is being able to adjust the load and keep a constant current. For a range of 1-about 30 omhs it stays constant anything higher it begins to fail. I believe my limitation is my rails but i would like more concrete evidence. The circuit i have currently built is attached, feel free to post a better circuit or give suggestions. Thanks in advance.

    [​IMG]
     
  2. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Why the restriction on op-amps?

    You should recall that MOSFETs are voltage controlled, it may be easier to use a BJT if you're allowed. If I recall correctly, it would be linear according to Ic = β × Ib. However, remember that β (hfe) varies widely between transistors, it's also very dependent on temperature and Vce/Ic, so you might want to use feedback somehow (it's up to you.)

    Now doesn't an op-amp sound easier... ;)?
     
  3. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Thank you for the quick response, much appreicated. The restriction on op-amps is due to the professor wanting to make the design harder. Are you suggesting replacing X7 with a BJT or all the MOSFETS? The design i showed you actually is a differential amplifier connected with feedback (node V2). I agree that an op-amp would be a much easier design but replacing my circuit with one seems to still lead to the problem that different loads create different currents and to my professor this is unacceptable =/.
     
  4. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Here is a current sink using a MOSFET and an op-amp.

    See if you can use your diff-amp to create feedback, imitating the op-amp.

    If you want to simulate this circuit then use Paul Falstad's excellent Java circuit simulator: http://www.falstad.com/circuit/ (paste this code into File > Import.)

    Code ( (Unknown Language)):
    1.  
    2. $ 1 5.0E-6 10.20027730826997 50 5.0 50
    3. a 224 240 320 240 1 15.0 -15.0 1000000.0
    4. R 176 176 176 144 0 0 40.0 5.0 0.0 0.0 0.5
    5. 174 176 176 224 272 0 1000.0 0.0050 Current
    6. g 176 272 176 288 0
    7. f 320 240 368 240 0 1.5
    8. r 368 256 368 320 0 0.01
    9. w 368 320 224 320 0
    10. w 224 320 224 256 0
    11. r 368 320 368 384 0 10.0
    12. g 368 384 368 400 0
    13. R 368 176 368 144 0 0 40.0 20.0 0.0 0.0 0.5
    14. w 368 176 368 224 1
    15. x 416 297 582 303 0 24 Sense resistor
    16. x 416 359 475 365 0 24 Load
    17. x 416 252 574 258 0 24 Pass element
    18. x 67 235 157 241 0 24 Current
    19.  
     
  5. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,962
    1,097
    Federov
    In your circuit is impassible to get more current then 10V/30 = 333mA
    Becomes there is not enough positive voltages.
     
  6. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Tom66 thanks again for the quick response. From what you sent it looks like your actually sensing the voltage across the load? Maybe im mistaken. I simulated it and varying the load resistor varies the current being produced, which was my original problem. Are you not seeing the same thing when you simulate it?

    Jony130
    The circuit I submitted was just one attempt at making it work i have tried 20v-ground rails as well but it just led to a smaller load variance for a constant current.
     
  7. Jony130

    AAC Fanatic!

    Feb 17, 2009
    3,962
    1,097
    Try modify the circuit
     
  8. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Sorry about that - In my clutter, I mixed up the sense and load resistors. Also, the sense resistor was too small. 1 ohms works better.

    Try using this circuit:

    Code ( (Unknown Language)):
    1.  
    2. $ 1 5.0E-6 10.20027730826997 42 5.0 50
    3. a 176 192 272 192 1 20.0 0.0 1000000.0
    4. R 128 128 128 96 0 0 40.0 0.5 0.0 0.0 0.5
    5. 174 128 128 176 224 0 1000.0 0.0050 Current
    6. g 128 224 128 240 0
    7. f 272 192 320 192 0 1.5
    8. r 320 208 320 272 0 10.0
    9. w 320 272 176 272 0
    10. w 176 272 176 208 0
    11. r 320 272 320 336 0 1.0
    12. g 320 336 320 352 0
    13. R 320 128 320 96 0 0 40.0 20.0 0.0 0.0 0.5
    14. w 320 128 320 176 1
    15. x 368 249 424 255 0 24 Load
    16. x 368 311 438 317 0 24 Sense
    17. x 368 204 524 210 0 24 Pass element
    18.  
    The sense resistor drops a voltage according to the current flowing through it.

    It regulates the current independent of the load resistor up to a point (about 20 ohms for 500mA), because it is a non-ideal current source. As such it cannot generate a voltage and push 1A through a 10 megohm resistor.

    The load resistor can also be moved to the top in which case it acts as a current sink. With a base resistor, a BJT can be used.
     
    Federov likes this.
  9. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Alright so i started to realize why with my rails the current source would only support a certain load.. I just wasn't thinking clearly. Tom the circuit you sent works perfectly for the specified load. I changed my circuit to almost exactly the same that you demostrated and it works well for the the load I desire, thank you. Now its time to work on the actual load circuit. Oh how i love electronics =) Thanks again for the help!
     
  10. Federov

    Thread Starter New Member

    Sep 14, 2010
    26
    0
    Actually speaking of the load circuit do you by any chance have a good scheme for a switch across an LED. I was thinking of using a darlington pair but im not that familiar with their operation besides that they amplify the base current. The load in the circuit will consist of 3 LEDs and each one needs to be able to be switched off an on. Maybe i should start a new thread? lol
     
  11. tom66

    Senior Member

    May 9, 2009
    2,613
    214
    Well, a simple NPN transistor would work. Or a MOSFET. With a logic level signal, no Darlington pair is necessary.

    If you use it with the circuit posted and control the noninverting input, you can also turn the whole output on and off. This has the advantage that the op-amp does not try to saturate when it maxes out the voltage going through a nearly infinite resistance, which could also damage the LED if the voltage is briefly applied (causing a surge of current.)

    Remember when paralleling LEDs, or any device for that matter, load current is shared. So if you were to put switches on each of the outputs, then it would cause the two or one remaining LED to take the output current, which wouldn't be what you want.

    The simplest solution to this problem is to have three current sources/sinks.
     
Loading...