Precise constant current source

Discussion in 'The Projects Forum' started by jining, Oct 18, 2012.

  1. jining

    Thread Starter New Member

    Oct 18, 2012
    1
    0
    Hello everyone,

    Now I need to build a precise constant current source to supply a NTC.
    In the attachment is my circuits.

    Now the problem is , it is able to provide a constant current source, but during I change the value of NTC, the constant current source has a 1uA tolerance. Any idea to get rid of this , and to be more precise?

    Thanks!!

    Jining
     
    • 4.png
      4.png
      File size:
      11.8 KB
      Views:
      123
  2. MrChips

    Moderator

    Oct 2, 2009
    12,452
    3,371
    The problem with any resistive temperature sensor is the self-heating effect. It is better to measure resistive sensors with a very low current that can be turned off in between samples.
     
  3. eblc1388

    Senior Member

    Nov 28, 2008
    1,542
    102
    The easiest way is to increase the gain of the current regulating transistor by forming a Darlington pair with another transistor.

    See Image.

    [​IMG]
     
  4. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Change the NPN to a small N-channel MOSFET for basically perfect regulation and higher voltage compliance (the output can go closer to +2V on the low side).
    You can use 2N7000, 2N7002, BSS123, BSS145, etc.
     
  5. DickCappels

    Moderator

    Aug 21, 2008
    2,664
    634
    You circuit looks fine to me, especially considering that you have not provided any desired specifications, so it is not possible for one of us to decide what cost-performance tradeoffs might be appropriate.

    By the way, the LM358 is a very old though cheap opamp, you can probably find one with less input bias current and lower drifts at a low price.
     
  6. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    He said the current varied 1uA over a range of load values (I paraphrased). Apparently that was too much.
     
  7. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    R3 is the resistor which has to be temp stable for the circuit to get maximum accuracy. There are some very stable .01% precision resistors available, I have a set I use for calibrating meters.

    BTW: the 2V reference also has to be dead on stable. I suspect your error problem is coming from that more than the resistor.

    You will also need a precision op amp since the input offset voltage is in the error term as well. LM358 won't cut it.
     
Loading...