High impedance DC monitoring

Discussion in 'The Projects Forum' started by arpex, Sep 20, 2009.

  1. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    I want to measure DC voltages over long wires, so I intend to use a twisted pair with a shield to eliminate induced noise. The voltage will be up to 15VDC maximum. I want to feed this into the ADC input of a microprocessor, so the input needs to be scaled down to be 4.5VDC max, say. I only have a 12V and a 5V supply. I have no experience with op amps, but I am sure that a low pass filter, coupled with a high impedance op amp should be able to do this easily. I don't know where to start. :confused:

    Can someone help me please?

    Thanx in advance, Arpex
     
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    You don't really need to use an op amp to scale down the voltage. A simple voltage divider with two resistors can be enough. Use 1K and 3K resistors to scale the voltage down by 3.
     
  3. nirmal_rockin

    Member

    Aug 7, 2009
    19
    0
    try using voltage regulators or appropriate resistors....
     
  4. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    My concern is noise filtering to get a stable reading for the ADC. I expected that a suitable op amp would have desirable common mode noise rejection. These are terms I know a little about and I presumed that these would be good features to have in this situation.
     
  5. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    With an op amp you may introduce more noise than the existed because op amps generate some noise themselves.

    If you are interested to measure only the average value of the DC voltage and not spikes etc just put a 10uF capacitor across the ADC input of the uC.
     
  6. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    Okay. An op amp may introduce some noise, but it also protects the ADC input from the outside world. The resistor divider also loads the voltage source I am trying to monitor. If I use high value resistors, does that make the input more susceptible to noise as well? What if I introduce the incoming scaled down voltage to an op amp with a gain of 1 and the output goes to the ADC input of the uC? That would protect the uC. If so, what type would be suited?
     
    Last edited: Sep 21, 2009
  7. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    What kind of DC source is it?

    Why you want to measure it?
     
  8. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    You could also do the same thing with resistors and capacitors, passively. Is the input a Vin and ground?

    Since we don't know what you're measuring we also don't know how much loading is too much.
     
  9. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    I am trying to make it a generic input, so a resistor divider is okay if I keep it relatively high around 100k, I guess. As I said, it is the noise rejection I don't know about. For a measurement of a positive voltage ( Vin) referred to ground, the divider is adequate, but I want a clean signal and the ADC input isolated from the real world, protected by the high input impedance of an op amp.
     
    Last edited: Sep 21, 2009
  10. russ_hensel

    Well-Known Member

    Jan 11, 2009
    818
    47
    I like op amps. A high resistance voltage divider will also have a high output impedance, a no no for the adc. You might want to put the op amp near the voltage source so that the long leads are on a low impedance source. Or to go further use a differential output to help get rid of induced noise, or convert to a current loop. All this assumes you have power at the 15v source end. If the op amp is rail to rail on 5 v it cannot overdirve the adc. Using a 4 v reference signal on the adc will make sure you have the full dynamic range of the adc if the op amp is not really rail to rail.
     
  11. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    What I want is only one pair of wires bringing the dc voltage to this input circuit, nothing going out to pre-condition it. I want an op amp that works rail to rail on a 5VDC supply to provide the scaled down input DC signal to the ADC input. Alternately, an op amp working on 12VDC, but with the gain less than one to scale the 15VDC (max) input to 4.5VDC. I want the flexibility to measure anything up to 15VDC. If this op amp circuit is high impedance, then it should cover all these conditions.
     
  12. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    You cannot use an op amp power by 12V to measure 15V. Use a voltage divider resistor network and feed it to a rail to rail op amp wired as voltage follower and powered by 5V.
     
  13. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    That's right. We have already discussed that the input signal can be scaled down simply by a voltage divider ie two resistors. I'm sorry if my terminology is not correct - that's what I was trying to explain in my previous message. That's why I put forward that this scaled down voltage is the input to a unity gain buffer to protect the ADC input. I think we are saying the same thing. I have drawn 2 circuits, which are identical, except for power supply to op amp. Please advise if this would be a suitable design and which would be better. What types of op amps wpould be appropriate?

    Thanks, arpex.
     
    Last edited: Sep 22, 2009
  14. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    If you use a rail to rail op amp I suggest to use 5V supply because the output voltage will not go more than 5V and destroy your uC.

    A voltage follower does not need any external resistors connected to it. Just connect its output directly to its inverting input and apply your signal on the non-inverting input.
     
  15. arpex

    Thread Starter New Member

    Sep 20, 2009
    8
    0
    Okay. Thanks. I was of the impression that the high value resistors added protection to all inputs/outputs.:)
     
  16. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Using high-value resistors will add noise.

    You should have decoupling caps to ground to eliminate the noise.

    Voltage followers have the output connected to the inverting input using a zero-Ohm connection.

    10 MEG on the input will delay the charging/discharging time of the opamp's input.

    Remove R4, and connect the lower input to the opamp's ground.
    C1 is too large. Use 0.1uF or smaller.
    Replace R3 with a straight piece of wire. Opamps with CMOS or JFET inputs already have a very high impedance.
     
    Last edited: Sep 25, 2009
Loading...