mesuring low level ac voltage

Discussion in 'General Electronics Chat' started by aspectus, Dec 12, 2007.

  1. aspectus

    Thread Starter New Member

    Dec 12, 2007
    4
    0
    Hi
    I'm trying to make circuit to measure resistance of probe in liquid.
    EC if being exact.

    I have set up series of op-amps to generate clean sine wave.

    and my output sine wave range from 0.055 V ac to 5 V (Peak, not RMS but i guess it makes no difference .. i have to pick one and go with it)
    Frequency between 1khz to 2khz i can go up if needed but not much down...


    Well everything is great except i can't convert AC to DC to input it to 12 bit ADC
    Right now i have rectified ac output to full wave pulsating DC same amplitude (very close)

    but I'm stuck with it ... i tried all kinds of capacitor filters to smooth it out ...
    but it signal looses it's amplitude so i get false results ... plus the adc can't get steady measurement ... so here I'm am...

    Here what I'm thinking please help me out here:

    option 1: go with square wave from the beginning so then i rectify it, it will be very constant reading .... But i don't know how will it impact the accuracy of the measurement of EC (electrical conductivity of liquid)

    option 2: convert output sine wave to square and then rectify and be like option 1 but sine wave to take measurement ...

    i think option 2 is best but i don't know.... how much results be different after sine to square conversion ... I'm trying not to lose my accuracy at every single part of circuit ;)

    I have read on this forum thread http://forum.allaboutcircuits.com/showthread.php?t=5294
    but i can't understand the schematic included ...
    specifically the analog switches .. and the values to use ... for resistors and capacitors ....

    maybe someone could give me a different schematic or clarify this one

    PS: my Freq is 1-2khz ... i tried 20khz at first and i had problems with gain ... it wouldn't be even close ... what is the Magic freq for op-amp i have TL062CP Texas instrument ... ?
    I have looked up data sheet ... but couldn't understand it ... what is it called what I'm looking for ?

    Higher the freq the better (supposing) is the measurement from probe...

    sorry for so many q's .. but I'm lost right now
    Thank you ....
    It's nice to join this community ...
     
  2. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    I'm not sure what the best approach would be, to measure the resistance of a liquid. (By the way: What does "EC" stand for, in this case?)

    But, assuming you're going to do it per your description, with the sine waves, with analog circuits, it should work to full-wave rectify them and then low-pass filter the result. The low-pass filter will give you the _average_ value of the rectified signal. No, it's not the same as the peak value. It is lower, as you noted. But you should be able to account for that in the design of the "calibration" of the instrument, since there is still a known, unique one-to-one correspondence between the peak and average values. Since you're converting it to digital, anyway, you could do that in the software.

    Here's a post that has more details about (and links to schematics for) doing the analog RMS-to-DC conversion (the last post on page one of the thread):

    http://forum.allaboutcircuits.com/showthread.php?t=8348

    Alternatively, since the resistance should not change too quickly (I assume), you could use a (resettable) "peak detector" circuit, which is like an "envelope detector" or "envelope follower" but without a bleeder/droop resistor in parallel with the peak-holding capacitor. With the right setup, you should be able to get almost no droop between the peaks, with a true peak detector (probably uV per hour, if done right). But that requires very low-leakage devices. And you have to have a reset capability (in case the peak value decreases), like a mosfet or switch that can periodically drain the capacitor. And that also has to be a very low leakage circuit, to get good accuracy. You can find everything you need, eventually at least, by doing searches for "peak detector circuit OR schematic", and probably also "sample and hold circuit OR schematic", at http://www.google.com , AND AT http://groups.google.com . Also search at the EDN website. And search at national.com, analog.com, and linear.com, at least. And also go to the uspto.gov site and look up patent # 4362996, by Hewlett Packard.

    I would also make sure that you have something like LTspice from linear.com , so you can quickly simulate different setups.

    But also note that, if you could just sample the original sine waves fast-enough, it would then be trivial for your software to find the peak amplitude.

    - Tom Gootee

    http://www.fullnet.com/~tomg/index.html
     
  3. John Luciani

    Active Member

    Apr 3, 2007
    477
    0
  4. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    Conductivity through the liquid is probably poor. Along with the electrode resistance, it amounts to a source that is being loaded badly by the rectifier and capacitor filter. The cure is a good op amp to buffer the electrode output before applying it to an RMS circuit.
     
  5. aspectus

    Thread Starter New Member

    Dec 12, 2007
    4
    0
    Thanks.

    EC = electrical conductivity = 1microSimens = 1/1000000 ohms
    which you multiply by factor to get ppm or total dissolved solids (TDS)

    I like the idea of rms to dc IC .... i'm looking at different kinds right now .... not a lot of choice :(
    I'm never used rms to dc ... i never even herd of it before ...
    Will out put from it be strait dc without ripple voltage ?
    I will try to get by with Square waves while i wait for the part ....
     
  6. RiJoRI

    Well-Known Member

    Aug 15, 2007
    536
    26
    Would it be possible to use the resistance of the liquid as the input resistance to an op-amp amplifier? The source could then be a sine wave instead of a square wave. Amplify the signal more, if needed, then rectify and filter it.

    --Rich
     
  7. aspectus

    Thread Starter New Member

    Dec 12, 2007
    4
    0
    what's the way i have ...
    out put is from 0.055 to 5 V AC freq about 1.5 Khz ..
    and i have it rectified . but can't filter it right ... riple voltage is messing up the adc ... the reading is not steady ...
     
  8. aspectus

    Thread Starter New Member

    Dec 12, 2007
    4
    0
    what's the way i have ...
    out put is from 0.055 to 5 V AC freq about 1.5 Khz ..
    and i have it rectified . but can't filter it right ... riple voltage is messing up the adc ... the reading is not steady ...
     
  9. John Luciani

    Active Member

    Apr 3, 2007
    477
    0
    If it is working properly ;) As the name says it has a DC
    output. Analog Devices doesn't have a part that will work
    for you? You could try Linear Technology. The Analog Devices
    application notes look detailed enough so that you could build
    one.

    (* jcl *)
     
  10. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    If what you are measuring cannot change too quickly, then simply add more lowpass filtering! i.e. Either lower your filter's cutoff frequency (e.g. increase RC's capacitance), or add more filter stages (poles), or both. You can make the ripple as small as you want. i.e. It can virtually completely disappear.

    Note that if you eventually get the lowpass filter's cutoff frequency so low that your measured (i.e. filtered) output level cannot track i.e. follow, quickly enough, changes in what is being measured, and there is still too much ripple, then you would need to either increase the number of poles in the filter, or, you could try use a higher frequency signal, so it would be easier to filter without affecting lower-frequency dynamic tracking capability.

    I recommend downloading the free FilterPro software, from Texas Instruments, at http://www.ti.com . It will design and give you opamp-based filters' schematics and response plots, in seconds, for filters with up to ten poles.

    Or, alternatively, since your output is being sampled, maybe you could just average it out with software.

    - Tom Gootee

    http://www.fullnet.com/~tomg/index.html
     
  11. BeeBop

    Member

    Apr 25, 2006
    17
    0
    Use a precision rectifier. (a bridge in the feedback loop of an op amp.)
    You will probably have to come up with a way to get rid of any off set.
    You will have to scale it after this, so you might want to run your ac closer to the rails of the op amps you are using, and after rectifying, use a final stage to scale your voltage. You didn't say what voltage you were running them at, did you? Single or dual polarity?
     
Loading...