Connect with us

differential op amp to measure current

Discussion in 'General Electronics Discussion' started by eem2am, Nov 10, 2012.

Scroll to continue with content
  1. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    Hello,

    I wish to measure 1 to 10Amps into LEDs.
    I am using a 1mR sense resistor.

    I am using a x100 differential op amp, but should I best have the current sense voltage going below the op amp ground, or staying above the op amp ground?

    Current sense voltage going below op amp ground
    http://i45.tinypic.com/2rxbk93.jpg


    Current sense voltage staying above op amp ground
    http://i48.tinypic.com/29qbps2.jpg
     
  2. john monks

    john monks

    693
    1
    Mar 9, 2012
    Neither will work.
    You need a 200K resistor to join the junction of R2 and the + input of the op-amp to V1.
    Then you need to change R2 to 200K.

    How do you intend to look at the output of the op-amp?
    Is it connected to a meter?

    Please redraw and let's look at it again.
     
  3. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    651
    May 8, 2012
  4. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    This topic i thought was different because in this topic i consider the sense voltage going outside the op amp rails....that is in fact the question....John has brought up a situation though which appears that i have made a mistake which is a mistake similar to my other thread.

    but the subject of this thread, "the voltage going outside the opamp rails" is original

    Here is the schem redrawn as per John Monks advice.
    http://i46.tinypic.com/975y09.jpg

    ...The current source on the left is a 15Amp peak sine at 50khz(!)...the load is LED and i just represent by a 5V source.
    The LED load might be 1A , 2A, 3A.......up to 10A.

    I have to measure it so that i can set the emergency current to 10% of the power received in non-emergency.

    The output of the opamp will be read by an ADC on a pic micro


    i hear you say neither work but unsettlingly they work fine on the simulator...LTspice ....and the LM258 model is one i got from onsemi.com

    current reading doesnt have to be massively accurate, i just have to see if its 1A, 2A, 3A.....10A.
     
    Last edited: Nov 10, 2012
  5. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    651
    May 8, 2012
  6. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    Thanks CDrive, but i already went through hi side with linear.com and diodes.com, ...they dont do it for 100V high side....which i need......and if they do do it for 100V, then you dont get enough amplification.....that is why none of that stuff works for my requirement, which is 1mR sense resistor due to the 10A max current.......becasue of my low rsense, i need 100x amplification and 100v capability...and hi side monitors dont do that

    Also, please could this third option also be considered?...

    http://i49.tinypic.com/rm4dph.jpg
     
    Last edited: Nov 10, 2012
  7. CDRIVE

    CDRIVE Hauling 10' pipe on a Trek Shift3

    4,960
    651
    May 8, 2012
    I deleted my first reply before I posted it because I had so many questions that I was confusing myself! This is because this topic and your last one are definitely related. They both share a common fault though. Extremely pertinent information is being disseminated in dribs and drabs like a politician testifying before the supreme court! Now is not the time to tell us that your supply is 100VAC @ 50KHz.

    Please tell us what this supply is going to power. You stated LEDs, yet your current can be as high as 10A! :eek: And why would you need 50KHz Sinusoidal AC to power LEDs?

    I don't want to suggest anything else until all the particulars are posted.


    Chris :confused:

    Edit: No, I don't see anything wrong with the last low side OpAmp circuit you posted.
     
    Last edited: Nov 11, 2012
  8. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    Yes, i have to admit that this circuitry is very different from you or i or anyone has ever seen before.

    The load is LEDs.....it can either be a 5V LED or a 50V string.....and the max power is 50W, so yes, the 5V LED is a 5V, 10A monster LED.

    The bus that delivers power to the LEDs.....is where i want to measure the LED current...........

    this is going to sound unbelievable, but it is a rectified 50KHz sine wave!!!!....its from a current source, which is 50KHz sine....(please do not ask why)

    Anyway, i have to measure the led current becasue when the mains fails, the 50KHz source shuts down, and i have to power the leds with a battery.......and the power i must give the leds is 10% of what they were getting bofore the power cut....thats why i need to measure the currnt.

    .....i am told that so far the maximum (longest string ) led load is 50V...but there might be a situation of OV transients,

    i am trying to give little away due to "Industrial secrecy", but all that i can say is out now, i assure of this
     
  9. duke37

    duke37

    5,364
    769
    Jan 9, 2011
    You could put a diddy current transformer before the bridge rectifier. This would give isolation and would drop less voltage than a sense resistor.

    The output of the current transformer will need to have a resistive load and a smoothing capacitor.
     
  10. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    thanks, thought the sense resistor is cheaper, and 1mR only dissipates 100mW with 10A...i find current transformers, even small ones , are quite pricey
     
  11. duke37

    duke37

    5,364
    769
    Jan 9, 2011
    You can get a transformer from a dead computer power supply at no cost.

    You appear to want to change supply when the 50kHz supply fails so you do not need to measure it, just detect whether it is there. Use the recitified current transformer output to drive a changeover relay.

    Do you want to revert back to 50kHz when this supply comes back. Measurement of current will not do this if the leds are switched to the alternative supply. A relay to detect the 50kHz voltage may be better.
     
  12. eem2am

    eem2am

    414
    0
    Aug 3, 2009
    So, to summarize, I have a 50KHz sinusoidal current which I must detect. The current is full wave rectified, so its actually pulsating DC half-sines.
    The current is sinusoidal, but may be an average of 1A, or 2A, or 3A, or 4A…………..to 10A.

    I must detect whether it’s a 1A, or a 2A , or a 3A……to 10A, that’s flowing………so I don’t need great accuracy, just to depict which one of these current levels it is that’s flowing.

    Here is a picture of the 10A average rectified current that I must detect…..
    http://i46.tinypic.com/2mxpjdg.jpg


    ….Do you think this is best done with a current sense transformer?

    now …..The DC level for the 10A one, is, exactly that ..10 Amps, and that seems an awful lot to put through a current transformer…transformers don’t like DC, surely.

    I was wondering of the B82801B0205A100 (1:100) current sense transformer is up to this job?…..

    B82801B series C.T. datasheet:-
    http://www.epcos.com/inf/85/ds/b82801b.pdf

    ….the frequency range for almost all off-the-shelf current sense transformers is either 20KHz-50KHz or 50KHz-1MHz.
    …..this makes it awkward for me because I am right on the boundary…i.e. 50KHz.

    Also, the equation on page 5 of the B82801B series datasheet has a “duty cycle” parameter in it……..though my waveform, (shown above) is just a continuous train of half-sines…….so I don’t know what to put in for the “duty cycle” parameter?

    So anyway, do you think the B82801B0205A100 C.T. will be OK for this job, and not saturate?
     
  13. duke37

    duke37

    5,364
    769
    Jan 9, 2011
    I do not know whether the transformer will be suitable. I would guess that the duty cycle should be taken as 1.

    If the load is resistive, you could measure voltage.

    If you put the current transformer before the rectifier, there will be no DC component.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-