RMS Current Control Based on LTC1968

Dear Team,

 

I want to make current control via MOSFET based on LTC1968. The load will be full wave rectified sinusoidal ( not filtered ) and frequency will be max 400hz.

My design is in the attachement.

  1. I have used U1-LTC1968 and U2-LT1782 for RMS to DC convertion with DC accurate post filter according to Figure 13 of the LTC1968 datasheet.
  2. I have used U3-LTC6226 as gain amplifier for comparator and ADC input.
  3. U4- ADCMP600 is comparator for driving the mosfet via U5- 1EDN7512B gate driver. Comparator’s reference voltage supplied by U6- AD5760 DAC.
  4. U9- AD4001 ADC for readback value of the current.

 

According to application needs is this design reliable and fast enough for fast and precise current control? Do you advice any improvement or changing?

Best regards,

PDF

Parents
  • 0
    •  Analog Employees 
    on Feb 26, 2021 7:26 PM

    Hi aliaykan:  The LTC6226 seems like overkill on the bandwidth compared to the LTC1968 and LT1782.  Also, it has high bias current which will compromise the accuracy of the LTC1968.  For a gain of 2, I would recommend another LT1782.

    Regarding the speed of the LTC1968 in such a control circuit, while the input bandwidth is quite high, the output bandwidth is relatively low, as it is a DC output after all.  The output bandwidth can be increased by reducing the output capacitor and increasing the f3dB of the filter, but this will cause output ripple and output noise to rise.  I think an upper practical limit on the speed of the LTC1968 resolving a sinusoid RMS would be about 4 cycles, or 10ms at 400Hz.   For an example of how these RMS-to-DC's work in a loop, you can see the compressor application here:

    https://www.analog.com/en/technical-articles/ltc1966-true-rms-to-dc-converter.html

    The loop using LTC1966 settled the amplitude in about 50ms.  

Reply
  • 0
    •  Analog Employees 
    on Feb 26, 2021 7:26 PM

    Hi aliaykan:  The LTC6226 seems like overkill on the bandwidth compared to the LTC1968 and LT1782.  Also, it has high bias current which will compromise the accuracy of the LTC1968.  For a gain of 2, I would recommend another LT1782.

    Regarding the speed of the LTC1968 in such a control circuit, while the input bandwidth is quite high, the output bandwidth is relatively low, as it is a DC output after all.  The output bandwidth can be increased by reducing the output capacitor and increasing the f3dB of the filter, but this will cause output ripple and output noise to rise.  I think an upper practical limit on the speed of the LTC1968 resolving a sinusoid RMS would be about 4 cycles, or 10ms at 400Hz.   For an example of how these RMS-to-DC's work in a loop, you can see the compressor application here:

    https://www.analog.com/en/technical-articles/ltc1966-true-rms-to-dc-converter.html

    The loop using LTC1966 settled the amplitude in about 50ms.  

Children
No Data