• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 315 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AMC1306E25: Bandwidth and max input voltage of isolated modulator

Joel Holland
Intellectual 765 points
Part Number: AMC1306E25
Other Parts Discussed in Thread: AMC1306M25, AMC1336

I am looking to sense the voltage across a MOSFET in a flyback power converter. I start by stepping the voltage down to about 3V.

Since my input voltage is 300V, I need to then isolate this signal such that I can pass the information to my DSP from TI and measure the voltage across the FET safely. The switching frequency is 1MHZ, so I believe an analogue Linear isolator With 300kHz-400kHz bandwidth will not be sufficient in achieving the response speed that I need to fully reflect the high frequency voltage signals.

I came across this device which is an isolated modulator which has a high clock speed.

Since digital signals can have theoretically unlimited bandwidth, would these devices work well in replicating a 1MHZ analogue signal with fast response time, rise and fall time?

The voltage that I will measure will not be a differential voltage. The data sheet mentions that the maximum differential voltage is in the mV range. If we tie the negative analogue input to ground and simply pass a 3V analogue signal to the positive analogue input, will the device still work as intended and replicate a 3V  output at the digital side which accurately reflect a 3V analog at the input side? The data sheet does mention that the max analogue input is equal to the 5V supply rail of the high voltage side, but the application examples only really use it for differential current sensing - I want to use it for ground referenced voltage sensing.

Thanks in advance!

  • 0
    TI__Mastermind 31655 points

    Hi Joel,

    A 1 Mhz input signal is quite fast for this type of device. In the case of the AMC1306, the input is actually bandwidth limited to up to about 800khz due to an internal filter on the differential input. Additionally, the SNR of the device rolls off quite heavily with an input frequency above about 10Khz. This is mostly due to the decrease in the oversampling ratio (OSR). Delta sigma modulators like the AMC1306 operate at a high sampling frequency, but their output is essentially a single bit ADC which produces a bitstream of 1s and 0s related to the input signal. This bitstream is filtered and averaged to produce a multi-bit output at a slower rate. The ratio between the sampling rate and the output rate is the oversampling ratio. You can experiment with the effect of the oversampling ratio in the Delta Sigma Modulator Filter calculator here: 

    You can technically get an output with a 1Mhz input, but it will be extremely noisy and not very accurate, since the maximum sampling frequency for this device is 21Mhz, limiting your OSR severely. 

    As far as the input side goes, you would be better off choosing a modulator with a larger input range like the AMC1336 which is designed with voltage sensing applications in mind. You would still need to divide your input down to fit the +/-1V input range.

    The main concern is definitely the speed. If you need a high degree of accuracy, these isolated modulators will not be able to reproduce such a fast signal.

    Regards,

  • 0 in reply to Scott Cummins
    Intellectual 765 points

    Hi Scott,

    This was my worry as well. I’m glad you clarified the same.

    Do you happen to have any possible alternatives for isolating a 1MHZ analogue signal to my Texas Instruments DSP That can do so more accurately than the isolated modulators while achieving at least 600V isolation?

    Thank you,

    Joel

  • 0 in reply to Joel Holland
    TI__Guru**** 168485 points

    Hi Joel,

    Sorry for the delay - with a low OSR, you might be able to get the signal over the isolation barrier.  Let us look around a little to see if we can figure out another option for you.

  • 0 in reply to Tom Hendrick
    Intellectual 765 points

    Hi Tom,

    Not a problem. I will investigate the matter of OSR a bit more.

    I welcome suggestions for this. I am currently thinking along the lines of using a pulse transformer, similar to which is used for gate drives, to transmit the 1MHz signal. 

    I also thought to use an isolated gate driver or digital isolators but I do not think these will work as the signal is technically analogue. 

    Thanks for taking the time to help.

    Best wishes,
    Joel

  • 0 in reply to Joel Holland
    TI__Guru**** 168485 points

    Hi Joel,

    If you do go down the modulator path, I would suggest you look at the AMC1336 over the AMC1306, you would have a +/-1V input span versus the +/-250mV you get with the 1306. 

  • 0 in reply to Tom Hendrick
    Intellectual 765 points

    Hi Tom,

    Would a capacitive isolator like this work well?

    Thanks,

    Joel

  • 0 in reply to Joel Holland
    TI__Guru**** 168485 points

    Hi Joel,

    Those isolators have faster throughut but they are digital in/digital out.  They would not be able to translate your analog waveform.