Loading ...
Sorry, an error occurred while loading the content.

Re: I and Q: at what frequency is the phase difference 90 degrees?

Expand Messages
  • alan_r_cam
    My own $0.02 worth: In a SSB receiver, you can get audio by mixing the IF signal with an oscillator, then running thru a low pass filter. The SDR uses a
    Message 1 of 15 , Jan 30, 2012
    • 0 Attachment
      My own $0.02 worth:

      In a SSB receiver, you can get audio by mixing the IF signal with an oscillator, then running thru a low pass filter. The SDR uses a similar trick.

      A pure 7.01MHz signal, mixed with a 7MHz clock, should give you a 10KHz signal (and a 14.01MHz signal- a low pass filter will get rid of that). Sounds easy...

      BUT we don't have a pure 7.01MHz signal. We have a whole band o' crud from the antenna, including noise at 6.99 MHz. this will ALSO get mixed down to 10KHz and interfere with our desired signal. How to eliminate this?

      Analogy: Stereo FM. The main FM signal is L+R, and there's a second signal L-R (up a few KHz and rarely mentioned in FM theory classes). Add these to get L, subtract to get R. That's how the I and Q signals work: I+Q = 7.01MHz signal, I-Q = 6.99MHz crud.

      OK, how to create I and Q? They are samples taken with a 90 degree phase difference. We get them by manipulating the CLOCK signal. A 4x clock (28MHz), gives us one clock cycle for every 90 degrees of a 7MHz signal. It gives us 4x signals, which are split as:

      (1)I, (2)Q, (3)-I, (4)-Q (and repeat).

      Signals (1) and (3) are summed and filtered in a Tayloe detector, and form the I channel. Signals (2) and (4) form the Q channel.

      Q: At what frequency is the phase difference 90 degrees?
      A: At the centre frequency of the band being sampled.

      Note: There are other ways of doing this, besides a Tayloe detector. If you use a single fast ADC, you can "split" this into I and Q signals at a quarter of the ADC clock rate. The problem: the ADC clock is constant, so you need a constatnt IF stage - which complicates the design.

      EXAMPLE-
      1. Use SI570 at 34-63 MHz, create IF1 at 33MHz.
      2 Mix again with 32MHz XTAL, for IF2 at 1MHz.
      3. Sample with TI ADS1675 (24-bit 4MS/sec)
      4. Alternate samples between I and Q,
      and make every 2nd channel sample negative.
      You now have 24-bit samples at up to 500KHz, rivalling the best sampling available in a conventional sound card.

      Regards to all,

      - Alan Campbell
      Adelaide, South Australia
    Your message has been successfully submitted and would be delivered to recipients shortly.