[time-nuts] WWVB Dephaser Question

[Graham / KE9H]

Paul:

The easiest way to convert a single sample into an I-Q representation is to
have a two phase local oscillator.
You multiply the incoming signal against the primary phase output of the
oscillator for the I channel,
and multiply the same sample against the 90 degree offset output for the Q
channel.

The quadrature oscillator can be as simple as two sine look-up tables,
whose entries are shifted by 90 degrees.
The only math is in the sample multiply, and creating the look up tables to
start with.

If there is a ratio of integers relationship between the sampling rate and
the desired local oscillator frequency, then you can get a real clean
implementation with reasonably small look-up tables.  When running there is
no transcendental math involved, just a simple multiply.

--- Graham

==

On Sat, Oct 24, 2020 at 7:37 PM paul swed <paulswedb at gmail.com> wrote:

> Antenna and filtering are not a problem for me. I use the 10' X 10' square
> loop and about 800' of wire with a cap to resonate at 60 KHz and then a
> preamp really to drive 140' of coax. Currently using a modified KD2BD
> receiver frontend but using a 350 Hz xtal filter. Yes indeed that hamfest
> find was amazing.
>
> Looks like you have to build a costa loop SDR. Looked at lots of documents.
> My first question to the group.
> Can a single incoming sample be converted to I & Q through math. I sort of
> think so. The original sample might be I and then use math to make Q. That
> would save an entire analog chain. The teensy has a single chain for mic
> level input. Though it does feed left and right channels. Granted there are
> two audio chains at line level.
>
> The teensy has a NCO that can create a quadrature output. The alternate
> would be to use that to drive analog switches as a multiplier to form I & Q
> samples.
> Regards
> Paul
> WB8TSL
>
>
>