[time-nuts] new WWVB BPSK dev board

[jimlux]

On 12/4/18 2:28 PM, John Moran, Scawby Design wrote:
> I'm a little puzzled as to why people keep calling 60kHz - 'RF'. Many Hi-Fi audio amplifiers go higher than that.
> 
> As for an 'RF' front end, there are dozens of off-the-shelf op-amps that can amplify the signal from a tuned ferrite rod aerial sufficiently to then feed into a decent A-D and then a micro.

yes, and someone has to pick one of them and put it on a board, with a 
voltage regulator and other parts.

It's not complex, but it's also not something you can buy off the shelf, 
plug into your RPi and start coding up the filter.


> 
> Analog Devices have A-Ds that digitize the carrier of 12GHz RF signals so I think we should be able to manage 60kHz ... the problem then being to process the resulting data stream.

I don't think the ADC is the challenge - a decent tuned front end with 
some gain so that the signal is around a volt, and you can feed the 
onchip ADC of most microcontrollers and you'll be all set.

If you had a wide open front end, with AM broadcast, all manner of RFI, 
and so forth, you'd probably need more bits.  But a few kHz sample rate 
would probably work, if your sample and hold was good - if it isn't then 
maybe a couple hundred kHz sample rate.

Easy to do with most microcontrollers on an off the shelf board.



> 
> However, fast A-Ds are not particularly cheap and you would need circa 50MHz sample rate to resolve 1deg of the carrier, and TVB has already stated that there are no time benefits to BPSK, so this is all just an interesting technical exercise, isn't it?

Your sample rate doesn't have to be anywhere near the resolution - what 
you do is fit a sine wave to the samples you have and you can easily get 
sub-sample accuracy.  The important thing is the SNR.   Higher sample 
rates do help, because they basically give you that Sqrt(N) improvement 
in SNR over a single sample.