[time-nuts] WWV/CHU

[jimlux]

On 3/29/18 3:49 AM, Attila Kinali wrote:
> On Thu, 29 Mar 2018 03:12:24 -0700
> Hal Murray <hmurray at megapathdsl.net> wrote:
> 
>> What do I need in in order to get time from WWV or CHU?
>>
>> Do I need a fancy receiver as a front end?  Do I have a chance with one of
>> the low cost USB thumb drive size receivers?
>>
>> Is there an obvious software package to start with?  (Linux)
> 
> I think the easiest is GnuRadio... A quick googling lead
> to https://github.com/jasonabele/gr-wwvb
> I don't know anything about it so use at own risk :)
> But at least it seems like something that can be done easily
> on a rainy evening.
> 
> Normal RTL-SDR's do not work for WWVB as they have a lower cut of
> frequency in the range of 20-50MHz...unless you bypass the tuner
> chip and feed the signal directly to the ADC. As IIRC all RTL-SDR
> give you something like 2Msps, that should be more than plenty to
> decode WWVB and related signals. If you feed the RTL-SDR from an
> external frequency source, you should be able to related that
> frequency source to WWV.

The RTL-SDR is an interesting device - I'm putting together a hobby HF 
interferometer with GPS to provide time tags.

Yes, most of the newer parts (RTL-SDR v3, for instance) provide a 
programmable bypass of the front end downconverter (the part is actually 
designed to tune TV signals and the L-band output of a consumer dish LNB)
The backend chip (RTL2832U) is a digital downconverter which mixes and 
filters the nominal 3.5 MHz IF which is sampled at 28.8 MHz

You can actually adjust the output sample rate - something around 2 
Msample/second is the default, but there's lots of other rates 
available.  For WWV you could crank it down, but..
The ADC is 8 bits (7 ENOB) and the output is 8 bit I/8 bit Q.

Folks have modified the RTL-SDR to accept an external frequency 
reference, so you could take the output from your ensemble of Cesium 
references to discipline a hydrogen maser (so your close in phase noise 
is better),then use that to drive a 28.8 MHz discrete divide/multiply 
chain, and run that into your $30 receiver to improve the frequency 
accuracy.  (not for nothing are we called time-nuts)




> 
> 			Attila Kinali
>