[time-nuts] Near-perfect chip for Loran-C frequency receiver

Poul-Henning Kamp phk at phk.freebsd.dk
Fri Jul 4 06:20:45 EDT 2008

In message <486DF228.6000202 at xtra.co.nz>, Bruce Griffiths writes:
>Poul-Henning Kamp wrote:

>> Look at the subject, it says "... frequency receiver".
>The meaning diverged/became ambiguous after Magnus made his post.
>My comments apply to the extended capability receiver that Magnus 
>envisaged and not to a Loran only receiver.

Not for me.

I guess you assumed Magnus wanted to receive the DCF77 timecode,
but that's not how I read it, I assumed he wanted to use DCF77 as
a frequency reference, in which case you can integrate for days if
you want to.

>But locking to a 1uV signal in the presence of a 1V signal [...]

What is your point here ?

I'm not trying to build the receiver to end all receivers, I'm
trying to build a simple (one chip), cheap (<EUR100) SDR that will
allow us to lock local OCXO's or Rb's to LoranC or possibly VLF
signals like DCF77.

I have found out that 12bit/1Msps and a trivial low-pass filter
copes nicely with both the local 246kHz LW boomer, the 200 km close
Loran-C on Sylt and I can still drag out pretty much anything
transmitted in the LF/VLF range anywhere in Europe, that might be
of interest to a time/frequency nut: (see: http://phk.freebsd.dk/loran-c)

Obviously such a receiver will not lock my Rb to WWWB which is
halfway around the world from here, but I don't care to that in the
first place: the wander from the ionosphere would make it useless
from a time&freq point of view.

(But see also: http://freenet-homepage.de/df6nm/MSFgaps.htm and don't
miss the rest of his site.)

>Actual signal levels for MSF, WWVB etc as well as signal levels would be 
>useful in sorting out the limitations of various approaches to combined 
>Loran, MSF/WWVB/etc receivers.

It is actually impressive what you can do if you throw enough CPU
power at it.  You can trivially get 16 bit 1MSPS ADC's today,
giving you close to 100dB dynamic range.

But that would be quite a high-end project: somebody would have to
knit together a moderately complicated PCB with enough DSP power
on it to make it worthw,hile and programming would be a non-trivial

My proposal for this project requires a ferrite-bar, a couple of
OP-AMPs and a prototype board you can buy for EUR50, and you
can write the software in plain C which can be compiled with GCC.

If you live a place where you have 1V/m signals and try to derive
frequency fom a 1uV/m signal, or if you live in a faraday-cage
at the bottom of a iron-mine, this project may not be for you.


Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the time-nuts mailing list