Searching \ for '[EE] Radio Clock receiver' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/timers.htm?key=clock
Search entire site for: 'Radio Clock receiver'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Radio Clock receiver'
2005\06\14@165010 by Jake Brownson

picon face
I was just reading about how the radio clock works, and the protocol
used... seems like this would be a neat thing to do in a PIC project,
anybody attempted this before? As far as I can tell.. Would need a
60KHz receiver and a timer to fire 3 (or probably more) times a second
to check for high or low power, and some code to decode the signal.
I'm actually working on a project involving a RTC, and do not want to
allow the user to change the time manually, so this might be good if
it doesn't involve to many external components...

For the receiver, Wouldn't a simple loop antenna tuned for 60KHz and a
LC resonator, work? Then maybe use a cap to even out the signal, and
transistor/opamp to amplify, maybe not in that order? Then just bring
that into an analog input? Am I on the right track?

Here's some info on the signal:
http://electronics.howstuffworks.com/question461.htm

Thanks for any guidance

~Jake B

2005\06\14@175239 by marcel

flavicon
face
Hi,
not exactly what you were asking, but I do know that Telefunken, and probably
some others, make chips that take care of all that stuff. You tell it which
continent you're on and they give you the proper time and so forth.
- Marcel

Jake Brownson <spam_OUTjbrownsonTakeThisOuTspamgmail.com> wrote:

{Quote hidden}

2005\06\14@175249 by marcel

flavicon
face
Hi,
not exactly what you were asking, but I do know that Telefunken, and probably
some others, make chips that take care of all that stuff. You tell it which
continent you're on and they give you the proper time and so forth.
- Marcel

Jake Brownson <.....jbrownsonKILLspamspam@spam@gmail.com> wrote:

{Quote hidden}

2005\06\14@175830 by Peter Onion

flavicon
face
On Tue, 2005-06-14 at 13:50 -0700, Jake Brownson wrote:
> I was just reading about how the radio clock works, and the protocol
> used... seems like this would be a neat thing to do in a PIC project,
> anybody attempted this before?

Lots of people LOL  It's a pretty standard "rights of passage" sort of
thing...


> As far as I can tell.. Would need a
> 60KHz receiver and a timer to fire 3 (or probably more) times a second
> to check for high or low power, and some code to decode the signal.

Once you are "in step" with the 1sec data, you only need 1 sample/sec,
but to keep in phase you need many more....

> I'm actually working on a project involving a RTC, and do not want to
> allow the user to change the time manually, so this might be good if
> it doesn't involve to many external components...
>
> For the receiver, Wouldn't a simple loop antenna tuned for 60KHz and a
> LC resonator, work?

You really want a hi-q (= narrow bandwidth) filter afterall the data
rate is only 1 bit/sec.

> Then maybe use a cap to even out the signal, and
> transistor/opamp to amplify, maybe not in that order? Then just bring
> that into an analog input? Am I on the right track?

You can do it that way.  The first one I ever built used op-amps all the
way.  But I did live within 50km of Rugby !  Now I use a commercial
module from galleon
http://www.atomicclockrugbymsf.co.uk/OEM-Receivers-MSF.htm

> Here's some info on the signal:
> http://electronics.howstuffworks.com/question461.htm

Official info for UK station http://www.npl.co.uk/time/

Latest display on my MSF clock can be seen at
www.btinternet.com/~Peter.Onion/PhaseData/p002148.jpg
The signal shown at the bottom is the data from the previous minute.

GOOGLE for "MSF Clock PIC" for more examples

Peter



2005\06\14@182558 by Mike Hord

picon face
> For the receiver, Wouldn't a simple loop antenna tuned for 60KHz and a
> LC resonator, work? Then maybe use a cap to even out the signal, and
> transistor/opamp to amplify, maybe not in that order? Then just bring
> that into an analog input? Am I on the right track?

One of the right tracks, maybe, but a harder-than-necessary one.

First of all, 60 kHz is a very friendly frequency, much moreso than
regular "RF" operation.  My knee-jerk solution to this would be some
kind of antenna, run into a pretty hefty amplifier, then switch the
signal from the amplifier at 60 kHz.  A very simple low-pass filter
on the output to remove harmonics, and you'll likely find a pretty
good data signal.  At that data rate, you can decode it by hand
with a blinking LED. ;-)

I haven't done it yet, though; too busy on other things.  What's the
availability of that signal?  Am I likely to pick it up in my RF-proof
dungeon of a workshop?  I can't even get much local radio down
there...

Mike H.

2005\06\14@183948 by olin piclist

face picon face
Peter Onion wrote:
> You really want a hi-q (= narrow bandwidth) filter afterall the data
> rate is only 1 bit/sec.
>
>> Then maybe use a cap to even out the signal, and
>> transistor/opamp to amplify, maybe not in that order? Then just bring
>> that into an analog input? Am I on the right track?
>
> You can do it that way.  The first one I ever built used op-amps all the
> way.  But I did live within 50km of Rugby !

I've thought about building a WWVB receiver as a fun project.  However,
since it's just a "fun" project, it's not likely to get very far anytime
soon.  The thought was to use a dsPIC to do the RF demodulation and also
present a NTP server over ethernet.  The whole thing could be easily powered
over ethernet or from a wall wart.

The part that intrigues me is doing the carrier detection and demodulation
in firmware.  A dsPIC can sample at up to 500KHz, although even half that
should be enough to process the 60KHz carrier.  Start with a magnetic pickup
antenna such as is found in AM radios and make a resonant circuit of it at
60KHz.  After that, the carrier can be treated pretty much like an audio
signal.  Either select reasonable gain by trial and error depending on where
you live, or use a selectable gain amplifier and have the PIC switch as
necessary.

Anyway, the interesting part is detecting just the right frequency in
software. I was thinking of multiplying the incoming signal by both sine and
cosine at 60Hz derived from the PIC oscillator crystal.  If the PIC
oscillator is good to within 50ppm (within cheap crystal specs), then it
will be off by no more than 3Hz.  If the carrier were constant and the PIC
60KHz oscillator phase locked to the carrier, then each series of sine
products and series of cosine products would be constant.  Due to them not
being in phase, each series will actually be a sine wave at the beat
frequency, which is up to 3Hz.  Now low pass filter each stream to remove
most components beyond around 5Hz.  This means the 3Hz beat frequency
survives, but not much beyond that.  For each sample point, the resulting
filtered sine and cosine products are squared, then added together.  That
yields the square of the carrier strength with a total bandwidth of about
10Hz.  I theory you would take the square root of that for true AM
demodulation, but that shouldn't be necessary if just looking for high and
low levels.

To be a little more fancy, you could have a separate slow speed servo
algorithm try to null out the beat frequency.  Because of phase noise due to
multipath distortion of various atmosphere layers, I don't think it
reasonable to try a true phase locked loop.  I think it would jump around
too much.

Anyway, I've never heard of this type of AM demodulation before, but I don't
see any theoretical problems with it.  Maybe some day I'll actually get
around to trying it.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\06\15@014741 by Buehler, Martin

picon face
i'm not sure if the american radio clock signal is the same as the
european (dcf-77) is.
i'm actually working on a project, where the dcf-77 signal is used to
sync the rtc. there are cheap single chip receiver modules (receiver
chip + antenna) for the dcf-77 signal. so i'm using one of these. they
give you the demodulated signal as ttl, and all you have to do is
capture and decode the pulses received.
tino

************************************************************************
******************************


>{Original Message removed}

2005\06\15@045805 by Alan B. Pearce

face picon face
>> You really want a hi-q (= narrow bandwidth) filter afterall the data
>> rate is only 1 bit/sec.
>>
>>> Then maybe use a cap to even out the signal, and
>>> transistor/opamp to amplify, maybe not in that order? Then just bring
>>> that into an analog input? Am I on the right track?

All the commercial receivers use chips designed for the purpose, and most
seem to have one or two crystals as a Hi-Q filter on the carrier frequency.
A lot of information on the signal formats is available from
http://www.hkw-elektronik.de/produkte/ and the data sheets contain info on
the different modulation formats. The Temec (Telefunken) chips do not seem
to be available anymore.

>I've thought about building a WWVB receiver as a fun project.  However,
>since it's just a "fun" project, it's not likely to get very far anytime
>soon.  The thought was to use a dsPIC to do the RF demodulation and also
>present a NTP server over ethernet.  The whole thing could be easily
powered
>over ethernet or from a wall wart.
>
>The part that intrigues me is doing the carrier detection and demodulation
>in firmware.  A dsPIC can sample at up to 500KHz, although even half that
>should be enough to process the 60KHz carrier.  Start with a magnetic
pickup
>antenna such as is found in AM radios and make a resonant circuit of it at
>60KHz.  After that, the carrier can be treated pretty much like an audio
>signal.  Either select reasonable gain by trial and error depending on
where
>you live, or use a selectable gain amplifier and have the PIC switch as
>necessary.

This reminds me of the discussion I originated about 1-2 months ago here
about doing exactly that, except I was looking at doing it as a software PLL
instead of an FFT. It was at that time I found out how many modulation
formats there are, as well as modulation depth percentages (some switch the
carrier right off, others go to levels that vary from 10% to 25% of full
power). Again refer to the data sheets on the site above for info on aerials
and modulation formats. The FFT route would probably be the way to go in
hindsight. At that data rate it could probably even be done in an 18F,
although probably easier in a dsPic.

2005\06\15@050843 by Peter Onion

flavicon
face
On Tue, 2005-06-14 at 17:25 -0500, Mike Hord wrote:

> I haven't done it yet, though; too busy on other things.  What's the
> availability of that signal?  Am I likely to pick it up in my RF-proof
> dungeon of a workshop?  I can't even get much local radio down
> there...

There are some coverage maps for  WWVB (I'm assuming you are in USA) at
http://tf.nist.gov/timefreq/stations/wwvbcoverage.htm

A Google for WWVB or MSF or DCF77 will get you lots of useful links.

> Mike H.
>

2005\06\15@052450 by Peter Onion

flavicon
face
On Tue, 2005-06-14 at 18:39 -0400, Olin Lathrop wrote:
> Peter Onion wrote:
> > You really want a hi-q (= narrow bandwidth) filter afterall the data
> > rate is only 1 bit/sec.

> > You can do it that way.  The first one I ever built used op-amps all the
> > way.  But I did live within 50km of Rugby !

Just to explain that..  Most op amps have a high enough gain-bandwidth
product to be able to get useful gain at 60kHz, so a ferrite rod antenna
feeding a few stages of gain, a bit of filtering and then into a
envelope detector will produce a reasonable data signal.  

>
> I've thought about building a WWVB receiver as a fun project.  However,
> since it's just a "fun" project, it's not likely to get very far anytime
> soon.  The thought was to use a dsPIC to do the RF demodulation and also
> present a NTP server over ethernet.  The whole thing could be easily powered
> over ethernet or from a wall wart.

Yup... My MSF current receiver/decoder will probably end up on the end
of the wi-fi development board I'll be playing with soon.

[BIG SNIP]

Your "dsp theory" sounds fine Olin, but it really is a bit OTT for this
application ;-)   However as a learning exercise in using these
techniques it sounds great fun.  I always enjoy getting to the point
where you can see the 1 bit/sec data.

>
> The part that intrigues me is doing the carrier detection and demodulation
> in firmware.

What you described was essentially a "Direct Conversion" receiver where
the incoming signal is converted directly to base band without any
intermediate frequency (IF) stages.  

> Anyway, I've never heard of this type of AM demodulation before, but I don't
> see any theoretical problems with it.  Maybe some day I'll actually get
> around to trying it.

It's only a variation of good old fashioned morse code on/off keying
(class A1A).  As someone pointed out you can decode the data by eye with
a blinking LED or if you have a SSB radio receiver that will tune that
low you can do it by ear.  I've used my Yeasu FRG100 and fed the output
into a PC sound card to decode MSF.

Peter.

2005\06\17@131312 by Howard Winter

face
flavicon
picon face
Martin,

On Wed, 15 Jun 2005 07:47:41 +0200, Buehler, Martin wrote:

> i'm not sure if the american radio clock signal is the same as the european (dcf-77) is.

I'm sure it isn't!  :-)  The American WWVB signal uses Amplitude Modulation (the signal is emitted constantly, but its amplitude varies) whereas the British MSF signal uses switched carrier (the carrier is turned completely on/off).  I believe the German DCF-77 signal is similar to MSF, but on a different frequency (MSF and WWVB are 60kHz, DCF-77 could be 77kHz perhaps?).

Incidentally, where I live in South-East England, not far from London, I can receive MSF *and* DCF - I bought a DCF clock for GB£4 (US$7.30,  5.70) in a sale recently, and I just had to move the hour-hand back an hour and it works fine!  :-)  It will be a useful for an experiment to see if I can use Jinx' idea of picking off the second-hand solenoid pulses to use as a 1Hz timekeeping signal.  
This clock has a pair or solenoids, for the second and minute hands, with the hour hand geared to the minute hand.  The minute hand moves a quarter of a minute-mark each 15 seconds.  Previous clocks I've had use a single solenoid and all three hands were geared together.

Cheers,


Howard Winter
St.Albans, England

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...