View a Leap Second - WWVB - June 30, 1997

Below is a WWVB subcode record of the leap second in the middle of 1997. The question I was trying to answer is what subcode bit is transmitted during the leap second; a '0' bit or a '1' bit, or a position marker bit. The answer is a position marker bit. Since the positive leap second is inserted into UTC at the top of the minute, this is the only time three position markers are broadcast back-to-back.

From: Tom Van Baak <tvb>
Subject: Re: WWVB to Logic RF Circuit / Chip
Date: 1997/10/21
Message-ID: <>#1/1
References: <>
Organization: VERITAS Software
Reply-To: tvb

Brooke wrote:
> I would like to receive the time code from WWVB on 60 kHz (not the
> carrier frequency).  Does anyone know of a simple circuit and/or
> chip that outputs the time code?
> Thanks,
> Brooke


An excellent 3 part article by Bob Roehrig on building a
WWVB receiver can be found in the Jan/Feb/Mar issue of '73

The old HP 117A VLF receiver has a signal level output that
can be used to decode the subcode. The manual has an example
where a strip chart recorder is used to decode the signal.

for a description of the 60 second WWVB time code format.

for a bibliography of time and frequency radio receivers.

Perhaps the easiest way these days is to extract the raw
subcode from one of the many battery-powered desktop WWVB
radio clocks on the market.

I have been unable to decode the VLF receiver output of
the Oregon Scientific, Brookstone, or Junghans WWVB radio
clocks, but have had great success with the Arcron Zeit.

There are two ICs inside the desktop model of the Zeit:
one is the 60 kHz VLF receiver and the other, hidden
under the LCD, is the uP which decodes the 1 Hz subcode
and drives the display. One of the traces that runs
between the two ICs is the receiver enable and the other
is the 1 Hz subcode output. Logic low (0 V) on this line
corresponds to full WWVB power and logic high (3 V) to
10 dB reduced power.

When the receiver is on and reception is good the 0.2,
0.5, and 0.8 second width bit frames are clearly visible
on the 1 Hz subcode output. This output can be sampled
and decoded by a PC or microcontroller (such as a Basic

At the end of June 1997 I used this technique to observe
how WWVB handled the leap second. In the following log
the date and time are PC local time (PST), "P" indicates
a frame or position marker (0.8 second of 10 dB power
reduction), "1" indicates a binary one (0.5 second power
reduction), and "0" indicates a binary zero (0.2 second
power reduction):


Use SP432 to decode the :59 and :00 minute frames:

 59 minutes, 23 hours, 181 days, UT1 = -0.5 seconds, year 97,
 leap second pending, and DST active.

 00 minutes, 00 hours, 182 days, UT1 = +0.5 seconds, year 97,
 no leap second pending, and DST active.


The above was an old posting of mine found in the Google Usenet archives.

Visit home page
Send questions or comments to tvb