DGPS correction codes

[The Stickman]

HI all, thanx for reading this. I'm ina search for info regarding DGPS
correction codes. I'm aware that there is a number of low frequency
NDB's that do provide correction information along with their ident.
I was wondering if anyone knows what data format is used to
encode this info and ways to extract it from the NDB signal.
ANy info would be appreciated.
Thanx.
--

("`-/")_.-'"``-._ Alex Wiecek - Email : mag...@ionsys.com
. . `; -._ )-;-,_`) Webpage : www.ionsys.com/~magnum
(v_,)' _ )`-.\ ``-' Packet Radio: VE3GOP @ VA3BBS
_.- _..-_/ / ((.'
((,.-' ((,/

[Sebastian Birnbach]

The Stickman <mag...@ionsys.com> wrote:

> HI all, thanx for reading this. I'm ina search for info regarding DGPS >
correction codes. I'm aware that there is a number of low frequency >
NDB's that do provide correction information along with their ident. > I
was wondering if anyone knows what data format is used to > encode this
info and ways to extract it from the NDB signal. > ANy info would be

appreciated. > Thanx. The data stream is encoded in RTCM 104. You can
get the specs for it (for about 70 bucks) from the Radio Technical
Commission for Maritime Services 655 fifteenth Street, NW, Suite 300
Washington D.C., 20005 U.S.A.

RTCM 104 consists of 21 different messages, the first 17 of which are
designed for use with C/A code tracking receivers, and result in 10m
accuracy. The other 4 messages have to do with carrier phase
corrections, which are used in surveying. You can get sub-meter accuracy
with those, but need a specially equipped field-GPS as well.

RTCM 104 is transmitted via AM radio beacons along the coastline of the
USA, Canada, and some of the Northern European Countries. It is
transmitted via a 300 kHz carrier. The following is a repost from: "Tom
Clark (W3IWI)" <cl...@tomcat.gsfc.nasa.gov>:

***repostal starts***

Michel Stoop wrote: > > Mark Fraser asked me this: > How to demodulate
the beacon MSK. > > Wel there are 2 ways to do that. > > You can use a
SSB receiver and use an AFSK demodulator to recover the bitstream. > or
use a FSK detector connected to the IF signal from a communications
receiver. > For this you can use a Phase Detector or a Costas-loop PLL >
> I know nothing of the data format, and what must go to the garmin unit. >
Presumably they transmit Satellite number and x,y,z offset values. >
Sinse they are using 200bps I think that the datastream is synchronous >
and packed in frames. And that the data is send in async (rs232) frames
to the > garmin unit. > > Since the datarate is 200bd the shift is 50Hz,
very narrow! > This is why the garmin can retune the beacon receiver.
Maybe there is also data > transmitted for tuning help. (ie transmitter
drift value) > But I rember seeing 75Hz somewhere. But then the datarate
would be 300bd. > > Michel

The ~300 kHz beacons transmit MSK encoded binary data. The basic data
formatting is similar to that used with the GPS 50 b/s downlink, with 24
bit "words" and 6 bits of error checking, with the data highly
compressed. See the GPS ICD-200 for the basic formatting spec, and then
read the RTCM SC-104 document for details. The data bit rates used at
VLF are 50, 100 and 200 b/s, i.e. 1,2 & 4 times the GPS satellite
downlink rates. At 100 b/s, "fresh" DGPS info arrives about every 8
seconds. The update rate for the commercial services using FM broadcasts
(i.e. DCS and AcquPoint) is faster than is sent by the Coast Guard.

The typical DGPS beacon receiver (this is also true of the FM broadcast
RDS units) convert this ugly binary data into "legal" ascii characters,
sending 6 data bits per 8-bit character, mapped into the
<space>...A....Z.....a.....z character space which is what your typical
DGPS receiver wants to see (usually at 4880 b/s rate). [When I have my
W3IWI-13 packet DGPS becons on the air sending the bits so that they can
be copied with VHF amateur equipment, the DGPS data looks like pure
heiroglyphics, understood only by a DGPS-equipped GPS receiver. But the
characters are "user friendly" in that they don't cause the CRT of a
causual user sharing the frequency to flash, or send "beeps".]

The RTCM spec defines a number of DGPS message types sent at different
rates. The one with most of the info has the pseudorange and
pseudorange-rate error for each PRN in view at the differential station.
Also sent are an ID number for the station (0000-1023 in binary), the
specific "serial number" (called IODE=Issue of Data Ephermeris) sent by
the GPS satellite to identify the ephemeris version in use, and the
typical GPS receiver will reject any info for which te IODE's don't
match (an IODE usually is applicable for ~6 hours time, depending on
when the USAF controllers uplink new data. So it is NOT unusual to have
IODE mismatch).

All in all, the detailed format is pretty complicated and the "Cap'n
Midnite Magic Decoder Ring" code to generate or disentangle the DGPS
messages runs several pages of bit-diddling in "C". If there is REALLY a
need for some code to generate of decode the bits, and if someone are
successfully decoding the MSK bits from the beacons, I can probably put
you in touch with a person (not me) who has developed suitable software.

IMHO, kludging a DGPS beacon receiver that works will be a ~6 month
intensive effort -- not just something you can do with a TNC!

73, Tom

***repostal ends***

-- Please answer to one of the following addresses:
birn...@informatik.tu-muenchen.de sebastian...@rberger.de