DARC
Berg. Universitšt Wuppertal
Arbeitsgemeinschaft Amateurfunkfernsehen e.V.

Digital  Amateur  TeleVision



 

Second part of a series in the magazine TV-AMATEUR of AGAF (German ATV Club)
written by Prof. Uwe Kraus, DJ8DW, AGAF-Member #50


Digital HAM TV
---------------
(Uwe E. Kraus, DJ8DW)

During HAM RADIO 2000 fair in Friedrichshafen, Germany, we had a live
Digital ATV demonstration. This text describes present and future
developments of the DATV working group.

All day long we transmitted a camera view in hall 9 at the
DARC-Distrikts stand with 10 mW into an eleven element yagi antenna on
434 MHz with 2 MHz hf bandwidth. DF2DS, DJ3DY and DC5QC supervised the
transmitter devices and answered on questions from many interested
visitors. The receiving devices were installed in hall 9 too about 20 m
away at the AGAF stand. There DJ8VR, DJ1CU, DL4KCK, DC6MR and DJ8DW
cared for equipment and visitors. Two colour TV monitors showed the
scene at the camera side in realtime, and many german as well as foreign
radio amateurs payed much interest. The highlight of the first day was a
visit by the DARC administrative committee accompanied by prominent
guests.

present DATV technology

The second generation transmitter and receiver devices are shown in the
block scheme. A video camcorder supplies a PAL signal to the
PAL/MPEG converter where a MPEG-2 transport stream with 2 Mbit/s is
produced including FEC (error handling). The effective usable data rate
is variing around 1,2 Mbit/s, the rest is FEC and adaptively added
stuffing packets. This is unsufficient but the used single-chip MPEG
encoder does not allow a better solution without huge additional effort.
The chip produces only a variable data rate elementary stream without
full MPEG-2 standard elements (no B-frames). The next development steps
will add better video quality maintaining the overall data rate.

The GMSK coder comprises of a digital part supplying the I- and
Q-signals depending on the input bitstream and a following analog I/Q
modulator that produces the GMSK signal at 36,2 MHz. This IF was chosen
in order to use commercially available SAW filters with 2 MHz bandwidth
deriving from the Digital Audio Broadcasting (DAB) development. For
steeper filtering edges and better selectivity two filters with buffers
are cascaded. A step-up-mixer with a SBL1X converts the IF signal to 434
MHz with the aid of a 470,2 MHz oscillator. Next is a helical-filter
bandpass PA producing 10 mW on 50 Ohm, and a 30 dB power modul with 10 W
output is useful for middle range tests or driving bigger PAs.

The receiver uses lambda/4 tubular tuned circuits at the input and
between preamplifier and mixer (SBL1X). As IF stages two
limiter/FM-demod ICs are used with SAW filters in front and between
them. From the demodulated signal a slicer produces the serial data
stream and a 16 MHz crystal PLL with divider the appertaining 2 MHz
clock signal. For error handling, MPEG decoding and PAL signal
processing a commercial settop box with digital satellite TV standard
(DVB-S) is used. Therefore the received serial MPEG-2 transport stream
is remodulated on a 1152 MHz carrier as narrow band QPSK modulation
(SCPC). The settop box must be able to process a symbol rate of 1
Megasymbol/s where some devices have difficulties because of their
internal oscillator`s phase noise.
Other boxes only allow 2 Megasymbols/s minimum.

Transmitter and receiver are built on modular europe-format printed
circuits. Digital signal processing is done mostly in programmable logic
ICs.

Third generation DATV

The next generation is formed clearly at least on the transmitter side.
Here the modular concept is left in favour of a compact, cheaper and
homebuilder friendlier construction. The aim is one europe-format
printed circuit processing an MPEG-2 transport stream and producing 10
mW hf on 434 MHz. Discussions at the HAM RADIO fair especially with
DJ3OI resulted in the plan for a built-in test pattern generator for
long time test set-ups. After a first analysis it should be possible to
integrate a programmable permanent storage device containing moving
MPEG-2 sequences 10 seconds long which are repeated cyclicly.

The compact and extended construction is reached by highly integrated
programmable ICs, a digital I/Q modulator and digital filtering of the
GMSK spectrum making SAW filters and drivers obsolete. The PAL/MPEG
transcoder will be constructed with an encoder IC that includes sound
coding, B-frame processing, MPEG-1 and MPEG-2 capability and complete
transport stream processing.

The third generation receiver will maintain it`s modular design
temporary. The converter circuits are prolonged, and the mixer
oscillator is crystal controlled with frequency multiplier (no
synthesizer). IF, slicer and clock regenerator remain unchanged, between
IF and Slicer a channel equalization is planned in addition.

The current solution of MPEG decoding by settop box and QPSK remodulator
is driven by time pressure and seems to be problematic in a future view.
Maybe the boxes are getting cheaper, but the cheaper ones are not able
to handle the low data rate of 1 MS/s. A self made improvement of the
boxes internal is impossible due to lack of proper documentation. One
possibility could be enlarging the received transport stream with
stuffing packets producing a wider and box friendlier QPSK. This
requires complicated error decoding before and standard error handling
after the procedure.

Maybe it is better to use a self built MPEG-decoder in the long term
with a newer MPEG decoder IC that processes the transport stream and
delivers RGB or even PAL/NTSC. Such decoder IC have more onboard storage
capabilities reducing the overall construction expense. The error
handling could be simplified because of unneeded DVB standards
compatibility.

Higher bands and data rates

The higher amateur radio bands allow transmissions with higher data
rates giving a better video quality. An effective data rate of 5 or 6
Mbit/s provides the excellent quality known by digital satellite and
cable broadcasting. An overall data rate of 7 Mbit/s with GMSK means hf
bandwidth of about 8 MHz. Following the technological development FM-ATV
could be replaced by the digital transmission concerning a transition
period with both modes. DATV would allow more parallel channels and
duplex traffic at the same ATV allocations in the end.

Devices of the second and third generation DATV are able to use higher
data rates. The GMSK coder clock frequency can be higher, filters in
step-up mixers and amplifiers and in the converter are wide enough, only
the SAW filters must be exchanged against 8 MHz wide filters with 36 MHz
center frequency. The wider GMSK signal on 70 cm is converted to the
higher bands, the 70 cm receiver can be used as broadband receiving end.

Another modulation scheme could be QPSK, on 23 cm this would enable
direct reception with a settop box provided that a DVB standard signal
is transmitted. Furthermore QAM (from DVB-C cable networks) and
especially OFDM from DVB-T (terrestrial digital TV) are possible, the
latter reducing problems with hf multipath reflections. These modes
allow data rates of several MBit/s in 2 MHz hf bandwidth but require
very linear amplifiers and higher signal to noise ratios at the receiver
frontend.

Low usage of amateur frequencies will produce desire otherwise. For
modern digital broadband modes like DATV suitable segments in all
amateur radio bands above 430 MHz are important. This was discussed in
Friedrichshafen too using a diagram with the presently allocated ATV
segments from 70 to 3 cm. It would be useful to reach a european
agreement on DATV frequencies on each band. OM who are advocating our
interests at frequency conferences would be in a better position.
Happily more and more knowledge is found that digital broadband modes
will be essential working fields in a future amateur radio world. These
are very suitable to get young people into modern communication and
multimedia technology and direct their interest to appropriate
engineering careers.

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

Starting and Testing

First considerations and experimental tests on Digital ATV began about
five years ago. Most important was the appropriate modulation for the
starting period, and GMSK (Gaussian Minimum Shift Keying) was chosen.
GMSK is successfully used with mobile phones (GSM), it has a compact
spectral density and a constant amplitude even after band-pass
filtering. Similar to FM an effective power output is possible with
class-c amplifiers.

Demodulation in the receiver is simple with an FM demodulator, a costly
coherent demodulation would give 3 dB more. A GMSK disadvantage is the
relative low data rate of 1 bit/s per Hertz bandwidth. Until now
experiments took place only on 70 cm (433-435 MHz), this band is
interesting for propagation tests (long distances without repeaters).
Antennas give much gain here with moderate dimensions, cable losses are
relative low and middle range power output is produced easily. 2 MHz
bandwidth for DATV on 70 cm gives sufficient space for other modes, each
of them has equal rights and is used by amateurs according to their
personal liking and possibilities. AM-ATV on 70 cm covering most of the
band has no future any more.

For data compression MPEG1 or MPEG2 was chosen because of worldwide
standardisation and in view of cheap ICs for coding and decoding from
consumer electronics. So the first generation DATV was developed setting
an emphasis on evaluations of the digital modulation and of propagation
effects. MPEG1 data files were sent from a Video-CD through a specially
developed PC slot card to the digital modulator and stored at the
receiving end through a similar slot card on the PC hard disk.
An MPEG1 software decoder displayed the video on the PC monitor. These
devices were shown and explained in a lecture at the HAM RADIO 1999.
First successful transmission tests were performed in the nearfield
aerea, over 50 km and at last over 100 km (with 80 W and stacked 19
element yagi antennas). The development got speeded by support coming
from DARC, AGAF, DARC districts G, L, O and R, some private sponsors and
by foundation of the DATV working group in december 1998. Four DATV
stations should be built within a year and used for tests by the
district teams, and they were distributed in time in december 1999.
Hermann, DF2DS, reported results of the field trials in district O in
his lecture at the HAM RADIO 2000 fair.

References:
(1) Projektgruppen DATV gegründet: Uwe Kraus, DJ8DW, CQ DL 2/99, S. 92
(2) Modulares DATV-Konzept läuft: Uwe Kraus, DJ8DW, CQ DL 11/99, S. 904

(from TV-AMATEUR 117 and CQ DL 9/2000)
translation: Klaus, DL4KCK, AGAF e.V.



E-group home: http://www.eGroups.com/list/dhtv
Free Web-based e-mail groups by www.eGroups.com