[Moon-net] JT2 and JT4 modes

[Leif Asbrink]

"Alberto E. Zagni" wrote:

> Unfortunately the requested hw+sw for MAP65 is quite demanding.
> JT2 and JT4 should offer a similar approach with NO added cost or 
> complexity. Please let's go this way.

Hi Alberto,

I am afraid you overlooked the dynamic range problems.
EME-ers tend to form clusters with several stations not
too far away from each other. In the SM5 area, when I was
active in CW, there was half a dozen stations that
were too strong to be allowed within the IF passband of
a conventional radio while receiving signals at the noise
floor. Having all the EME activity within 2 kHz will not 
work for a large number of stations and I do not think it 
will be any good idea at all to try to move in that direction.

I do however think that a lower bandwidth is a very good 
thing. The number of stations is growing and even though
JT65 can detect several overlapping stations when they
have similar signal strength I do not think overlapping
stations that differ in amplitude by 30 dB will allow 
detection of the weaker station when it is near the detect 
limit. (The stronger would be terrestrial from another
EMEer who is not local, but has a reasonable signal.)
In such cases JT2 will solve a problem that is likely to
arise in the future with JT65.

It is however not quite true that "the requested hw+sw 
for MAP65 is quite demanding." You can use a softrock
or time-machine together with the builtin soundcard
of a standard computer at negligible cost. This will 
give a single channel only and a limited dynamic range, 
but I am pretty sure the dynamic range will be better 
as compared to what you find at the loudspeaker output 
of many popular transceivers. In case you want good 
dynamic range you can use an SDR-IQ for USD 499. It
will also give you one channel only so you will not be
able to take full advantage of X-pol antennas.
(I do not know whether MAP65 currently accepts data
at the rates that SDR-IQ can supply, but I am pretty
sure it will in the future.)

I have not operated MAP65 myself, but it seems to me
that saying "sw for MAP65 is quite demanding" is not
fair. You would have to use Linrad to feed MAP65 with 
data, but that is extremely easy, there is an executable 
for Windows, linrad.exe, and you do not have to know
much about Linrad to run it just to send data to MAP65.

BUT, most important of all, it will not be a good idea to
take decisions about future operational standards based on
equipment on the market today. Things change rapidly.
Price and complexity of various hardware solutions will
change dramatically in a not too distant future. The 
Perseus will give nearly 500 kHz bandwidth and an extremely
good dynamic range at 799 Euro. At such bandwidths it
will be possible to remove static rain and it will be 
straightforward to make MAP65 detect all JT65, JT2 and 
JT4 transmissions on any amateur band. What should become
available next is multi-channel digital systems. They will
open new possibillities for interference suppression on HF
bands, but also for EME-ers. No details here, but believe me,
the effects will become dramatic for some of those who 
suffer from man-made noise. Spurs, broadband over powerline,
noise from computers and screens etc. Anything.

Personally I favour several well defined building blocks over
a "everything in one box" solution and my digital dream is
as follows:

1) A PCB with an A/D converter (LTC2208 or similar with randomization
at both input and output.) To this PCB, one glass fibre with a
clock that is at 32 times the A/D sampling frequency. The PCB 
would have a divide by 32 to supply the A/D with its clock
and a parallel to serial converter that would be driven by 32 times the
A/D sampling rate to send the digital data onto another glass fibre.
Galvanically isolated, and with no other frequency than the A/D clock
and its overtones, this PCB will be free totally of spurs.

2) A box that can generate the clock signal for some (large) number
of glass fibres. Possibillity for GPS locking.

3) A PCB to put into the PC that can receive the serial data from
a number of type '1' boards. This board would be "standard digital 
technology" with FPGA(s) that do the decimation of N channels and
supply I and Q to the PB data bus. An architecture that would allow
several such PCBs in one computer running synchronously would be
very desireable.

I think all three units are easy to design and I am sure they would
allow a compromise-free performance. Personally I do not have the skils
needed for the PC unit, but I do know it would be standard technology
that a large number of people know how to handle.

Personally I am convinced that digital technology will have a much
greater impact on HF bands than on VHF (EME). It will not become
visible in quite some time yet because HF communication is so much
easier than EME so the driving force is not so large.... 

It is obvious to EME operators that digital technology has lead 
to big changes. When the advantages become noticeable to HF 
operators, the cost for wideband hardware will drop due to 
increasing volumes. EMEers will have access to good 2-channel 
wideband hardware at modest cost in the future:-) I think 
narrowband transmissions (JT2) that are scattered over a wide 
frequency range will be the optimum for the future. For many 
bands and modes! Trying to squeeze all activity into a narrow
frequency range would be a dead end.


73

Leif / SM5BSZ