Message-ID: <28702.200103221619@gemini>
From: "James Moritz" <j.r.moritz@herts.ac.uk>
Organization: University of Hertfordshire
To: rsgb_lf_group@blacksheep.org
Date: Thu, 22 Mar 2001 16:24:40 +0000
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII; format=flowed
Content-Transfer-Encoding: 8bit
Subject: LF: <Tech>WOLF Transmit
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

Dear Mike, LF Group,

Mine is certainly not the only way of generating a BPSK signal, and 
it is certainly true from experience that little trouble results from key 
clicks on 136k, in spite of the fact that a lot of TXs on the band 
have less than ideal keying. But in my own position, being in an 
area relatively densely populated with LF stations, having a high 
ERP compared to most, and transmitting beacon signals over 
extended periods means I feel obliged to try and produce as clean 
a signal as possible.

Actually, key clicks are never likely to be a huge issue on LF; when 
compared to HF standards, even a full 1W ERP is a sub-QRP 
signal, and the higher noise level will tend to swamp clicks from all 
but the most local stations anyway.

Both "WOLF" and (usually) "Coherent" BPSK use a 10 bits/second 
rate; so the number of phase transitions per second will be similar 
to the number of on-off transitions of a 12wpm CW signal. Cutting 
the drive before a phase transition would eliminate the click due to 
the phase change, but would add 2 more, when the drive was 
switched off and then back on again, so little would be gained.

Having "redundant" envelope modulation when no phase transition 
occurs would actually increase the sideband levels - the 
explanation I gave was something of an over-simplification. It 
would still be neccessary to synchronise the amplitude modulation 
with the phase modulation somehow, so little saving in complexity 
would be acheived.

A perfectly good way of generating BPSK would be to first low-
pass filter the logic level signal, and then mix it with the carrier in a 
linear balanced mixer (rather than an ex-or gate). A linear PA would 
be required. This would be a good option if you do have a linear, 
but do not have an SSB exciter that will give output on 136k.

To put the complexity into perspective, my prototype modulator is 
built on 2 100 x 160mm prototyping boards, plus a heatsink 
assembly carrying the MOSFETs. Implementing the waveform 
generator part on a PIC - type microcontroller would roughly halve 
the number of components and interconnections. If a PCB were 
designed, the small-signal parts of the circuit would be no more 
difficult to build than, say, a fairly elaborate audio filter kit. The 
MOSFET pass element obviously needs to be quite big to handle 
the power. The Heatsink assembly could be simplified by using a 
really big MOSFET power module instead of 8 small ones, but I 
doubt if this would be cost-effective. 

cheers, Jim Moritz
73 de M0BMU