LF: WOLF Test results from DF0WD

[DL4YHF@aol.com]

Hello Jim, Markus, and group,

Some on-air testing of Stewart's implementation of WOLF took place on sunday 
evening, which was partly successful (tnx Markus, DF6NM and Jim, M0BMU for 
their patience and to Stewart, KK7KA for the new software).

Here is the new setup used at DF0WD:
- An LF-transverter which was once intended to be a linear transverter
   (until I blew the first set of matched MOSFETS a few months ago). 
   Transverter mixes 10.136MHz down to 136kHz and vice versa.
- An oven-controlled 10MHz XO which feeds the transverter 
    and synchronizes the 30MHz 'master' oscillator inside the IC706
- An old IC706 shortwave TRX, slightly modified 
  (30MHz clock 'voltage controlled'). 
- WOLF signal taken from soundcard, fed to IC706 in USB mode 
  (to test amplitude shaping of new WOLF software by KK7KA).
- TX output power about 10 Watts (a little more later with "rough keying")
- Antenna 200m horizontal wire (no straight line), end fed, 
   about 8..10 meters above ground. 
   EIRP once estimated by PA0SE about 80mW with max. drive, 
   depending on WX (see www.qsl.net/dl4yhf/lf_index.html).

I received the WOLF signal with the IC706 in "CW" at (10MHz+)137.500, using a 
the narrow CW filter.
To transmit, the IC706 runs in "USB" at (10MHz+)135.500. The audio frequency 
should have been 2000.000 Hz, so the transmitted WOLF carrier should have 
been on 137.500 kHz (but it wasn't, see below). The soundcard's A/D 
conversion rate was once measured as 7938.086, this value was passed to WOLF 
as command line argument (because this is quite 'far off', I will verify it 
later with the 10MHz-ref divided down to audio freq. Thanks Markus for the 
point).

To Jim, M0BMU: The "test carrier" was transmitted "CW", so it was not 
generated the same way as the WOLF signal. This may explain the problem.. I 
don't know how you generated the "leading carrier" before your WOLF 
transmissions a few weeks ago.

One intention for the test was to compare the new 'amplitude shaped' WOLF 
signal. I found that with the amplifier currently use here, I can not 
generate such a 'crystal clean' WOLF spectrum like the one shown on Stewart's 
updated WOLF page (http://www.scgroup.com/ham/wolf.html).

It seems that each MOSFET (IRF540) in the PA needs an indivudual carefully 
adjusted bias point, and a little more source resistance (as negative 
feedback) to make the small PA work linear. Each FET carries a different 
current, especially when the driving power is very low. Result: the damn 
thing looks like a linear when the power is 10%...70% of the output, but it 
does not behave linear when drive is less than 10% (because of different FET 
behaviour) or more than 70% (compression point).  The low-power linearity ot 
the PA was better, when I used the 4 "best-matched" FETS which I selected 
from a bundle of 50 low-cost FETS.  I blew two of these 4 FETS a few months 
ago when was experimenting with a slowed-down variant of PSK31 (which 
produced the sample modulation spectrum as WOLF with the option 't 1').

The best way to drive a high-power BPSK transmitter is the way M0BMU 
described (do the phase reversal by "XOR-gating", and do the envelope shaping 
by "modulating the PA supply voltage" or similar).  On the other hand, a 
'completely linear' concept will allow low-power experiments with QPSK, MT 
HELL etc. As long as I am still resctricted to the crazy power limit in DL, I 
don't have to care for a highly efficient TX design (there's enough "spare 
power" to compensate the losses in the ATU and in the cable which I consider 
as part of my transmitter hi).


Hope I'll have better luck next time, and thanks to everyone involved. 
Also thanks to all stations active in fast CW, which will remain my favorite 
mode. Nice to meet you again; no matter if in CW, any kind of BPSK or 
whatever.

73's Wolf (DL4YHF).