|
Ok Wolf,
you're right. That
is a simpler solution, requiring only minimal hardware changes for Stefan. One
channel sends a continuous uninterrupted 17.94 kHz tone to the divider, the
other channel conveys modulated Opera at an arbitrary frequency to
a VOX rectifier.
Wolf, I'm looking forward to learn more details
about your modulator and MF PA setup when it's ready!
Best 73,
Markus
Sent: Sunday, December 29, 2013 12:18 PM
Subject: Re: LF: VLF exciter
Hello Markus,
Ok .. i had assumed the problem
was the flipflop (or binary frequency divider) which would be in an
unpredictable state when not fed with a phase-coherent signal "in the gaps"
(zeroes in the Opera- or whatever amplitude keyed signal).
This could be
solved by providing the auxiliary output from the soundcard (at 2 * f_tx), to
keep the divider chain running.
Depending on the MOSFET gate drivers, it may
also be possible to feed the 'keying signal' to an output-enable-signal, but I'm
not sure if that would put the driver output in high-impedance state (which
would result in a catastrophic 'bang' if both FET gates were left floating).
Anyway your suggestion with the xor gate may be the simpler variant than
using two different output signals from the soundcard.
73 and
HNY,
Wolf .
(p.s. just managed to build a transformer for a
low-voltage MF PA, which transforms 0.6 Ohms drain load impedance to 12.5 Ohm
with very low stray inductance.. using copper foil instead of wire, and thin
adhesive tape ("Tesafilm") as insulator. But that's another
story..)
Am 29.12.2013 12:00, schrieb Markus Vester:
Hi Wolf,
I think Stefan was only trying to solve the
simpler problem of generating a dual squarewave drive for his VLF PA, and was
not aiming for amplitude modulated output yet. So when he transmits Opera
we'll just have to live with ultra-strong keyclicks, devastating the 9 kHz
band all over Europe ;-))
But I remember that in 2011 we discussed about
transmitting high-quality Hell piccies. As you say, the Kahn
technique with a separate AM modulator is certainly a good and efficient
option. At VLF, outphasing (aka Chireix or LINC technique) between two
half-bridges would be a valid alternative. There's some software for the
purpose (including Basic source code ;-) on
This works in conjunction with your SndOutpt
audio utility, generating the two outphasing carriers at the left and right
stereo output.
Best 73,
Markus (DF6NM)
Sent: Sunday, December 29, 2013 9:12 AM
Subject: Re: LF: VLF exciter
My suggestion:
Use two outputs from the
stereo soundcard:
An "amplitude" and a "phase" channel. The output from
the phase channel will be a square wave, coherent, constant amplitude, fed to
the divider chain. The "amplitude" channel delivers, well, the amplitude (as a
waveform with variable voltage) which, after rectification, drives the final
voltage via an additional mosfet. I think the software option is already
there, and iirc it has been used (by Jim ?) for an LF PA.
This way, you
could even use amplitude shaping if the bandwidth (from the hard-keyed PA) was
a problem - which I don't think it is, on VLF.
Cheers,
Wolf
(p.s. problem with the "very long command lines" already fixed,
new SL version will be uploaded soon).
Am 29.12.2013 02:43, schrieb
Stefan Schäfer:
Am 28.12.2013 23:33, schrieb Markus Vester:
[...]
PS Any progress on bypassing the divider in
your VLF transmit chain?
Aaah, no, i forgot that. So i cannot transmit OP32 on VLF
because it would be incoherent then. So i have to key the PA via a VOX and
OP32 tones, still :-/
In this H bridge PA config, an exact 50% duty
cycle isn't needed, so i may actually remove the divider. But then i need a
solid and defined input signal wave form! How do i get an inverted signal
then? A transformer isn't possible since i like to have a wide frequency
range of the PA. So far it works from 10 Hz to > 100 kHz. With the
frequency divider, the wave form can be quite distorted, no problem unless
the peak reaches the necessary voltage level. So maybe it is better to build
the VOX circuit instead? Also i want to leave the PA circuit as it is, for
nostalgic reasons :-)
Or do you have another idea? I remember that
circuit with the XOR gates...
73,
Stefan/DK7FC
|
|
