| |
PA2AGA > HDDIG 05.12.99 17:17l 178 Lines 6596 Bytes #-9661 (0) @ EU
BID : HD_99_306C
Read: GUEST
Subj: HamDigitalDigest 99/306C
Path: DB0AAB<DB0SL<DB0RGB<DB0MAK<DB0ERF<DB0SHG<DB0BRI<PI8DRS<PI8DAZ<PI8GCB<
PI8HGL<PI8VNW
Sent: 991205/1404Z @:PI8VNW.#ZH2.NLD.EU #:29468 [HvHolland] FBB7.00g $:HD_99_30
From: PA2AGA@PI8VNW.#ZH2.NLD.EU
To : HDDIG@EU
Received: from pa2aga by pi1hvh with SMTP
id AA25427 ; Sun, 05 Dec 99 13:19:24 UTC
Received: from pa2aga by pa2aga (NET/Mac 2.3.67/7.5.3) with SMTP
id AA00017049 ; Sat, 04 Dec 99 23:40:14 MET
Date: Sat, 04 Dec 99 23:34:52 MET
Message-Id: <hd_99_306C>
From: pa2aga
To: hd_broadcast@pa2aga
Subject: HamDigitalDigest 99/306C
X-BBS-Msg-Type: B
Bob
>.
------------------------------
Date: Sun, 28 Nov 1999 20:32:41 -0500
From: "Rob" <NoEmail@NoWay.com>
Subject: Decoding packet with a soundcard in Windows
Hi Greg,
I was looking for these Linux programs. All I could find were programs for
AX.25 and Linux.
I couldn't find any Linux Programs for RTTY or Pactor. (All I see were
references that they were TO BE developed)
Were can I find them?
73's
Rob
"Greg Blair" <siggraph@dgp.toronto.edu> wrote in message
news:38185412.C25AADD1@dgp.toronto.edu...
> Probably not relevant to your question, but certainly interesting:
>
> Check out LINUX. The late model kernels have support for AX25, RTTY,
> PACTOR-I and more all for free, complete with source code.
>
> Since you're homebrewing your own s/w, take a look at the kernel source
> and see how they did it.
>
> Cheers,
> Greg, ve0acr
>
> Byon Garrabrant wrote:
>
> > I'm sure that this thread has been beaten to death in the past, but I
> > searched around and could not find a Windows app to decode/encode 1200
> > or 9600 baud packet with a sound card.Can anyone confirm that this
> > doesn't exists, or point me in the direction of one?
> >
> > I have, in fact, started writing one of my one. If it turns out to
> > become functional, I'll notify the group.
> >
> > Thanks!
> >
> > Byon
>
>.
------------------------------
Date: Sat, 27 Nov 1999 08:33:38 -0500
From: Gary Coffman <ke4zv@bellsouth.net>
Subject: German packet radio
On Sat, 27 Nov 1999 01:15:07 +0100, Nico Palermo <nicopal@microtelecom.it>
wrote:
>Gary Coffman wrote:
>> Cathyrn wrote:
>> >Is MSK better also than PSK too? Or is MSK just better than
>> >all FSK with higher deviation? Is MSK just what it sounds like? FSK with
>> >that specific deviation?
>>
>> Better is a slippery word. MSK does have a smaller occupied spectrum
>> than ordinary PSK for signals of the same baud. OTOH, BPSK has a larger
>> Hamming distance than MSK, so it can offer a weak signal detection
>> advantage if coherent detection is employed. Which one is better is a
>> matter of particular spectral circumstance and system objectives.
>
>No! MSK performs even better than BPSK. Since MSK is a modulation with
>memory, you can demodulate it with a Viterbi decoder (or alternatively
>with a matched filter with 2T integration period) and get the same BER
>performance of BPSK ( BER=10^-5 @ Eb/No = 9.6dB).
I think we're talking apples and oranges here. If you code the data, you
can use a Viterbi decoder to get coding gain, but you lose effective
throughput
in doing so. Dale added that as an option to our modem, but the effective
throughput drops to 38.8 kbps when you do. I don't understand what you're
saying about using a 2T integration filter. We already use an integrate and
dump filter in our adaptive slicer.
>MSK performance is worse than BPSK only if demodulated by a coherent
>detector which takes symbol-by-symbol decisions (3 dB penalty), either
>by an incoherent one (6 dB penalty), or by a frequency discriminator. In
>the latter case the optimum modulation index is 1.25 (not 0.5) and there
>is a ~7 dB penalty over Viterbi decoding.
There is a 3 dB penalty between MSK and BPSK because the Hamming
distance is twice as large for BPSK as for MSK, ie there's only a 90 degree
bit to bit phase shift with MSK while BPSK has a 180 degree bit to bit shift.
This gives a bigger decision space for BPSK, and hence the difference in
channel performance. That assumes coherent detection. If MSK is decoded
incoherently, then the disadvantage for MSK is 6 dB. We agree on that.
(Decoding incoherently may be a worthwhile tradeoff, however. Lock time is
much faster, and the circuitry is much simpler. Dale chose to use incoherent
demodulation with our modem to achieve faster lock time in a half duplex
packet radio environment.)
>Furthermore, MSK requires smaller bandwidth than BPSK and, beeing its
>envelope constant, it can be nonlinearly (C class) amplified. Filtered
>BPSK envelope is not constant and the power amplifier should be operated
>far below saturation.
MSK does *not* have a constant amplitude envelope. It does not pass through
zero when the vector goes from bit to bit as BPSK does, but there is a 3 dB
variation as the vector moves from bit to bit. (This is the chord to arc
amplitude
difference as the vector rotates 90 degrees.) If you amplify it with a Class C
amp,
the spectrum will balloon to a larger value than if you use a linear amp.
As a specific example with our 56 kB modem, using a linear amp, the spectrum
is 1.4 Hz per baud (at the -22 dB points). Using a Class C amp, the spectrum
expands to a bit more than 3 Hz per baud (at the -22 dB points).
BPSK is worse, of course. The spectrum at the -22 dB points is about 15 Hz
per baud even if a linear amp is used. That's because the secondary and
tertiary
sidelobes of BPSK are much larger in amplitude than those of MSK due to the
vector passing through zero on state transitions.
Note, both examples assume square bits. If you do baseband shaping, you can
narrow the occupied bandwidth of either system considerably. (PSK31 is an
example of that.) But MSK still wins for a given amount of shaping.
There are more complex modulations which can occupy a smaller bandwidth
than either BPSK or MSK for a given baud, but they give up either throughput
or Eb/No as a result. They also generally require channel equalization which
requires a training sequence. That's not workable in a multipoint to
multipoint
packet radio environment, though it can be practical in a full duplex fixed
point
to point envirionment (indeed, phone modems work that way).
Gary
Gary Coffman KE4ZV | You make it |mail to ke4zv@bellsouth.net
534 Shannon Way | We break it |
Lawrenceville, GA | Guaranteed |
>.
------------------------------
Date: Mon, 22 Nov 1999 10:44:33 -0500
From: Gary Coffman <ke4zv@bellsouth.net>
Subject: German packet radio
On Thu, 11 Nov 1999 18:16:33 -0600, "Steve Sampson" <ssampson@usa-site.net>
wrote:
>There's probably a chart somewhere (I'm not that much of an expert) that
>shows the optimum deviation. I suspect it should be at an index slightly
To be continued in digest: hd_99_306D
Read previous mail | Read next mail
| |
