LF: Fwd: CMSK Beta released

[Andy Talbot <andy.g4jnt@googlemail.com>]

---------- Forwarded message ----------
From: Murray Greenman <denwood@orcon.net.nz>
Date: 5 August 2010 07:14
Subject: CMSK Beta released
To:
Hi,
there has been a lot of interest in the new software, and I'm sure my answers will be of interest to all. I've posted some of your comments below.

First of all, the Beta version of the CMSK software and supporting web page has now been released:

http://www.qsl.net/z/zl1bpu/CMSK/cmsk.htm

Please pass this on to your friends. Andy, could you also post the link to the RSGB_LF_Group for me please.

Jay, what you found is very likely a small bug. I recall Con commenting about something similar in an early version. I do know that if the RX shuts off before all the received text has flushed, it is likely to be flushed out with the first of the TX text, and so come out the wrong colour (exacerbated by tight Squelch setting). Perhaps you could send Con (zl2afp@xtra.co.nz) a screen-shot and explanation, as it might help narrow down the problem. I run it on Win XP SP2 with no sign of your problem.

Yes, Con and I are of the opinion that MSK has the right credentials for LF and MF. You probably have more test opportunities than we do, and your assistance is appreciated. If you have any suggestions that might enhance the design or the performance of the protocol, we're listening.

John,
I'm pleased to hear that CMSK31 worked well for you. It's a bit slow of course, but if you're used to QRSS3, it's like lightning in comparison! The interleaver is has the same time spread as CMSK63, but because the frames are smaller the PN-sequence overhead is higher - thus its less than half the speed. The bandwidth is pretty good. I've had excellent 600m copy into VK2 (2200km) with CMSK31 - and in fact it's often better than the 500km N-S path to ZL2. Signals seem particularly strong around sunset in VK2.

Your suggestion of saving all the settings at the end of a session is a good one (are you listening, Con?). The RX centre and baud rate are easily set, but I agree, if you are operating the same frequency each day or perhaps waiting for someone to call on sked, it's nice to be prepared. The sampling rate should definitely be saved.

I'm interested that you don't find WSPR too good on 137kHz. While it wasn't intended for LF when Joe and I conceived it, it would be interesting to know what you think is the cause. Just too much QRM?

I've not managed a two-way QSO on LF or MF yet either, so you might beat me to it! I'm the only guy down here running an SSB exciter on LF/MF with digital capability at present. Con and I have had plenty of two-way QSOs on 80m, but that doesn't count!

By the way, I'd be interested in comments on how much documentation we need to put in the public domain for the mode to be accepted in other countries. I've described the technique and posted the varicode, but I'm sure that's not enough to replicate our work, which would be desirable if we want an enduring interest in CMSK. I am encouraging Con to write a little morsel about the algorithms used etc, which I'll perhaps 'ghost write' and post on the web site. I also plan to write a technical article for publication some time soon, once we've knocked out the loose bolts, rattles and squeaks.

We've only had access to 600m for a few months here, but I'm really impressed by how useful the band will be. We have grabbers in VK watching our every step, and there have been several un-scheduled 2-way QSOs, which is really heartening. There are at present 7 ZLs with 600m capability, and an 8th due any day.

73,
Murray ZL1BPU/ZL1EE

======================================================================
Jay W1VD:
-----------
John W1TAG forwarded a copy of the pre beta software as we work together on LF and MF wringing out
new software and hardware.

Found what may be an anomaly or (more likely) user error. Running Windows XP SP2 on Dell 4600 2.8
GHz, 1g  ram using a Delta 44 sound card. This setup works fine on all other available digital modes
software.

When switched to transmit the main screen shows a number of characters (~ 3 - 10) of the transmitted
message in red,  then the text switches to green for a number of characters. Some of the characters
make sense for hat is going on but others don't. No apparent rhyme or reason to the switching. I've
verified that the program is transmitting the correct message and when on receive will decode
correctly.

It may be something I'm doing wrong - if so, there's no need to be gentle ...

Thanks for your time and hard work developing the software. John and I have been convinced for some
time that MSK is the way to go. We've tried just about every other 'digital com' software and we've
determined MSK is the way to go. We normally judge software by 'static resistance' and 'fade
recovery' on 137 and 500 kHz. Some previous examples of MSK from other folks have exhibited rather
slow 'fade recovery'. Looking forward to running your FEC MSK. Sounds like it will be just the
ticket.

Jay W1VD  WD2XNS  WE2XGR/2

John W1TAG:
--------------
Jay, W1VD, and I have had a couple of one-way opportunities for testing
over a 317 kM path at 511 kHz, and I'm very encouraged.

Last night, Jay ran 100 watts at the 31 baud rate, and copy was flawless
here through static crashes and fading. We are going to try lower power
tonight, and the 7.8 baud rate as well if weather permits. We're in an
active thunderstorm season right now!

This noon, I copied the same 100 watt signal over a surface-wave only
path. The signal was barely audible, and would have made difficult CW
copy. Again, the 31 baud copy was completely solid. Not much for static
crashes or fades, just steady background noise.

My only wishlist for receiving at this point would be to have more of
the settings preserved -- sampling rate, baud rate and the rx center
frequency, in particular. The sampling rate on this laptop is really
off, and it takes a while to enter each time.

Jay will have some wishes on the transmit side, but I'll let him speak
for himself! We will do some more one-way playing this week, and can
graduate to 2-way next week when I get home from holidays. I'm
particularly curious to see how it does on 137 kHz with all of my
neighborhood SMPSU garbage. WSPR has been absolutely dismal down there
compared to 500 khz.

Nice job, and we promise to keep exercising the program.

John, W1TAG