Re: [linuxham] Re: [NBEMSham] Re: MT63 receive audio & waterfall, but no message decoding ... and more
- Guys,I recently downloaded and installed the latest Alpha release of FLDIGI (3.22.0AQ) and was calling CQ in MT-63/500 with TxID turned ON ( usually call CQ with TxID).The far end was a W7 station in Virginia ( I am in Central Florida). Once FLDIGI locked onto the signal ( I have RSID squelch time set for 2 seconds),there were no "hit" characters at all. In fact, we held quite a long QSO for more than an hour with no errors in the text on either side.I am currently being interfered with from an RFI source that produces 120 Hz harmonics on all bands (the source is AM rich) and "lines"appear almost all of the time. MT-63 had no issues or problems decoding in the presence of this RFI.Of particular interest to some may be that I use a VERY CHEAP Chinese version of the CM-109 USB headset FOB for the audio IN and OUT.I have found that some of these individual units can be very far off frequency when calibration is done with WWV and my MAC-PRO. Its particularly importantto note that each sound FOB is different! Some exhibit rates that require 100-200 ppm corrections, and some as little as 10-12 ppm.Even the internal sound card in the MAC-PRO is off by just a little bit (41 ppm). Just make sure that you don't overload the audio input! Try to run your audio "volume" at mid rangeand keep the level indicator out of the yellow and red zones. In my case I have the waterfall gain set at zero for most normal conditions and I go down to about -13 during quiet conditions.Each radio and interface will have its own best set of operating conditions, but I try to adjust the levels on all to be mid-scale which should result in things being operated in the most linearportions of their range.I have found that using the 48000 sample rate on these cards provides the most stable conditions and reduces the drift that is perceived.With respect to the MT-63/500 QSO mentioned above, I observed the frequency error to be less than 1 Hz all of the time and less than 0.5 Hz for the greaterportion of the time (both ends using both TxID and RxID on each "over"). Throughout this QSO there was evidence of selective fading and even some multipathfading going on, yet FLDIGI did not hiccup even once.Take away from this the fact that for the major portion of the user community, you SHOULD do the sound card calibration, for those interested in the most stable operation,you MUST do the sound card calibration. Take the time to do this and you will be rewarded with perfect copy on those stations that have also taken the time to do the calibration.On those that have just begun to use digital modes, they can be gently reminded that there are ways to improve decoding. On those others that don't care, at least you will bein the happy medium and secure that you are calibrated.For those interested - I usually can be found on MT-63/500 on 20m at 14.073 to 14.074, and on 40m at 7.038 to 7.040. I try to stay well out of the PSK31 areas as a 500 Hz wide signalwill cause lots of problems to the PSK users.On the other hand MT-63 will tolerate a fair amount of overlap by a PSK31 signal into the MT-63 signal and still copy perfectly.Dave has done wonders with FLDIGI and it keeps getting better. If you have or see any issues, bring them up here and a friendly chap will respond to help you out.Mitch Baum / AF2M
From: w1hkj <w1hkj@...>
Cc: firstname.lastname@example.org; email@example.com
Sent: Wednesday, September 21, 2011 10:12 AM
Subject: [linuxham] Re: [NBEMSham] Re: MT63 receive audio & waterfall, but no message decoding ... and moreOn 09/21/2011 08:32 AM, David Kleber wrote:
> I recently ran into a fellow with MixW where on Olivia I was getting very poor copy and we had no luck at all on MT63 1k long (this was all on HF). I suggested he try FLDIGI and we had no problem.
> I have seen an issue, only rarely, usually where you get about 10 good characters followed by 4 or 5 bad and it's usually with folks running MixW.
> I have been exchanging email with the fellow who I had decode problems with most recently and he found from the Tigertronics (Signalink) site that there is an issue with the USB Signalink when it is set to 11025, which, for some reason, is the default sample rate for MixW. The site suggests that you change the sample rate to 12000. He did this and it straightened out the problem.
That is a very interesting piece of news Dave. That means MixW is
relying on the operating system to perform sample rate conversion. So
long as the device sample rate is set to Native in fldigi's sound card
setup all of the sample rate conversions are performed internally using
very high quality software algorithms. Depending of the OS version (XP
...) and the sound card drivers the rate conversion can be quite
horrible. That apparently is what is going on with MixW on some Windows
computers. This phenomenon will not be limited to the SignaLink USB.
> In any case, I've never seen any problems we couldn't figure out with MT63 between stations running FLDIGI. Â The most common issues are calibration (which CheckSR solves although I'm not sure what you do if you're running Linux or Mac),
I have written a sound card calibration program that is similar to
CheckSR. It's been tested on Linux, but never OS X. My recommendation
is to use the WWV calibration routine in fldigi for calibration. It is
far superior to CheckSR in that all stations are then calibrated to the
same external standard. That doesn't help the operator who is limited
to VHF/UHF equipment. CheckSR (and my cal program) calibrate the sound
card against the computers cpu clock. The assumption is that the cpu
clock is accurate ... you know all about ASSUME. A secondary issue with
CheckSR and other such programs is that they usually rely on software
loop timing measurements. That type of measurement goes down the tubes
in a multi-tasking environment such as Linux, OS X, and most recent
versions of Windows. That said, CheckSR etal do put you close enough
for government work. You might expect to see some problems with MFSKpic
mode unless both ends of the transfer were either calibrated to an
external standard or to each other.
It is entirely feasible to create a time-tick modem in fldigi. A master
station could set it's PPM to 0 for both Rx and Tx. It would then send
a scheduled time-tick broadcast which all of the other stations in a
group would calibrate their sound card to using the WWV modem type.
I'll give some thought to creating such a modem type.
> bad sound levels (you can go way too low with no problems but overdriven audio on tx/rx or both is a BAD), DSP /Compression / Noise Reduction / Enhanced Stereo, Echo or Room Effects (Compression and DSP should be disabled on the radio and on the computer under the advanced settings of the manufacturer's ICON in the control panel, disable effects in record (input) and speaker (output). And, of course, make sure on HF that both stations are on USB even on 40 and 80m and extended characters and long (64 bit) interleave are set as the defualts.
> If you make sure all this points are adhered to, and you work within the acceptable S/N for the mode, MT63 is bullet-proof.
> -Dave, KB3FXI
> From: w1hkj<w1hkj@...>
> To: NBEMSham@yahoogroups.com
> Sent: Wednesday, September 21, 2011 7:52 AM
> Subject: Re: [NBEMSham] Re: MT63 receive audio& waterfall, but no message decoding ... and more
> I routinely test MT63 send and receive between MixW on a Vista machine
> and fldigi on a Linux (or OS X) machine. The two programs produce
> absolutely compatible sound streams. You must be aware that MixW and
> the latest versions of fldigi can transmit and receive on frequencies
> other than the standard low frequency of 500 Hz. I recommend setting
> both up so that they use the fixed audio frequencies. Next you must
> adjust both MixW and fldigi configuration for long interleave to be the
> same. This may be a bit confusing on MixW as it has more than 2 levels
> of interleave.
> Work out the differences in the software first between two computers on
> the same desktop. Then play with them in an on the air exchange.
> On 09/21/2011 04:56 AM, Charles Brabham wrote:
>> Come to think of it, one of the fellows in our weekly VHF net has experienced the same problem described here with MixW.
>> It sounds familiar, at any rate.
>> A lot of the guys in MARS use MixW. It's good software, but there may be a compatability issue of some kind.
>> Charles, N5PVL
>> ----- Original Message -----
>> From: Jim
>> To: NBEMSham@yahoogroups.com
>> Sent: Wednesday, September 21, 2011 12:06 AM
>> Subject: [NBEMSham] Re: MT63 receive audio& waterfall, but no message decoding ... and more
>> Problem was simple, sort of. My co-experimenter was using MixW at his end and both ends were getting garbled MT-63 output. Interesting note re is that the MT-63 waterfall intelligence was yellow against the otherwise blue background noise. He installed fldigi today and our results were the same as advertized for MT-63 and fldigi. Waterfall rx signal was slightly darker blue against the blue background noise. Just to verify, we last tried fldigi at my end and MixW at his end. Same symptoms as as described earlier: garbled Rx (at my end, no decoding at all).
>> Now that we know what works, we don't intend to troubleshoot why the delta between fldigi and MixW, but, if anyone has any ideas, maybe the group would be interested in case anyone is mixing programs.
>> BTW: we conducted all our tests over 20M. Next tests will be to acoustically couple over VHF.
>> BTW2: our intent is to use MT-63 for EmComm between the affiliated hospitals (Sutter Health) over a footprint extending from the Monterey Coast to the Sierra and up to the Oregon state line. Thus, we are testing HF to be used for the long stretches and VHF around the greater Sacramento area.
>> I've added to our punch list something (a script maybe?) to (semi)automatically take an "unwrapped" file and open it in its creator program. What I have in mind is to create HICS forms in an Excel spreadsheet with protected fields as labels for the text fields. Either the user or the message sender (i.e., the radio op) would save the file as a commma delimited CSV file so that its size would be minimized and transmitted as pure ASCII.. Perhaps large files (e.g., long lists of pharmaceuticals) could be zip'd to reduce file size even more. If anyone has done anything similar, I certainly would be interested to hear about it.
>> --- In NBEMSham@yahoogroups.com, w1hkj<w1hkj@...> wrote:
>> > On 09/20/2011 01:11 PM, Jim wrote:
>> > > Still working thru my problems. In testing with a compadre last eve, we both have receive audio and waterfalls from each other, but no intelligence decoded from the received signal. My fldigi "Overload Indicator" is green as it should be. Sound card calibration done. As with my transmit issue, I'm sure I have a cockpit problem here again. Have carefully read the fldigi manual and am still frustrated.
>> > >
>> > > Get excellent copy on RTTY but PSK31 slightly garbled, jamming words together without any breaks or spaces.
>> > >
>> > > Gotta be something simple, but ...
>> > Make sure that both ends have the same settings on the MT63
>> > configuration tab for
>> > 64-bit long interleave
>> > 8-bit extended characters
>> > Dave