{k3} PSK31 tuning

[Ron Herman]

How does one tune PSK31 on the K3 such that the built in decoder can
read it. Tuning seems to be very selective and without a waterfall
nearly impossible to tune. The capture range of the AFC doesn't seem
to be very wide. The CWT spotting function sometimes helps but only
on stronger signals. Sensitivity to weak signals doesn't seem to be
very good either. With digipan on the PC I can decode stuff that I
can hardly hear. Any ideas?.

Ron, K1PDY

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[KK7P]

  1) Use CWT.

2) Use a narrow filter.  Once you find the station you want to copy,
tighten the filter and retune as necessary.  50 Hz is not too tight.  
300 Hz is too wide.  If you are hearing two stations, then the
selectivity is not tight enough.

3) Use auto-spot.  You may have to hit it two or three times.  Tuning is
critical and an error of more than 5 Hz changes from solid copy to no copy.

73,

Lyle KK7P

> How does one tune PSK31 on the K3 such that the built in decoder can
> read it...
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[John Ragle]

     Please don't take this comment to be frivolous or intentionally
snippy...but my take on the decoding functions built into the K3
firmware is that at this stage they are gimmicks or toys.

     As Ron observes, even an old warhorse like Digipan works much
better (...more easily, less fussily...) than the K3 routines (and I
love my K3!). When one considers the range of other digital methods
available in, e.g. fldigi, it is apparent that Elecraft is thinking in
the right direction but is not all the way there, YET.

     I use Digipan or fldigi (3.12.5) routinely with my K3 without
problems. For digital mode operation, SpectraVue is also a very nice
adjunct program.

     I have every faith that Elecraft will eventually enhance the
digital decoding functions in the K3. One will still want something like
SpectraVue to provide a window into the digital ops part of the spectrum
(or an LP-PAN or a P3, if funds permit). Eventually it will be nice to
be computer-free if one desires it...portability would be enhanced
somewhat, though a small laptop (e.g. a netbook) is not much different
in volume from an LP-PAN or a P3. My small ASUS Eee901 runs software
like Digipan and SpectraVue without flinching.

     I am not at all a Luddite WRT new solutions...just very oriented
toward functionality. This is also the reason that I rid myself of my
Flex 3000.

John Ragle -- W1ZI

=====

On 8/12/2010 11:46 AM, Ron Herman wrote:
> How does one tune PSK31 on the K3 such that the built in decoder can
> read it. Tuning seems to be very selective and without a waterfall
> nearly impossible to tune. The capture range of the AFC doesn't seem
> to be very wide. The CWT spotting function sometimes helps but only
> on stronger signals. Sensitivity to weak signals doesn't seem to be
> very good either. With digipan on the PC I can decode stuff that I
> can hardly hear.
>    

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[KK7P]

  They are intended for casual operation and within the confines of the
LCD panel of the K3.

I've found both the RTTY and the PSK decoder to work well when checked
against DM780, Digipan, and MMTTY.

But certainly not as quick as clicking on a waterfall :-)

73,

Lyle KK7P

>       Please don't take this comment to be frivolous or intentionally
> snippy...but my take on the decoding functions built into the K3
> firmware is that at this stage they are gimmicks or toys.

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Kok Chen]

On Aug 12, 2010, at 8/12    9:09 AM, Lyle Johnson wrote:

> ... Tuning is critical and an error of more than 5 Hz changes from  
> solid copy to no copy.

With a binary PSK demodulator, you can be off tuned up to an amount of  
plus or minus 90 degrees of perfect phase coherency between bit  
periods (what is sometimes called a "modulation chip").

BPSK31 has a bit rate of 31.25 per second. That means the carrier or  
pseudo carrier can be off up to 90 degrees in 1/31.25 of a second,  
i.e., 7.8 Hz is how much a demodulator can be off tuned under good SNR  
conditions.

With QPSK31, everything is halved since the signal constellation is  
doubled but all still lie on the circumference of a single circle.

If a higher order DPSK demodulator is used (for example, an Okunev  
demodulator), then the tuning range will be reduced accordingly.  A  
second order Okunev for BPSK31 requires 5.2 Hz tuning accuracy, a  
third order Okunev requires 3.9 Hz, and so on.

Higher order demodulators can give very significantly fewer decoding  
errors, but you also need to tune more carefully. Higher order  
demodulators (ditto coherent demodulation schemes such as the Costas  
Loop) also perform poorly when there is HF multipath.

Under non-ideal propagation, the simplest non-coherent differential  
PSK demodulator often works best for copying BPSK31.  It is also the  
most forgiving when it comes to tuning accuracy.

73
Chen, W7AY


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Dennis KB7ST]

Ron,

I share the same challenges with PSK and RTTY decoding.  My baseline is to operate my K3 while RV camping and to explore as much as I can without using a computer.  I don't have full answers and folks like Lyle have been most helpful.  I was last on HF 25 years ago so my K3, SN 4335, is pretty new.  I think my challenge has been to  better identify some aids to help in the tuning.  See http://www.netdave.com/wa0ttn/Tuning.asp but note he demonstrates some concepts using a PC and I still want to tackle while camping.  As a work in progress, here are some things I've noticed and some questions for the group.

WA0TTN article above mentions to "keep the PSK31 signal in the center of the most sensitive audio region of your receiver."  I have 8-pole 2.8 and 8-pole 400 Hz filters but still uncertain where the most sensitive audio region of the K3 is.

The tactics I'm using right now is that I set one VFO on data mode and the other VFO mode on CW, using 440 Hz (the A above C on the musical scale--the 'perfect pitch' A.  It isn't clear to me yet, despite some poking around on the net what two frequencies are used for PSK signaling.  I'm assuming that if I knew the 2 freqs then I could pick a center freq and use some PITCH setting in CW mode to get me close?  For example, one of the PSK macros submitted by Matt, W6NIA, in the K3 Utility Command Tester Help file shifts the IF center freq to 1100 Hz and changes filter bandwidth to 1800 Hz.  But there isn't explanation of why those numbers were chosen and I haven't tried IF shifting before.  In other words, when I can't get a quick decode using the recommended techniques, what can I do in CW mode to 'get me in the ballpark."

I do notice that I can decode much weaker PSK signals than RTTY signals -- and that is still a challenge for me (have some disabilities).

Within the last day or two, it occurred to me that I should tune into the W1AW data transmissions 14.095 at 3PM and 6PM Pacific time.  Then I assume that once I get a solid decode, I could switch into CW or some other mode to compare what the audio should sound like.  It would be neat to pick a certain CW pitch freq then SPOT it in CW mode to get me closer.  SPOT in PSK mode sometimes works and sometimes doesn't, but without a computer I can't analyze why it works or why it doesn't work.  I haven't run into too much QRM as an issue and SPOT PSK mode certainly works pretty good on strong signals.  SPOT CW mode on actual CW signals pretty much is nearly 100% in matching whatever PITCH tone I select.  But, as the more experienced folks say, the SPOT PSK technique doesn't get a PSK decode with the same reliability as SPOT CW when actually doing CW QSOs.

I'll be following this thread most carefully.

Dennis KB7ST Snohomish WA

[Greg - AB7R]

One thing that is important to remember is to make sure the threshold setting is
set so that the signal indicator on the CWT meter is set to beat with the signal.  
If you have it set for copying moderately strong signals and you try to copy a weak
one you may need to adjust the threshold a bit.

-------------------------
73,
Greg - AB7R
Whidbey Island WA
NA-065


On Thu Aug 12 12:04 , Dennis KB7ST  sent:

>
>Ron,
>
>I share the same challenges with PSK and RTTY decoding.  My baseline is to
>operate my K3 while RV camping and to explore as much as I can without using
>a computer.  I don't have full answers and folks like Lyle have been most
>helpful.  I was last on HF 25 years ago so my K3, SN 4335, is pretty new.  I
>think my challenge has been to  better identify some aids to help in the
>tuning.  See http://www.netdave.com/wa0ttn/Tuning.asp but note he
>demonstrates some concepts using a PC and I still want to tackle while
>camping.  As a work in progress, here are some things I've noticed and some
>questions for the group.
>
>WA0TTN article above mentions to "keep the PSK31 signal in the center of the
>most sensitive audio region of your receiver."  I have 8-pole 2.8 and 8-pole
>400 Hz filters but still uncertain where the most sensitive audio region of
>the K3 is.
>
>The tactics I'm using right now is that I set one VFO on data mode and the
>other VFO mode on CW, using 440 Hz (the A above C on the musical scale--the
>'perfect pitch' A.  It isn't clear to me yet, despite some poking around on
>the net what two frequencies are used for PSK signaling.  I'm assuming that
>if I knew the 2 freqs then I could pick a center freq and use some PITCH
>setting in CW mode to get me close?  For example, one of the PSK macros
>submitted by Matt, W6NIA, in the K3 Utility Command Tester Help file shifts
>the IF center freq to 1100 Hz and changes filter bandwidth to 1800 Hz.  But
>there isn't explanation of why those numbers were chosen and I haven't tried
>IF shifting before.  In other words, when I can't get a quick decode using
>the recommended techniques, what can I do in CW mode to 'get me in the
>ballpark."
>
>I do notice that I can decode much weaker PSK signals than RTTY signals --
>and that is still a challenge for me (have some disabilities).
>
>Within the last day or two, it occurred to me that I should tune into the
>W1AW data transmissions 14.095 at 3PM and 6PM Pacific time.  Then I assume
>that once I get a solid decode, I could switch into CW or some other mode to
>compare what the audio should sound like.  It would be neat to pick a
>certain CW pitch freq then SPOT it in CW mode to get me closer.  SPOT in PSK
>mode sometimes works and sometimes doesn't, but without a computer I can't
>analyze why it works or why it doesn't work.  I haven't run into too much
>QRM as an issue and SPOT PSK mode certainly works pretty good on strong
>signals.  SPOT CW mode on actual CW signals pretty much is nearly 100% in
>matching whatever PITCH tone I select.  But, as the more experienced folks
>say, the SPOT PSK technique doesn't get a PSK decode with the same
>reliability as SPOT CW when actually doing CW QSOs.
>
>I'll be following this thread most carefully.
>
>Dennis KB7ST Snohomish WA
>--
>View this message in context: http://elecraft.365791.n2.nabble.com/k3-PSK31-tuning-

tp5416698p5417461.html
>Sent from the Elecraft mailing list archive at Nabble.com.
>______________________________________________________________
>Elecraft mailing list
>Home: http://mailman.qth.net/mailman/listinfo/elecraft
>Help: http://mailman.qth.net/mmfaq.htm
>Post: [hidden email]','','','')">[hidden email]
>
>This list hosted by: http://www.qsl.net
>Please help support this email list: http://www.qsl.net/donate.html


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[KK7P]

  The K3's PSK encoder and decoder (used in DATA: PSK D mode) uses  a
pitch just above 1,000 Hz.

If you narrow the badnwidth and use the CWT feature, you'll get the PSK
signal.  YOU can use the SPOT function in conjunction with the CWT
indicator in this mode as described in the K3 manual.

For RTTY, you have choices for the MARK tone using the PITCH button
while selecting DAT AFSK or DATA FSK modes.  You must manually tune the
RTTY signal using the tuning indicator as described in the K3 manual.  
The MARK pitch you choose need not be the same the MARK pitch used by
the sending station.

73,

Lyle KK7P

> I share the same challenges with PSK and RTTY decoding....

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Dennis KB7ST]

Lyle and Ron,

Lyle, thanks on the PITCH freq for PSK mode.  That would explain why that macro has the IF shifted to 1100. I've spent last two days using W1AW twice per day data transmissions as a reference and have now developed an approach to more quickly gain success as PSK decodes and reduce frustration. I need to experiment a bit more but the method has significantly increased my ability to decode weak PSK (S1-S2 levels) on 20M during West Coast afternoons for South America.

Ron, perhaps you might want to experiment with this (unless you've given up)?  Again, I want to increase PSK/RTTY operations while camping using QRP levels without a computer.  I use a DSK9SQ telescoping mast and an endfed wire with 28-gauge teflon insulation.  Wire about 130-200 feet long (I have a splice to lengthen it depending on camp trees, etc).  Ground stake with various length radials thrown into bushes 6-10 feet high per one of latest QST articles on elevated radials.  Endfed goes up vertically along mast to about 25 feet and then horizontal so height above ground not much more than 20 feet.  I'm building a compressed air launcher to get higher and try some different antennas with the mast.  Also using the amazing T1 auto tuner.

My testing only been on 20M so far. I have 400 Hz CW and 2.8 SSB filters.

I set VFO B to CW and use PITCH of 800 Hz (highest it can go).  Starting at 14.070 and B/W of about 1000 Hz, I scan the PSK segment for strongest signals.  I then narrow the CW B/W progressively down to narrowest 50 Hz, SPOTing several times.  When I play with THRESHOLD settings, I notice three modes -- high thresholds and the CWT tuning bar disappears, middle numbers and it flickers a lot.  Low theshold numbers and the tuning aid more solid. After getting a good CW SPOT, I check the frequency.  Let's say the last three digits are 500.

Flip over to VFO A, now in PSK data mode and tune to the same frequency.  Technique now varies a bit so experimenting fun.  I have progressively narrowed PSK B/W and do several SPOTS.  The upshot is that I'm getting good decodes with PSK B/W of around 50-100 Hz.  Based on what Lyle said about the PSK tone being around 1000 Hz and my use of CW SPOT of 800 Hz, I suspected the PSK freq would be several hundred Hz away from the CW freq.  The last 10 successful decodes I got, however, shows the PSK frequency to be 20-50 Hz below the successfull CW SPOT freq on the PSK signal.  I'm not sure why that should be the case.

In summary then, my success went way up by 1) SPOT the PSK signal on VFO B using CW mode and PITCH of 800 Hz and narrowest B/W on CW spot then 2) go to VFO A in PSK mode, tune the freq to the VFO B CW spot freq and SPOT the PSK signal progressively, trying to get B/W as narrow as possible.  And, again, I notice good PSK decode freq has been wit 20-50 Hz below the good CW SPOT freq on VFO B.

Lyle, re RTTY,  the manual and K3 operating tips are not very clear to me as why there are different RTTY pitch frequencies.  When I was last on HF, 2125 was the MARK freq.  What is the advantage of choosing a MARK tone other than 2125?

Here is where I'm stuck on decoding RTTY.  I tune the signal from lowest freq and note the point at which the leftmost three bars are solid and note the frequency.  I assume the SPACE freq would be 170 Hz higher so I tune up half the distance (85 Hz) and don't get a decode even varying the tuning.  When I tune higher so that the rightmost three bars are solid, still no good decode.  I use the FSK RTTY mode that I can use the paddles with.  Maybe I should drag the laptop downstairs (part of my disability has to do with back pain).  It used to be that a number of non-ham band stations would transmit RTTY for long periods of time and it was a great way to practice decoding techniques.  When I telephoned the Seattle Coast Guard station here last week, they laughed when I asked for the freqs and times they transmit RTTY!

Thanks for your patience, Lyle, and good luck, Ron.

[KK7P]

  All that really matters is the shift (170 Hz) and rate (45 bps/60WPM,
or 75bps/100WPM).  The radio takes care of the rest, so you can use the
ptich that you prefer.

I've heard lots of stories about why the 2125/2295 Hz tone pair was
used. Some of the stroies make sesne, but I haven't taken the time to
research and satisfy myself about the "why" so I won't speculate here.
  In the early days of packet on HF, we picked 1700 Hz as the center
frequency because tyapical radios of the day with "CW" IF filters put
the center of the passband there when used in SSB mode.

The four pitch selections we provide in the K3 are based on defaults in
programs like MMTTY.

To tune RTTY, you want both indicatrors to be of similar width in the
CWT.  If it still doesn't decode, then go to the opposite sideband, but
that shouldn't be necessary.

Practice onthe W1AW RTTY bulletins.  Those transmissions are long enough
to allow you to practice tuning.  Many  RTTY "CQ" transmissions are too
short to tune in, especially during contests!

73,

Lyle KK7P

> Lyle, re RTTY,  the manual and K3 operating tips are not very clear to me as
> why there are different RTTY pitch frequencies.  When I was last on HF, 2125
> was the MARK freq.  What is the advantage of choosing a MARK tone other than
> 2125?
>
> Here is where I'm stuck on decoding RTTY...
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Kok Chen]

On Aug 14, 2010, at 8/14    9:57 AM, Dennis KB7ST wrote:

> KB7ST: What is the advantage of choosing a MARK tone other than 2125?

Many RTTY ops don't like to listen to such a high pitched tone pair in  
their headphones.  If you are tuning by ear, it is also easier to  
figure out how far and which direction you are off tuned when you use  
a lower tone pair (Weber-Fechner law).

The real question should be what is the advantage of choosing the  
2125/2295 Hz tone pair?

In the bad old days, the harmonics of a poor 2 kHz AFSK tone pair are  
better suppressed because they fall outside of the SSB transmit filter  
passband.  This has not really been a problem with current radios if  
you use the output sound card properly -- using close to full scale of  
the output sound card, but never saturating either the sound card or  
the rig's analog audio stages.

> KB7ST: I assume the SPACE freq would be 170 Hz higher ...

The Mark tone (the "rest" tone (a.k.a. stop bit) if you are not  
sending diddles) is by RTTY convention the higher of the two carriers  
on the RF spectrum.

If you are using LSB, Mark therefore becomes the lower of the two  
tones in the audio spectrum.  If you are using USB, Mark remains the  
higher of the two tones in the audio spectrum.

The Mark carrier is also by convention the reference frequency to spot  
an RTTY signal.  This convention is often overlooked by RTTY newbies,  
who would spot less than useful information, often their rig's SSB  
suppressed carrier frequency - causing some spots to be 2 kHz off :-).

73
Chen, W7AY

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Dennis KB7ST]

Hi Ron,

I'm pretty much done with my testing on decoding PSK31.  Using the method I tried, now I don't have to use the CW mode to help spot the right frequency to tune to in data mode.  I trained my ear to recognize the 1010 Hz spotting tone that Lyle, KK7P, pointed out.  And it's been much more successful since.  And, I'm pretty impressed at how well the K3 can pick out weak signals and decode them.  Next step is for me try it out on our next camping trip.

73 -- Dennis