Elecraft › [K3]

K3 TX delay

Classic

List

Threaded

12 messages Options

vk4tux

K3 TX delay

Is it possible to reduce the K3 TX delay down to 4ms from the current
8ms minimum for qsk use?

Adrian ... vk4tux
______________________________________________________________
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

Vic K2VCO

Re: K3 TX delay

No, you can't change this, but the best QSK performance comes from activating QRQ CW mode.

On 2/16/2011 2:17 AM, Adrian wrote:

> Is it possible to reduce the K3 TX delay down to 4ms from the current
> 8ms minimum for qsk use?
>
> Adrian ... vk4tux
> ______________________________________________________________
> 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

Bill VanAlstyne W5WVO

Re: K3 TX delay

... Which (QRQ mode) unfortunately isn't available, in practical terms, in
the absence of the KRX3 second receiver, since working split or XIT is
impossible. What a wonderful experience using this mode is, though! If only
I could use it for chasing DX running offset pile-ups. Really justifies
having the KRX3. It's on my must-have list, but a few other things have to
come first. :-(

I've also advocated to Elecraft in the past to make available RF delay
settings below 8 ms for those of us who have fast QSK amplifier switching
using vacuum relays or switching diodes. (My homebrew 6m amplifier T-R speed
is well under 4 ms, for example. The new Elecraft amp, I assume, will be
even faster.)

I'd also like to see the CW WGHT parameter (optionally) configure itself to
conform to the amount of RF delay selected in concert with the speed
selected for the internal keyer. Right now, the default CW WGHT value of
1.10 is optimized for a QSK RF delay of 8 ms (default) and a keyer speed of
20 WPM. (Do the math.) However, an ideal CW WGHT value could easily be
computed automatically from these two parameters, giving you the best-shaped
CW elements at different sending speeds, on the fly. (Of course, such a
feature would only work when using the internal keyer, not an external
one -- unless autodetection of CW sending speed was also incorporated.)

Bill W5WVO

-----Original Message-----
From: Vic K2VCO
Sent: Wednesday, February 16, 2011 16:00
To: [hidden email]
Subject: Re: [Elecraft] K3 TX delay

No, you can't change this, but the best QSK performance comes from
activating QRQ CW mode.

On 2/16/2011 2:17 AM, Adrian wrote:

--
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco/
______________________________________________________________
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

______________________________________________________________
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

W5UXH

Re: K3 TX delay

I am under the impression that the TX DLY setting of "8 msecs" does not actually produce an 8 msec delay between KEY OUT going low and start of RF. In non-QRQ mode I think it is on the order of 15 to 17 msecs. The delay does increase approximately 1 msec for each unit increase in TX DLY.

In QRQ mode, I measure the delay between 4 and 5 msecs when TX DLY is set to 8, and I believe it does not increase when TX DLY is increased.

Chuck, W5UXH

Bill VanAlstyne W5WVO wrote

I've also advocated to Elecraft in the past to make available RF delay
settings below 8 ms for those of us who have fast QSK amplifier switching
using vacuum relays or switching diodes. (My homebrew 6m amplifier T-R speed
is well under 4 ms, for example. The new Elecraft amp, I assume, will be
even faster.)

P.B. Christensen

Re: K3 TX delay

> I am under the impression that the TX DLY setting of "8 msecs" does not
> actually produce an 8 msec delay between KEY OUT going low and start of
> RF.

At the default 8 ms setting of TX DLY, I measure ~ 9 ms of delay in QRQ mode
between the K3's key *input* line and RF. So, I agree that the 8ms figure
is not associated with the time measured from the K3's amp key line to the
onset of RF. At least, that's the way it measures here.

In non-QRQ CW with TX DLY still at 8ms, I measure ~ 16 ms between key
closure and RF.

Paul, W9AC

______________________________________________________________
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

P.B. Christensen

Re: K3 TX delay

In reply to this post by W5UXH

> I've also advocated to Elecraft in the past to make available RF delay
> settings below 8 ms for those of us who have fast QSK amplifier switching
> using vacuum relays or switching diodes. (My homebrew 6m amplifier T-R
> speed is well under 4 ms, for example. The new Elecraft amp, I assume,
> will be
> even faster.)

I believe the larger issue is keeping good oscillator stability (to prevent
transmitter chirp) between T/R and R/T transitions. That's a reason why
Elecraft has not immediately implemented RIT in the K3 while in QRQ mode.
Folks with a KRX3 can get around the RIT issue with the second Rx. Even
when the K3 does have RIT in QRQ mode, I believe Wayne indicated that the
frequency delta will be quite restricted for this reason.

In the Ten Tec Omni Six series, CW chirp gets worse as the frequency delta
between the Tx and Rx increases. It can get pretty bad with RIT/XIT
excursions nearing 10 kHz. Most folks would not use RIT/XIT with such a
large offset and this effect mostly goes unnoticed. In the Omni V, T/R
times are actually slowed down a bit with larger VFO A/B split excursions
(no RIT/XIT in the V model) and chirp is non-existent.

I bring up the Ten Tec examples because the K3 is not immune from the same
oscillator settling effects. There's a definite trade-off between T/R
switch times and oscillator settling time when using a common oscillator for
Tx and Rx

Paul, W9AC

______________________________________________________________
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

W5UXH

Re: K3 TX delay

In reply to this post by P.B. Christensen

According to the manual TX DLY description:

"...sets the time from KEY OUT jack .. to first RF in 1-ms steps"

I agree with your measurement of 9 msecs from KEY IN to RF. Approximately 5 msecs is lost in the K3 between KEY IN going low and KEY OUT going low. At least in QRQ mode. I have not spent time looking at non-QRQ mode since QRQ is the only mode I use.

I normally use the KEY OUT jack to control the qsk boxes for amps and it is this delay that is important to me. I have implemented my own KEY OUT line that goes low coincident with KEY IN but goes high again approximately coincident with KEY OUT. This is needed with my miniature relay qsk boxes (a la AD5X) since the ~4 msecs setup time is too close for comfort to the relay switching times. It does not work with my PIN qsk boxes because the diodes switch before the K3 RX is muted, producing nasty clicks! I normally use diode qsk, so the KEY OUT line is fine for that.

Chuck, W5UXH

P.B. Christensen wrote

> I am under the impression that the TX DLY setting of "8 msecs" does not
> actually produce an 8 msec delay between KEY OUT going low and start of
> RF.

At the default 8 ms setting of TX DLY, I measure ~ 9 ms of delay in QRQ mode
between the K3's key *input* line and RF. So, I agree that the 8ms figure
is not associated with the time measured from the K3's amp key line to the
onset of RF. At least, that's the way it measures here.

In non-QRQ CW with TX DLY still at 8ms, I measure ~ 16 ms between key
closure and RF.

Paul, W9AC

W5UXH

Re: K3 TX delay

Oops. I attributed my relay qsk boxes to AD5X but forgot to attribute some of the diode ones to K7FJ!

Chuck, W5UXH

W5UXH wrote

....This is needed with my miniature relay qsk boxes (a la AD5X) since the ~4 msecs setup time is too close for comfort to the relay switching times. It does not work with my PIN qsk boxes because the diodes switch before the K3 RX is muted, producing nasty clicks! I normally use diode qsk, so the KEY OUT line is fine for that.

Guy, K2AV

Re: K3 TX delay

In reply to this post by Bill VanAlstyne W5WVO

I respectively submit that the K3 does NOT have a QSK problem.

It sounds more like a discussion on how-many-angels-on-the-head-of-a-pin
with the QSK switch time substituted for angels. It did have limitations
for very QRQ, but there is an option for that now. I like the hearing with
the QRQ mode QSK. And OF COURSE I would like that in effect at all times.
But that is the difference between breakfast and gourmet breakfast. Some
are talking as if there is no breakfast at all.

In going below 8 ms state change time, one is working against many
state-change processes in the K3 and the need for offsets to settle when
working split and other similar issues. The QRQ mode throws some of those
overboard in order to get really really fast QSK. (The QRQ QSK mode *IS*
really really fast.) Split and XIT are issues in this mode because there is
only one transmit string, extensively state-switched and reused for main RX.
This cute switching on the main board saves a TON of circuitry and board
space and COST for the transceiver. (You LIKED not having to spend ten
grand for your K3, didn't you....)

I have NEVER had an issue hearing in-between for either up-freq pileups or
on-freq listening for the station. The REAL pathetic timing issue is *MY*
brain's inability to stop mid-baud when I hear DX come back. The finish of
the letter (or two) seems to be on its own un-interruptible channel in my
cerebral cortex. I'd be tickled pink if *I* could respond in 8 ms. But
this response time is typical. Even the grandkids are not that fast, though
they are plenty faster than me. The weak element in turn-around time will
always be the HUMAN, by far. No contest. Nowhere near close. 8 ms is an
audio frequency (125 Hz). No human will be flipping decisions in the space
between two cycles of a 125 Hz sound wave.

In one very interesting (turned out to be human factors) set down about a
K3, where the complaint was not hearing between bauds, the situation was
remedied by changing the front end setting of "PRE" to "ATT" (yes on 160m,
too), backing off the RF gain until the band noise was a tad less than
moderate, and setting the AGC to slow. All of a sudden the QSK "sprang to
life". The operator was told to advance the AUDIO gain if a signal was in
the noise. It appears that QSK does not "sound right" when using fast AGC
(my own impression as well), though looking at audio spectrum recodings of
what I was hearing proved that the IMPRESSION was ALL IN MY HEAD. Was a
little deflating, but educational.

A penchant to blame the K3 at first perception leads to an inability to
drive it at the max.

Beyond all of that, most of the QSK restriction issues I have worked on
ultimately had to do with the transfer time **in the amplifier**, IMHO the
killer limiting factor in too-slow-QSK. This has been true with high end
Acoms, SPE Experts, older Alphas, and certainly any Ameritron not equipped
with their pin diode option (my AL1200 is extensively modified to get an
honest 5-6 ms switch time). In the case of two companies it took extended
arguments with their tech support to convince them there was an issue AT
ALL, and in both cases wound up in their having to invent a mod AFTER their
release, and sending the amp back to the factory to install it. (Wonder who
was testing things before release.)

Reducing transfer time to 4 ms had better be done with diode switches in the
amp because mechanical TR state change in amps is rarely finished within 4
ms...think consistent TWO milliseconds closure for the relay to support 4 ms
timing, relay closure intervals GRADUALLY LENGTHEN with time and use. EVERY
brand amp has this restriction with a mechanical transfer, whether they
admit it or not. Doesn't matter what they advertise. Will be watching the
KPA500, but it's pin diode I hear. I still wonder what happens to the
KPA500 pin diodes when you open with a dit into an unterminated coax and a
20:1 SWR. Wayne and Co are working their magic into a problem that has
taken most amps down. That problem IS Daniel into the lion's den,
regardless of who is playing the role of Daniel, even Wayne and Eric. They
have to actually beat the lions and so far there are many wounded out in
amplifier land.

I have had to replace fast vacuum relays twice in amps loaned out because
operators set their delay to minimum...less than 8 ms...thank you Wayne for
cutting that off at 8 ms, thank you, thank you. Transceivers were
transmitting RF before the relay contacts had settled and quit moving. This
is a formula for key clicks, relay contact burning, contact welding, and
burned bandswitches and carbon tracks everywhere. I have personally
repaired amps where nearly everything on the antenna side of the plate
blocking capacitor was ruined. The Orions will actually allow the delay set
at zero. The lame excuses from the relay-burner loanees went like "Well, we
were using the [insert brand] on the Alpha 89 just fine" (pin diodes). That
didn't cut it for me and all I got was "better sounding QSK" when I asked
them why they did it. And no, they didn't pay for the work.

I find that I am able to hear just fine between dits at ordinary speeds on
the K3, and I can be broken into at up to 35 wpm reliably WITHOUT needing
the QRQ mode. The very very fast QSK mode in the K3 allows for precision
code and RX/TX transfer at up to 75 wpm. In this mode all the various
features that kill the milliseconds in TX/RX state change are disabled so
that some can have their 75 wpm QSK.

73, Guy.

On Wed, Feb 16, 2011 at 2:04 PM, Bill VanAlstyne W5WVO
<[hidden email]>wrote:

> ... Which (QRQ mode) unfortunately isn't available, in practical terms, in
> the absence of the KRX3 second receiver, since working split or XIT is
> impossible. What a wonderful experience using this mode is, though! If only
> I could use it for chasing DX running offset pile-ups. Really justifies
> having the KRX3. It's on my must-have list, but a few other things have to
> come first. :-(
>
> I've also advocated to Elecraft in the past to make available RF delay
> settings below 8 ms for those of us who have fast QSK amplifier switching
> using vacuum relays or switching diodes. (My homebrew 6m amplifier T-R
> speed
> is well under 4 ms, for example. The new Elecraft amp, I assume, will be
> even faster.)
>
> I'd also like to see the CW WGHT parameter (optionally) configure itself to
> conform to the amount of RF delay selected in concert with the speed
> selected for the internal keyer. Right now, the default CW WGHT value of
> 1.10 is optimized for a QSK RF delay of 8 ms (default) and a keyer speed of
> 20 WPM. (Do the math.) However, an ideal CW WGHT value could easily be
> computed automatically from these two parameters, giving you the
> best-shaped
> CW elements at different sending speeds, on the fly. (Of course, such a
> feature would only work when using the internal keyer, not an external
> one -- unless autodetection of CW sending speed was also incorporated.)
>
> Bill W5WVO
>
>
>
> -----Original Message-----
> From: Vic K2VCO
> Sent: Wednesday, February 16, 2011 16:00
> To: [hidden email]
> Subject: Re: [Elecraft] K3 TX delay
>
> No, you can't change this, but the best QSK performance comes from
> activating QRQ CW mode.
>
> On 2/16/2011 2:17 AM, Adrian wrote:
> > Is it possible to reduce the K3 TX delay down to 4ms from the current
> > 8ms minimum for qsk use?
> >
> > Adrian ... vk4tux
> > ______________________________________________________________
> > 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
>
> --
> Vic, K2VCO
> Fresno CA
> http://www.qsl.net/k2vco/
> ______________________________________________________________
> 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
>
>
> ______________________________________________________________
> 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
>

______________________________________________________________
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

W5UXH

Re: K3 TX delay

And at least one of us has thanked Wayne for this. I don't run over 65 wpm, but the keying and qsk performance is excellent in QRQ mode at this speed! One does not need RIT, SPLIT etc. for a good QRQ ragchew, so I am very happy with the K3.

Chuck, W5UXH

Guy, K2AV wrote

In this mode all the various
features that kill the milliseconds in TX/RX state change are disabled so
that some can have their 75 wpm QSK.

73, Guy.

Andrew Moore-3

Re: K3 TX delay

And another thanks Wayne for this... It was a quick, clever and relatively
cheap (all in software) fix and a result of the Elecrew listening to
customers -- and it's a clear example of taking a fresh look at a problem
and coming up with an innovative solution. Not only did it improve QSK but
code timing at insane speeds too.

For the QRQ QSK crowd, it's high performance to match the rest of the K3.

Congrats to the Elecrew for this impressive engineering effort.

--Andrew, NV1B
..

On Thu, Feb 17, 2011 at 12:06 PM, W5UXH <[hidden email]> wrote:

>
> And at least one of us has thanked Wayne for this. I don't run over 65
> wpm,
> but the keying and qsk performance is excellent in QRQ mode at this speed!
> One does not need RIT, SPLIT etc. for a good QRQ ragchew, so I am very
> happy
> with the K3.
>
> Chuck, W5UXH
>
>
> Guy, K2AV wrote:
> >
> > In this mode all the various
> > features that kill the milliseconds in TX/RX state change are disabled so
> > that some can have their 75 wpm QSK.
> >
> > 73, Guy.
> >
>
> --
> View this message in context:
> http://elecraft.365791.n2.nabble.com/K3-TX-delay-tp6031150p6036729.html
> Sent from the [K3] 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: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

Bill VanAlstyne W5WVO

Re: K3 TX delay

In reply to this post by Guy, K2AV

Your explanation of the issues surrounding QSK in the K3 is well worth the
read, Guy, and I thank you for it. I’ll stop quibbling about milliseconds on
the T-R change-over. :-)

I am not a QRQ operator, far from it; I top out at around 30-35 WPM for
rag-chew copy in my head. But in regular QSK mode with default settings (8
ms delay, etc.), I do find it difficult to hear a weak DX station coming
back between my bauds. Your attribution of this to wetware factors is
compelling, and I’m sure you’re right, at least in part.

Another factor, however, might be that the pitch of the DX station is
typically the same as the pitch of my sending sidetone, since I have the DX
station centered in my passband -- which is usually set pretty tight. (I’m
not one of those guys who likes to listen to weak, QSB-laden CW in a wide
passband. It drives me nuts.) I could offset the DX station a few Hz by
moving the IF shift to one side, which I’ll try.

I also found this paragraph to be especially interesting:

“ ... where the complaint was not hearing between bauds, the situation was
remedied by changing the front end setting of "PRE" to "ATT" (yes on 160m,
too), backing off the RF gain until the band noise was a tad less than
moderate, and setting the AGC to slow. All of a sudden the QSK "sprang to
life" ... It appears that QSK does not "sound right" when using fast AGC (my
own impression as well) ... ”

I totally get it about reducing the front-end gain of the receiver on most
bands. However, with my current pathetic HF setup using a rather small wire
antenna (I’m primarily a 6-meter op, which is where the big aluminum is for
the time being), I’m really not troubled by too-strong signals, and the DX
stations I hear are typically below my AGC threshold. That may all change in
the near future, as I’m working on some new antenna ideas for my HF setup.
But for now, I don’t really see why changing the AGC response speed should
change anything. Nevertheless, I’ll try it.

I do find, as have others, I guess, that using the QRQ mode at regular CW
speeds makes detecting when the other station comes back much easier,
subjectively. I now have a much better understanding of what had to be
sacrificed to get that super-fast QRQ mode T-R switching, and why these
trade-offs are unavoidable. For this and a number of other reasons, like
being able to listen to the "up" DX pile-up in one ear, the KRX3 is going to
be put higher up on my must-have list.

Bill W5WVO

______________________________________________________________
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