Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
zBITX CW Keyer
|
Ken, you may be right. I looked at some benchmarks and it looks like the pinfour can be four times or more faster than the 3B plus.?
if that turns out to be the limitation then experienced CW operators might just have to wait for a faster processor or more RAM or something
It’s just surprising because a small change in the software of the SBitx, even tripling. sampling rate would likely work just fine. Johnshoy’s fix multiplied sampling by 20 times!! ? That was probably overkill, but it works great. You would think a similar solution on the slower processor could be arranged to give similar outcomes.
however, perhaps not!
?
A new processor comes out every so often so eventually, this will get fixed!!
?
gordon kx4z?
?
? |
|
On Mon, Mar 31, 2025 at 02:37 AM, Gordon Gibby KX4Z wrote:
Why only 18? sBitx has an RPi 4, the zBitx has an RPi Zero 2 W, effectively a low-memory RPi 3B+ (the CPU on the Zero 2 is the same as the CPU on the RPi 3B+).
?
The processing of the RPi 4 is significantly greater than the RPi Zero 2.
?
Ken, N2VIP |
|
?
Looks like a copy paste error here for mode B
?
? ? ? ? if (cw_mode == CW_IAMBICB){
? ? ? ? ? ? if (cw_next_symbol == CW_IDLE && cw_last_symbol == CW_DOT && (symbol_now & CW_DASH)){ ? ? ? ? ? ? ? ? cw_next_symbol = CW_DASH; ? ? ? ? ? ? } ? ? ? ? ? ? if (cw_next_symbol == CW_IDLE && cw_last_symbol == CW_DASH && (symbol_now & CW_DOT)){ ? ? ? ? ? ? ? ? cw_next_symbol = CW_DASH;???? this should be CW_DOT I think ? |
|
I cleared out /var/log and I updated the software this morning. The CW keying in Iambic B is still rough... at 22 WPM keying K gets me a Q maybe 20% of the time. I tested this with two paddles On Tue, Apr 1, 2025 at 11:44?AM John AE5X via <ae5x=[email protected]> wrote:
|
|
This just the kernel latency. On Tue, Apr 1, 2025, 9:05 PM Lee via <kx4tt=[email protected]> wrote:
|
|
Could this be as simple as debouncing the keying circuit? I have a 0.1 uf disc ceramic across the output of my J-36 paddles to prevent this. Lee KX4TT On Tuesday, April 1, 2025 at 11:00:00 AM EDT, Gerald via groups.io <hb9cey@...> wrote:
The new firmware v1.05 / v3.051 is slightly better than before, but since CW still doesn't work properly for me even with the new firmware, I compared it with the QCX-Mini. Both transceivers were keyed and monitored in parallel, at the same speed, using a paddle. I also tried unsuccessfully with another paddle. ? While the QCX-Mini didn't lose any characters, the zBITX repeatedly produced errors, dots and dashes were lost or appended. This happened at slow and also faster speeds. This reminds me of my simple diy touch paddle where I always had to wet my fingers otherwise exactly such errors occurred. ? It seems strange to me that everything works fine for other users. ? If I use a macro from the zBITX to transmit, the zBITX keys everything flawlessly up to 50 wpm. No lost dots or dashes. How can this be, and where is the difference between keying by macro and the paddle input? I'm neither an IT nor an electronics engineer, but I would first look for the problem with the key input. Perhaps someone can explain this to me. ? 73, Gerald |
|
The new firmware v1.05 / v3.051 is slightly better than before, but since CW still doesn't work properly for me even with the new firmware, I compared it with the QCX-Mini. Both transceivers were keyed and monitored in parallel, at the same speed, using a paddle. I also tried unsuccessfully with another paddle. ? While the QCX-Mini didn't lose any characters, the zBITX repeatedly produced errors, dots and dashes were lost or appended. This happened at slow and also faster speeds. This reminds me of my simple diy touch paddle where I always had to wet my fingers otherwise exactly such errors occurred. ? It seems strange to me that everything works fine for other users. ? If I use a macro from the zBITX to transmit, the zBITX keys everything flawlessly up to 50 wpm. No lost dots or dashes. How can this be, and where is the difference between keying by macro and the paddle input? I'm neither an IT nor an electronics engineer, but I would first look for the problem with the key input. Perhaps someone can explain this to me. ? 73, Gerald |
|
Today I installed and tested firmware version 1.05 / 3.05.
?
SSB and FT8 work as before. There is a delay in both modes of about 1 second when switching RX/TX. In SSB this is not ideal in contests and makes contacts difficult. I managed to log 16 Qs in the wpx contest, but it could have been many more.
?
CW still doesn't work correct in iambic A and B. I have one CW contact in the log, and this was made with the installed macros, no chance with a keyer. It doesn't work in straight mode with an external keyer either. Overall in my opinion, CW is in its current state unusable. I hope it's not a hardware issue and that a firmware update can solve the problem! Actually I wanted to use the zBITX primarily in CW which is, at least for the moment, not possible. It would be interesting to know if my device is an isolated case.
?
73, Gerald
?
PS. The base plate gets quite hot during operation, which makes it uncomfortable to hold with your hand when using it portable. |
|
toggle quoted message
Show quoted text
On Mar 30, 2025, at 22:52, Ashhar Farhan via groups.io <farhanbox@...> wrote:
|
|
The morse keying is now in a separate thread with critical priority. Pse read modem_cw.c . I can manage upto 18 wpm on the iambic key now On Mon, Mar 31, 2025, 7:25 AM Mike Johnshoy via <mike.johnshoy=[email protected]> wrote: Gordon and everyone? - |
|
Gordon and everyone -
You might be right. I don't have a zbitx, but looking in the zbitx repository, sbitx_gtk.c has code around line #4301 that looks like where the change worked on the sbitx. Looks like there have been some changes to ui_tick() and I can't say if this will be effective - but I would try it! Good luck. -- Mike KB2ML |
开云体育You might want to look at the Winkeyer by Hamcrafters.? This can take a string of characters from a host system, read them into an internal buffer, and handle correct output timing. Gerry Sherman On 2025-03-31 01:11, Ron, N6YWU via
groups.io wrote:
|
|
If the current source code posted is up-to-date, I don't think an external keyer will help. ?
?
The pin connected to the key input is being polled within a GUI process by code in sbitx_gtk.c . ?Normal processes running on a loaded non-real-time Linux kernel on a single-core Raspberry Pi can have maybe up to 20 milliseconds or more latency when polled without interrupt handlers or kernel mods. ?This is a bit too long for decent CW keying at common WPM speeds. ?The key input handling could perhaps better be done in an interrupt routine, or using pigpio time-stamps to de-jitter the timing. ?Or possibly with some real-time kernel modifications to the standard Raspberry Pi Linux distribution.
?
73, Ron, N6YWU |
|
The term is "iambic", pertaining to a dihDAH dihDAH dihDAH pattern, each of the three sounds in the above called an "iamb". This are "I" as in India, not "L" as in Lima. Also, "Lambic" is a real thing. A Lambic is a beer or ale to which wild yeasty beasties are introduced to the wort by exposing it to natural wind and breezes. It's the sourdough of beer. It's easy to find combined with fruit juice or extract, apricots, peaches, and cherries are popular. I love cherry Lambic, and the beer drink called a "Dirty Hoe", composed of raspberry or cherry Lambic and Hoegaarden wheat beer.? 73, Gwen, NG3P? |
Lee
Gwen Patton
to navigate to use esc to dismiss