|
For the CW operators here ... have you had any problems with the sBitx CW keyer?
The timing seems off for me when keying at higher speeds, and the TX isn't quite right even using an external keyer. I took a couple of videos showing what I mean:?
Using an external keyer at 25 wpm, the dits are much more staccato than they should be and sometimes entirely cut off. The TX is bad enough that the RBN will not pick up my call when using the external keyer:
https://youtu.be/GHbw9GroKAU
Using the sBitx internal keyer at 25 wpm, there are many keying mistakes introduced by the software, for example when keying the word "POTA", which is a word many of us will be keying of course:?
https://youtu.be/dtnPlcLXmdQ
Keying using a small keyboard or a save message seems to work fine, so that's good :)?
I noticed the code polls the GPIO associated with the key only about 10-20 times per second, which I think is much too slow for faster keying, so I was thinking about rewriting it to use GPIO interrupts to change the CW flags instead of repeatedly polling the GPIO.
Does anyone have thoughts on this as a possible approach??
Thanks,?
Steve W4SGN
|
I have an implementation to fix this. It is in the dev branch on the git
For the CW operators here ... have you had any problems with the sBitx CW keyer?
The timing seems off for me when keying at higher speeds, and the TX isn't quite right even using an external keyer. I took a couple of videos showing what I mean:?
Using an external keyer at 25 wpm, the dits are much more staccato than they should be and sometimes entirely cut off. The TX is bad enough that the RBN will not pick up my call when using the external keyer:
Using the sBitx internal keyer at 25 wpm, there are many keying mistakes introduced by the software, for example when keying the word "POTA", which is a word many of us will be keying of course:?
Keying using a small keyboard or a save message seems to work fine, so that's good :)?
I noticed the code polls the GPIO associated with the key only about 10-20 times per second, which I think is much too slow for faster keying, so I was thinking about rewriting it to use GPIO interrupts to change the CW flags instead of repeatedly polling the GPIO.
Does anyone have thoughts on this as a possible approach??
Thanks,?
Steve W4SGN
|
Confirming the farhandev branch does fix the issue. Thanks!?
|
Noticing the radio indicates power out of the transceiver using the firmware on the farhandev branch is lower, about 10-15 watts depending on band (according to the sBitx software). Is this an intended result??
|
Copy the hw_settings.ini over from the working one
Noticing the radio indicates power out of the transceiver using the firmware on the farhandev branch is lower, about 10-15 watts depending on band (according to the sBitx software). Is this an intended result??
|
?
Where is the corrected CW? - hw_settings.ini? -? available
for download ?
?
?
toggle quoted message
Show quoted text
From: Ashhar
Farhan
Sent: Sunday, July 7, 2024 11:49 AM
Subject: Re: [BITX20] sBitx CW keyer problems and idea for solution
#sBitx #sBITX_v3 #cw #firmware
?
Copy the hw_settings.ini over from the working one
On Sun, Jul 7, 2024, 12:52 AM Steve W4SGN via <stnagy@...>
wrote:
Noticing
the radio indicates power out of the transceiver using the firmware on the
farhandev branch is lower, about 10-15 watts depending on band (according to
the sBitx software). Is this an intended result??
|
There is no "corrected" version. Every radio gets their own tuned up file it is stored /home/pi/sbitx/data
?
Where is the corrected CW? - hw_settings.ini? -? available
for download ?
?
?
?
From: Ashhar
Farhan
Sent: Sunday, July 7, 2024 11:49 AM
Subject: Re: [BITX20] sBitx CW keyer problems and idea for solution
#sBitx #sBITX_v3 #cw #firmware
?
Copy the hw_settings.ini over from the working one
On Sun, Jul 7, 2024, 12:52 AM Steve W4SGN via <stnagy@...>
wrote:
Noticing
the radio indicates power out of the transceiver using the firmware on the
farhandev branch is lower, about 10-15 watts depending on band (according to
the sBitx software). Is this an intended result??
|
Took the sBitx out for a POTA activation this morning.?
The CW keyer on the dev branch is substantially improved, but I did find it was still making errors in both iambic modes. It did help to slow down to 20 wpm, but even at that speed I was still experiencing 1-2 errors per QSO.?
Given this seems like an area of active development, I am wondering if it makes any sense for me to attempt to improve the software, or if I should just hold tight for now. Happy to contribute in whatever way seems best.?
Steve, W4SGN
|
Steve, plunge in!
The development code is in farhandev branch.
I now have a seperate thread monitoring the keyer lines, it sets a? flag that is picked up by the signal processing loop in get_next_sample call (if I remember the function name correctly) to begin the key up/key down of the cw carrier.
There are many new easters in the farhandev branch. Some many not be usable. There is a qrp 60M band and AM mode too.
I hope you are comfortable forking on the github. Just start from there and work your own branch.
Took the sBitx out for a POTA activation this morning.?
The CW keyer on the dev branch is substantially improved, but I did find it was still making errors in both iambic modes. It did help to slow down to 20 wpm, but even at that speed I was still experiencing 1-2 errors per QSO.?
Given this seems like an area of active development, I am wondering if it makes any sense for me to attempt to improve the software, or if I should just hold tight for now. Happy to contribute in whatever way seems best.?
Steve, W4SGN
|
I got my sBitx yesterday and got it on the air today.? I've made three contacts with it already on 30m CW, but I'm glad to hear that someone is working on the keyer. Right out of the box, the keying is very rough.
?
Is there a thread here that will show me how to get this dev code, so I can try it out?
|
I'm also interested in CW improvements?to the straight key response (since I sometimes use an external keyer).? ?I tried looking at devfarhan and modem_cs but I couldn't really figure out what improvements have been made.
Can someone summarize any improvements made over, say, the last year, that would improve the CW response to the straight key?? ?I last tried that months ago and my impression from V3 code was that it still had some delays.? ?Does it still check only every 10-20 times per second, and is this adequate???
Sorry for the multiple questions, but figuring all this out is tedious. THANKS for any direction and wisdom, Gordon KX4Z
On Wed, Sep 18, 2024 at 8:02?PM Dan Romanchik KB6NU via <cwgeek= [email protected]> wrote: I got my sBitx yesterday and got it on the air today.? I've made three contacts with it already on 30m CW, but I'm glad to hear that someone is working on the keyer. Right out of the box, the keying is very rough.
?
Is there a thread here that will show me how to get this dev code, so I can try it out?
|
I don't think the main dev team has any current plans for improvements to the keyer, so I have been working on a solution that I hope they can integrate.?
?
The CW keyer does appear to rely on the 10-20 Hz modem polling function, which I think is too slow and is the source of our issues. I am trying to implement a different approach that relies on instantaneous RPi GPIO interrupts to bypass this function. This is the approach that many of the home-brew Arduino code oscillators use, and it seems to work very well in those devices. I am not sure yet whether interrupts will cause other issues in the sBitx software, we will see.?
?
I am not a professional coder, though, so it has taken me a bit of time to read and understand how the CW keyer code works. Will try to implement something in the next couple of weeks. Will report back on this thread soon.?
|
Steve
?
The next 64 bit release from our team does include fixes for polling and key types.
?
You can follow us on our discord server for updates and be part of a testing group for the new software.
?
?
I sent you a message there.
?
?
JJ
?
?
?
|
The keyer in JJ's dev branch feels promising. Played around with it on a dummy load and it seems to key correctly at 25 wpm. Planning to activate with it over the weekend and will report back.?
|
Would the keyer ever be able to handle the Ultimatic Mode rather than Iambic Mode. Or will I have to use my external ?keyer if I want the Ultimatic Mode?
|
I am new to this forum. I do a lot of reading of other forums especially on the SDR Radios and in particular to CW operation. The one common thing I am finding is as the speed exceeds about 15WPM the built in Keyers are not able to keep up. who would have thought that an old mode such as CW requires more computer time transom of the more advanced digital means of communications.
This explains why the CW Decoders have some difficulty as the noise increase and the speed increases above about 15WPM. Again proving the human ear connected to a brain wired (years of training) for CW is still faster and better at copying CW, than a current day processor.
|
Maybe this could be the solution:
--
Gyula HA3HZ
|
But even if you move to an external keyer, doesn't the sbitx still need to do the polling of the key connector to see when the keyer (instead of your fist) is pressing the key?
--
Mike KB2ML
|
An external keyer doesn’t solve the problem in the afarhan/sbitx main branch, aka the “stock” software. The radio still won’t accurately reproduce the correct keying timing at higher speeds.?
?
There’s an improvement to the CW keyer on GitHub in the drexjj/sbitx repository in the dev branch.?
?
I had intended to test it more fully last weekend, but unfortunately have had to deal with a personal matter that has kept me away from home and the sbitx.?
?
I do plan to circle back to this group once I’ve had a chance to put JJ’s software through its paces.?
Of course, anyone else is welcome try it as well. Keep in mind JJ’s software is 64-bit, but the stock sbitx OS is only 32-bit. There is a “how to” for installing JJ’s version I recommend following in his repository.?
?
?
|
And what about converting to the Newer Faster Raspberry Pi 5 running at 64 bits as well. From reports on just the differences between the RPI 4 and the RPI 5 should have a positive effect on the overall performance of the SBitX. However when I also read where other SDR run into similar issues with CW above 15 WPM and some have had improved performance using external keyers, is why I suggested giving this a try on the SBitX I cannot as I have not saved up enough to afford my own SBitX at the moment, takes time when you live in a fixed pension income. Plus at 83 reading is much easier than physically doing, so to speak. And I do not have any grand kids to do it for me as well. All this aside the SBitX concept offers the amateur radio operators who want a device they can actually open up and play with a much better option than the neatly packaged but tightly buttoned up commercial options. Ashhar has left many open doors in to the physical hardware and software allowing for so many possibilities. for me time and agility and ability are limiting factors. and as to keying and polling this is where pure CW transmitter as opposed to a digitally derived SSB transmitter has it advantages. This is a compromise that has to be made if one wants an all mode Transceiver. I for one do not want to turn back the clock. I have done my time lugging tube transmitters and tube receivers to field day. Even my power supply weighs more than my preferred transceiver. The SBitX concept offers so much not just for portable operation but for the average apartment dweller the SBitX remote operation built right into the design opens up the entire RF world to amateurs who have no hope of having and antenna farm at their residence. But perhaps on the roof of their apartment building with a battery supplied radio at the base, who knows. I do not mean to cause discord by any suggestions rather try a different approach to find if the first door is locked try the next door so to speak, we are Amateur Radio "Experimental" Service. Thanks for Litening
|
Jens/HB9JOI
Dan Romanchik KB6NU
VE6LX
HA3HZ