CW issue with first element in transmit.

[Dan Habecker]

In my setup with a Raspberry 5 and piHPSDR I get no output in CW for 50ms.  A short dit gets completely cut off.  First I thought it was related to a buffering issue and made an external CW keyer with its own sidetone.  Still the same problem.

Next I thought maybe it was an RX to TX turnover time issue, so I added a PTT output to the keyer with a 50ms delay.  That only caused that output to start at 100ms with the first element being cut off.

One thing I am doing differently that the standard Raspberry 5 install package is I'm using PulseAudio instead of ALSA so I can get the stereo output for two RX slices.  (Very handy for working split.)  According to the piHPSDR manual this will only add approximently 10ms of delay.  Nothing about chopping off the first element.  Also I can deal with the delay with the external keyer.

Are there any settings I'm missing or a differnt way to do the CW that might help deal with this issue?

73,

Dan W9EQ

[pa3gsb]

Hi Dan,

You are an fast CW operator.

What kind of configuration are you using, do you use the juice board ? 

How is your keyer connected?

Do you use the connection of the https://github.com/pa3gsb/Radioberry-2.x/wiki/Radioberry-In-Output pin 14? or are you using the pihpsdr cw config?

 

Please give some more insight in your configuration and some software settings.

73 Johan

PA3GSB

Op dinsdag 14 juli 2026 om 15:35:27 UTC+2 schreef dan.ha...@gmail.com:

[Dan Habecker]

Johan,

I'm using a Raspberry 5 with your image 5.13.  I have a preamp board as a driver for a 100W PA.  Also connected to it is a Raspberry Pi Pico 2 sending MIDI knob tuning, volume, etc.  The Pico also listens to CAT output and adjusts a low pass and band pass board accordingly.

I have the piHPSDR set to 'use radio for CW'.  They keyer is connected through the preamp board which is routed to pin 14.  I also tried a PTT connection from the keyer to pin 13.  That is, it is set up for pin 13 to go low 50ms before keying starts on pin 14.

Like I mentioned before I'm using PulseAudio instead of ALSA so I can get the stereo output for two RX slices.  This should cause a little more buffering delay, but not a cut off of the beginning.

Thanks for you time and 73,

Dan W9EQ

[Dan Habecker]

Here are some screen shots of what the RF output looks like vs keying and PTT input.  This time I didn't get truncated elements.  They were either delayed by the 50ms or not at all.  A single dit always produced nothing.  Two dits either produced the delayed output or nothing at all.  It seemed kind of random whether I got output or not.

73,

Dan W9EQ

[Dan Habecker]

I played a little bit more with it today and saw this strange behavior.  When doing something like sending 'K' it all gets run in to one continous RF on.  LIke the first element starts late, but the second starts on time.  Here's an image of what that looks like.

[pa3gsb]

Dan,

I have done the following experiment.

Using CL025 radioberry using pi-5 with latest image.

Only radioberry no preamp board. Measuring the HF out directly from tx port of radioberry.

Created an rp2040 code to generate ptt signal and 4 cw dot using 24wpm. Repeating this in a loop after each 2 seconds again.

I was not able to reproduce the problem with a small dot in rf signal.

But i have looked through the gateware code and found an improvement which i present in the pictures below.

This is the behaviour when only using the key in pin 14 rb connector and no ptt on pin 13.

The yellow trace is the output of the radioberry hf out (via 10db att). Output looks fine no shorter time for the first dot.

I did a small modification in the gateware and when starting keying also the ptt must be set; the ptt signal is applied to the ptt pin 13 rb connector.

Giving a direct ouput when doing the keying. I think having a side tone at your keyer 

Very nice i think.

Which FPGA are you using i can send you the gateware and you are able to test again. 

Hope this helps.

73 Johan

PA3GSB 

Op donderdag 16 juli 2026 om 17:13:24 UTC+2 schreef dan.ha...@gmail.com:

[Dan Habecker]

Johan,

Thanks much for investigating this issue.  I have the CL025 version.  I look forward to trying it out.

TNX and 73,

Dan W9EQ

--
You received this message because you are subscribed to a topic in the Google Groups "Radioberry" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/radioberry/MrY1kUbW4Gw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to radioberry+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/radioberry/446c0396-75f7-4836-82f4-7c1863c6fd62n%40googlegroups.com.

[pa3gsb]

Dan,

In the attachment the 75.3 version for CL025 with addition.

Besides that i added the test code and the excutable for a pico 2040.

Please copy the radioberry.rbf into /lib/firmware  and restart the pi or use :

in a command window:

1. sudo systemctl stop radioberry 
2. sudo modprobe -r radioberry
3. verify that the latest radioberry.rbf is in /lib/firmware
4. sudo modprobe radioberry
5. sudo systemctl start radioberry

Please use a limited test setup as i described ; only pi and rb2 board.

Enjoy

73 Johan

PA3GSB

Op maandag 20 juli 2026 om 04:11:33 UTC+2 schreef dan.ha...@gmail.com:

[Dan Habecker]

Johan,

I gave it a try today.  It works very nice.  Just a 3ms delay and very consistant.  Attached is what it looks like now.  Same results you have.  Thanks very much. for you excellent work.

THere is just one more things I'm trying to figure out.  In CW the TX Drive has no effect on the power level.  If I use the TUNE botton it works fine.  Do you know if that is a setting somewhere I am missing?

73 and TNX,

Dan W9EQ