CWX: CW PTT not sent to external amplifier

[lb3rerag-hub]

Report preparation

• I used the AI-assisted bug report tool (Help → Support → File an Issue)
• I have attached a support bundle or log file

What happened?

I try to send with CWX CW on macro, then the TX1 PTT is not engaged on CW ?
its working in Phone..

What did you expect?

No response

Steps to reproduce

1. I tried to QSX on and off.
2. I tried to set CW PTT on DTR and RTS.. nothing happen?

AetherSDR version

26.5.2.1

Radio model & firmware

Flex 6700 4.1.5

Operating system

Windows

OS version and hardware

Widows 11

[aethersdr-agent]

Thanks for the report. Tracing the code path:

What AetherSDR does when you fire a CWX macro:

CwxModel::sendMacro() (src/models/CwxModel.cpp:36) emits a single command, cwx macro send N, and CwxModel::send() emits cwx send "..." <block>. That's it — AetherSDR never explicitly engages MOX or sends xmit 1 for CWX. It relies entirely on the radio to auto-engage TX, which is gated by two radio-side settings:

1. cw synccwx — when 1, the radio auto-engages MOX for the duration of the CWX queue. When 0, MOX is not keyed and the CWX queue just plays out silently with no RF.
2. cw break_in (QSK) — when 1, semi/full break-in keys TX around each CW element; when 0, the operator must manually key PTT for any CW transmission (paddle, keyboard, or CWX).

If both are 0 on your radio, CWX will queue but the radio will never go to TRANSMITTING — which is exactly what you're seeing (TX1 RCA PTT-out and the DTR/RTS serial PTT lines on the USB cable only fire when the radio's interlock state is actually TRANSMITTING).

The Phone vs CW asymmetry fits this too: in Phone you have explicit MOX/PTT engaging TX before audio plays. In CW, the radio's CWX behavior is determined by synccwx + break_in. The cwx qsk_enabled toggle in the CWX panel's Setup view is not the same as cw break_in — it's a CWX-internal flag and won't unstick this on its own. Per FlexLib Radio.cs:9693 and CWX.cs, the gating settings are cw break_in and cw synccwx.

(Worth noting: commit 42246ac1 briefly forced cw synccwx 1 when the CWX panel opened, but it was reverted in 0f0547be the same day. So currently nothing in AetherSDR pushes synccwx to the radio — it inherits whatever state the radio came up with.)

To confirm the root cause, can you check three things:

1. Radio Setup → TX tab → "CWX" row → "Sync" button — is it highlighted (on)? If not, click it on and retry the macro. This sends cw synccwx 1.
2. In the Phone/CW applet, on a CW slice, the "Break-In" (QSK) button — is it on? Try with it explicitly on.
3. When you press a CWX macro, does the radio's UI show any TX activity — does the TX slice flicker to TRANSMITTING, does the RxApplet show a TX indicator, does the TX meter move at all? Or is the radio completely idle (RECEIVE/READY) the whole time?

The answers will tell us whether (a) it's a radio-state issue we can fix by re-introducing cw synccwx 1 on connect or panel-open, or (b) there's a deeper bug where the radio is going to TRANSMITTING but TX1/DTR/RTS aren't following.

[lb3rerag-hub]

Hi To confirm the root cause, can you check three things: Radio Setup → TX tab → "CWX" row → "Sync" button — is it highlighted (on)? If not, click it on and retry the macro. This sends cw synccwx 1. In the Phone/CW applet, on a CW slice, the "Break-In" (QSK) button — is it on? Try with it explicitly on. When you press a CWX macro, does the radio's UI show any TX activity — does the TX slice flicker to TRANSMITTING, does the RxApplet show a TX indicator, does the TX meter move at all? Or is the radio completely idle (RECEIVE/READY) the whole time? 1 yes 2 yes 3 yes But the 40ms on TX1 RCA trips my amplifier as there no PTT, but RF is engaged. I try to push on MOX on, then push CWX command to send CW. Then all is ok :) It works ok if i use external logger with Winkey to toggle CW input . Best Regards, Stein-Roar Zamora Brobakken ***@***.*** LB3RE LC0X K3RAG Join me on Slack --https://join.slack.com/t/lb3re/shared_invite/zt-rq7klws5-UEVdjQql2rH0q6bhRFo8dQ

…

19. mai 2026 kl. 20:04 skrev aethersdr-agent[bot] ***@***.***>: To confirm the root cause, can you check three things: Radio Setup → TX tab → "CWX" row → "Sync" button — is it highlighted (on)? If not, click it on and retry the macro. This sends cw synccwx 1. In the Phone/CW applet, on a CW slice, the "Break-In" (QSK) button — is it on? Try with it explicitly on. When you press a CWX macro, does the radio's UI show any TX activity — does the TX slice flicker to TRANSMITTING, does the RxApplet show a TX indicator, does the TX meter move at all? Or is the radio completely idle (RECEIVE/READY) the whole time?

[aethersdr-agent]

Thanks — that confirms the picture. Handing off to a maintainer.

Summary for maintainer:

• Env: Flex 6700 / fw 4.1.5, AetherSDR 26.5.2.1, Windows 11
• Repro: With CWX Sync on and CW Break-In on, fire any CWX macro on a CW slice. Radio enters TX and RF is produced, but TX1 RCA only pulses ~40 ms (no usable PTT lead-in) — amplifier mis-trips with RF present and no PTT. DTR/RTS serial PTT lines behave the same. Phone is fine; manually pressing MOX → then triggering the CWX macro is fine; external logger keying via WinKey is fine.
• Root cause hypothesis: Radio-side TX1/serial PTT sequencer behaves differently for synccwx-auto-engaged TX vs. operator-engaged MOX. AetherSDR's CWX path (src/models/CwxModel.cpp:36 sendMacro() and :17 send()) emits only the cwx … command and relies on the radio to bring TX up — which on this firmware doesn't produce the same TX1/RCA assert-before-RF that MOX does.
• Suggested change: Make AetherSDR mirror the working manual workflow — wrap CWX transmissions with explicit MOX:
  • On sendMacro/send/sendChar: emit xmit 1 (or set interlock to TX) before the cwx … command.
  • Drop MOX after the macro completes (use the existing transmissionRequested / local-keyer completion path, or hand off to the existing CW-trailing-delay logic so the tail still observes cw break_in_delay / cwx delay).
  • Gate the behaviour so it doesn't double-key when the operator already has MOX latched, and skip when QSK is providing per-element keying that the amp can follow.
• Historical note: Commit 42246ac1 pushed cw synccwx 1 on panel open and was reverted in 0f0547be; neither addresses sequencer/PTT timing — the proposed fix is orthogonal and operates per-transmission.

Marking maintainer-review.

[jensenpat]

@AetherClaude why was this change reverted in 0f0547b? was this due to a timer issue in staging PTT and then the cwx string?

[aethersdr-agent]

@jensenpat — checked the git history. The revert in 0f0547be was not PTT/timing related; it was orthogonal cleanup after a pivot.

Original intent of 42246ac1: push cw synccwx 1 on CWX panel open so the radio would mix CW sidetone into the audio stream. That was the only reason it was added.

Why it was reverted (3 hours later): commit 770afc93 discovered via pcap that the radio never mixes CW sidetone into remote_audio_rx at all — sidetone only goes to physical outputs / DAX, regardless of synccwx. So the team switched to client-side sidetone generation in AudioEngine (Morse table + sine + envelope, triggered by cwx sent=N status frames). With client-side sidetone landed, the synccwx panel-open push was vestigial — it was sending a radio-state change that no longer served any AetherSDR purpose, so it (and the sibling cw sidetone 1 push in 1cf604e1) were reverted as dead code.

So the revert message hides what's really a "we found a better way" pivot, not a regression. There's no historical evidence of a staging/PTT timing issue with synccwx 1 itself.

Bearing on LB3RE's issue: the proposed fix in the maintainer summary is per-transmission xmit 1 wrap (not panel-open synccwx), so it's a different lever entirely. LB3RE already has synccwx on at the radio and TX engages — his problem is the radio's TX1 RCA / serial-PTT sequencer not asserting cleanly when the radio brings TX up via synccwx vs. when MOX brings TX up. That's a sequencer asymmetry on the 6700's side, and the only thing AetherSDR can do client-side is mimic the MOX path by asserting xmit 1 before the cwx … command (and releasing on queue drain, which src/models/CwxModel.cpp already has hooks for via transmissionRequested / queue-empty).

One thing worth verifying before landing it: whether xmit 1 followed immediately by cwx send "…" gives the radio enough lead-in to assert TX1 before RF — i.e., we may need a small (~10–20 ms) delay between xmit 1 and the cwx … command, or it may need to wait for the interlock state=TRANSMITTING status frame before queuing the macro. WinKey works for LB3RE because it keys CW via the radio's paddle/keyer path, which goes through the normal break-in PTT sequencer — different from CWX. That's the timing question worth nailing down in implementation rather than hoping the radio sequences it for us.

A maintainer can authorize the orchestrator to land this fix by adding the aetherclaude-eligible label.

73, Jeremy KK7GWY & Claude (AI dev partner)