|
Hi all
I have spent a lot of time on the QMX firmware. I decided I wanted to do a release rather than go too too too far with it... In fact it might not look to you like very much has happened. But in fact this step forward is major for me and very crucial in the development path.?
Initially the QMX firmware was a bodge of QDX source code and QCX source code roughly squeezed together into the same device. Sometimes the two systems fought each other, some menu items in the old QCX menu were no longer relevant on QMX, and some items in QDX no longer relevant in QMX, etc. My initial goal was met; to launch a powerful hardware platform, with a functional enough firmware to be able to operate it as a standalone CW rig (like a multiband QCX) or as a standalone digi rig (like a QDX), and with HUGE potential for all the things to be added as we go forward. BUT, to be able to move forward, even to be able to write an operating manual, I had to do a lot of work to integrate these two subsystems into a coherent whole. I hope you can understand the significance of this - the proper integration now lets me move forward with the more important stuff of fixing remaining broken functionality and of course, adding NEW functionality.?
See attached, this is how the new terminal menu looks. The main menu now consists of several categories, you can go into the category and come out again; the menu system is nested. QMX will end up with a lot more configuration parameters than QDX and QCX so it is important for them to be well-organized, categorized so that they are easy to find. I put everything to do with Configuration in the "Configuration" category, everything to do with the hardware (such as RF Sweep, AF Sweep, Image sweep, Transmitter test) in the "Hardware tests" category, PC interface stuff in "PC and CAT" and system stuff such as firmware upgrade in "System".?
In the code the entire menu structure is very cleanly constructed so that it will be very easy for me to add and re-arrange menu items. A menu item can be either an actual configuration parameter with an associated non-volatile storage location in EEPROM (with several possible data types), or a sub-menu, or an "application" screen which means the things from the QDX world like RF sweep, AF sweep, TX test.?
This is a multiple login system, with locking resource locking; there are several types of terminal which are a USB Terminal emulator login, a connection from one of the three USART serial ports, a CAT connection, and a "UI" connection (User Interface) which means the knobs and buttons on the actual QMX front panel itself. The menu system adapts to the type of connection that it is serving. So if it is a CAT connection (from WSJT-X) then it only handles CAT commands. A terminal emulator/serial port connection gets the full menu screen (see attached); and a UI connection (knobs and buttons) gets a view of the "Configuration" category which is adapted to the small screen and limited controls, with a look-and-feel similar to the QCX.?
Therefore the whole menu system on both the terminal login or on the knobs and buttons is now consistent, behaves the same way (taking into account adaption for the different presentation full-screen vs LCD/knobs/buttons) and it is all driven from the same place in the code. Integration! The terminal emulator is efficient (responsive) and so is the menu system when you access it on the QMX itself. I also removed all the nonsensical configuration parameters and functions that belong to QDX and QCX but have no corresponding relevance on QMX.?
I can't overstate?how exciting this is for me because now the code structure is really very much cleaner, and I will be able to move forward faster and start fixing the remaining issues and implementing all my new functionality dreams.?
The only other two things I did, that are important functionality-wise, are:
1) I fixed the CW decoder - like in QCX, you can separately enable/disable it on Receive, on Transmit (decodes what you are keying in) and for editing (lets you conveniently key in message memory entries).?
2) I removed the Tx/Rx Param 1/2 from the Band Configuration screens, and hard code them to 0 and 1 respectively, because there isn't any reason they need to be changed.?
So let me know - any comments!
73 Hans G0UPL
|
Nice! ?Great to be able to refactor your code and clean up all the technical dept!
73, Willie N1JBJ
toggle quoted message
Show quoted text
On Jul 19, 2023, at 9:51 AM, Hans Summers <hans.summers@...> wrote:
Hi all
I have spent a lot of time on the QMX firmware. I decided I wanted to do a release rather than go too too too far with it... In fact it might not look to you like very much has happened. But in fact this step forward is major for me and very crucial in the development path.?
Initially the QMX firmware was a bodge of QDX source code and QCX source code roughly squeezed together into the same device. Sometimes the two systems fought each other, some menu items in the old QCX menu were no longer relevant on QMX, and some items in QDX no longer relevant in QMX, etc. My initial goal was met; to launch a powerful hardware platform, with a functional enough firmware to be able to operate it as a standalone CW rig (like a multiband QCX) or as a standalone digi rig (like a QDX), and with HUGE potential for all the things to be added as we go forward. BUT, to be able to move forward, even to be able to write an operating manual, I had to do a lot of work to integrate these two subsystems into a coherent whole. I hope you can understand the significance of this - the proper integration now lets me move forward with the more important stuff of fixing remaining broken functionality and of course, adding NEW functionality.?
See attached, this is how the new terminal menu looks. The main menu now consists of several categories, you can go into the category and come out again; the menu system is nested. QMX will end up with a lot more configuration parameters than QDX and QCX so it is important for them to be well-organized, categorized so that they are easy to find. I put everything to do with Configuration in the "Configuration" category, everything to do with the hardware (such as RF Sweep, AF Sweep, Image sweep, Transmitter test) in the "Hardware tests" category, PC interface stuff in "PC and CAT" and system stuff such as firmware upgrade in "System".?
In the code the entire menu structure is very cleanly constructed so that it will be very easy for me to add and re-arrange menu items. A menu item can be either an actual configuration parameter with an associated non-volatile storage location in EEPROM (with several possible data types), or a sub-menu, or an "application" screen which means the things from the QDX world like RF sweep, AF sweep, TX test.?
This is a multiple login system, with locking resource locking; there are several types of terminal which are a USB Terminal emulator login, a connection from one of the three USART serial ports, a CAT connection, and a "UI" connection (User Interface) which means the knobs and buttons on the actual QMX front panel itself. The menu system adapts to the type of connection that it is serving. So if it is a CAT connection (from WSJT-X) then it only handles CAT commands. A terminal emulator/serial port connection gets the full menu screen (see attached); and a UI connection (knobs and buttons) gets a view of the "Configuration" category which is adapted to the small screen and limited controls, with a look-and-feel similar to the QCX.?
Therefore the whole menu system on both the terminal login or on the knobs and buttons is now consistent, behaves the same way (taking into account adaption for the different presentation full-screen vs LCD/knobs/buttons) and it is all driven from the same place in the code. Integration! The terminal emulator is efficient (responsive) and so is the menu system when you access it on the QMX itself. I also removed all the nonsensical configuration parameters and functions that belong to QDX and QCX but have no corresponding relevance on QMX.?
I can't overstate?how exciting this is for me because now the code structure is really very much cleaner, and I will be able to move forward faster and start fixing the remaining issues and implementing all my new functionality dreams.?
The only other two things I did, that are important functionality-wise, are:
1) I fixed the CW decoder - like in QCX, you can separately enable/disable it on Receive, on Transmit (decodes what you are keying in) and for editing (lets you conveniently key in message memory entries).?
2) I removed the Tx/Rx Param 1/2 from the Band Configuration screens, and hard code them to 0 and 1 respectively, because there isn't any reason they need to be changed.?
So let me know - any comments!
73 Hans G0UPL
|
Fantastic, Hans.? I agree, it's hard to continue work when what you are working on is all a mess.? I'm sure organizing everything was a daunting task but hopefully you have set yourself up for success in the coming firmware upgrades.
Next is organizing the lab bench?
Joshua
|
Upgraded from 002 to 004 and now wsjtx will not tune for more than a second and transmit will not work for more than a second. Reverting back to 002 and tx works normally.? did a factory reset in 004 without joy.
What new parameter do I need to change in 004 to make wsjtx work again?? I tried a few things without luck.
Dave VE3KCL?
|
On 7/19/2023 3:02 PM, VE3KCL wrote: Upgraded from 002 to 004 and now wsjtx will not tune for more than a second and transmit will not work for more than a second. Observing the same thing here with 004. 73 de K9ID
|
I can confirm a new problem with FW 004 and fldigi. If using TxID the QMX reverts to receive in a second or two. The QMX has had an issue with Olivia mode with firmware 003 and before, but this is a much more limiting issue as it affects all of fldigi's digital modes.
Reverting back to 003 solves the issue. I really liked the new menu setups, but I can't use it for now.
Wsjtx works FB for me on 004.
toggle quoted message
Show quoted text
On Jul 19, 2023, at 15:02, VE3KCL < airdry@...> wrote:
Upgraded from 002 to 004 and now wsjtx will not tune for more than a second and transmit will not work for more than a second. Reverting back to 002 and tx works normally.? did a factory reset in 004 without joy.
What new parameter do I need to change in 004 to make wsjtx work again?? I tried a few things without luck.
Dave VE3KCL?
|
Hans,
Thank you! That is a LOT of work done in quick time on top of all the rest that you have going on. I really appreciate the frequent updates.
That said, yes, you do have a few more things going on yet.
* CW is still not usable on -004 now because the receiver is listening 1.5 kHz lower than it's transmitting. CW-R is now not doing anything that I notice: Both transmit and receive frequencies do not change (aside from that fixed 1.5 kHz offset) and the sideband does not change. All bands.
* The receive BPF does not switch to a newly selected band until you toggle the tuning but the LPF filter now seems to be recognizing the shift (Thanks!).
* Attempting to change bands after a filter sweep if you start on 40m seems to lock it up still however now I can work it all fine if I start at 30m. (Getting there!)
Much better behaved now.
I shall do more exploring and poking about now. Fun stuff (for me, at least)
73, Don
|
Congrats on the FW progress!
My feedback is that the FW update completed and the QMX rebooted and displayed 1_00_0004 on the screen.
But the QMX was in a "hang" state and could? not be powered off using the power button.
Windows was beeping and complaining about a bad USB device.
After cycling power, all appears well.
More testing later...
73, Mike KK7ER
|
I can confirm Cliff AE5ZA observations regarding 1_00_004 firmware and digi modes on my QMX (#133). WSJT-X loads up connects and tests fine, even passing PTT test. But when going to FT8 and trying to make a contact as soon as it switches to transmit it switches back into receive mode making 2 way contacts impossible. Reloading 1_00_003 eliminates the FT8 problem and functions normally.
And yes, the organization of the menus and ease of use is much appreciated. 73, Sandy KB3EOF
|
Ok ok ok so 1_00_004 doesn't work much does it ;-)? Look forward to an imminent fix TODAY...
On Thu, Jul 20, 2023 at 4:49?AM Sandy < mac3iii@...> wrote: I can confirm Cliff AE5ZA observations regarding 1_00_004 firmware and digi modes on my QMX (#133). WSJT-X loads up connects and tests fine, even passing PTT test. But when going to FT8 and trying to make a contact as soon as it switches to transmit it switches back into receive mode making 2 way contacts impossible. Reloading 1_00_003 eliminates the FT8 problem and functions normally.
And yes, the organization of the menus and ease of use is much appreciated. 73, Sandy KB3EOF
|
Glad it wasn't just me that saw this issue. ?Just part pf the growing pains. ?
Curt
WU3U
|
Hi all
I fixed it, see version 1_00_005 see
A single little typo :-/? ?After fixing that, testing on 20m with WSJT-X in FT8 mode, QSO with SP8NFO immediately on the first CQ.
Dan... I wasn't able to reproduce the other things you mentioned but I think that at least some may have been side-effects of the main issue.? ? * CW is still not usable on -004 now because the receiver is listening 1.5 kHz lower than it's transmitting. CW-R is now not doing anything that I notice: Both transmit and receive frequencies do not change (aside from that fixed 1.5 kHz offset) and the sideband does not change. All bands.
I was not able to reproduce this. I have a -60dBm calibration signal source that is on 7,003,513, measured on a frequency counter. I tune to 7.0035 and I hear it. LOUDLY. Then key (straight key mode) into dummy load with the frequency counter attached - it measured 7,0035 too...?
Remember that you can only use CW when the radio is in CW mode. And only use Digi when the radio is in Digi mode. And yes I need to think of the nicest, least intrusive way to use a little piece of the limited screen real estate to indicate to the operator which modulation mode he is in.? ? * The receive BPF does not switch to a newly selected band until you toggle the tuning but the LPF filter now seems to be recognizing the shift (Thanks!).
I couldn't reproduce this either. Actually on studying the code again I could not even see how such a thing as setting the LPF but not the BPF could even be possible. They both get setup in the same place in the code, the same function. So... ARE YOU SURE? ? * Attempting to change bands after a filter sweep if you start on 40m seems to lock it up still however now I can work it all fine if I start at 30m. (Getting there!)
Now this is quite possible - I need to check that filter sweeps put the radio back in the correct state (previous state) at the end of the sweep. Quite possibly something is a little off, here.?
73 Hans G0UPL
?
|
Hans,
If I could add something regarding the filter sweeps. Here is has started the sweep and stops somewhere about 30% in some cases or 18% in another case. Been doing this for some time. Only on 40M and 60M presently.
toggle quoted message
Show quoted text
Hi all
I fixed it, see version 1_00_005 see
A single little typo :-/? ?After fixing that, testing on 20m with WSJT-X in FT8 mode, QSO with SP8NFO immediately on the first CQ.
Dan... I wasn't able to reproduce the other things you mentioned but I think that at least some may have been side-effects of the main issue.? ? * CW is still not usable on -004 now because the receiver is listening 1.5 kHz lower than it's transmitting. CW-R is now not doing anything that I notice: Both transmit and receive frequencies do not change (aside from that fixed 1.5 kHz offset) and the sideband does not change. All bands.
I was not able to reproduce this. I have a -60dBm calibration signal source that is on 7,003,513, measured on a frequency counter. I tune to 7.0035 and I hear it. LOUDLY. Then key (straight key mode) into dummy load with the frequency counter attached - it measured 7,0035 too...?
Remember that you can only use CW when the radio is in CW mode. And only use Digi when the radio is in Digi mode. And yes I need to think of the nicest, least intrusive way to use a little piece of the limited screen real estate to indicate to the operator which modulation mode he is in.? ? * The receive BPF does not switch to a newly selected band until you toggle the tuning but the LPF filter now seems to be recognizing the shift (Thanks!).
I couldn't reproduce this either. Actually on studying the code again I could not even see how such a thing as setting the LPF but not the BPF could even be possible. They both get setup in the same place in the code, the same function. So... ARE YOU SURE? ? * Attempting to change bands after a filter sweep if you start on 40m seems to lock it up still however now I can work it all fine if I start at 30m. (Getting there!)
Now this is quite possible - I need to check that filter sweeps put the radio back in the correct state (previous state) at the end of the sweep. Quite possibly something is a little off, here.?
73 Hans G0UPL
?
|
The receive offset seems to have been caused by the sideband option in the SYSTEM CONFIGURATION terminal entry. I had changed it to LSB to try to fix a problem and had changed it back. It showed USB after the change (and re-entry and re-power afterward) but it still occurred until 3 or 4 power cycles later when it suddenly began working properly. I checked the configuration screen and it remains USB but the frequency is again proper. I tried the experiment again to confirm and, Yes, it causes that problem. It takes several power cycles to make it recognize the change back however.
The CW-R function does not do anything at all at the time (and completely unimportant) and is the reason that I tried the aforementioned option change.
It may not be the RX BPF, but it certainly appears so. When changing bands now the transmit power is normal but the receive sensitivity is more than 20 dB down. It requires nearly a microvolt to be heard at all. Jogging the tuning up or down a click and returning it restores the normal sensitivity every time.
Yes, for some reason 40m band change is problematic for both the double-click band change or changing bands in the RF sweep on the terminal. Easily overcome by using presets or, in the case of the terminal, physically setting the radio to another band before doing the test. Just a slight annoyance.
Changing side tone frequencies doesn't seem to work.
The S-meter works very smoothly but the maximum increment is 999 and still is full scale at 3 uV (-97 dBm) on 40m. Quite generous. And very unimportant.
Keying is smooth with even the battery monitor on. Well done!
Don't burn yourself out on this stuff, Hans. We appreciated the hard work but you need some time to yourself. I only bring this stuff up to your attention because you might be interested. I am NOT complaining by any means! This is a great project but it is just a fun thing and not serious whatsoever.
73, Don
|
Thank you Hans!!! 005 solved the TxID issue. The problem with fldigi and Olivia mode going back to receive uncommanded still exists, but doing a Tune before using it or using another mode first allows it to transmit.
toggle quoted message
Show quoted text
Hi all
I fixed it, see version 1_00_005 see
A single little typo :-/? ?After fixing that, testing on 20m with WSJT-X in FT8 mode, QSO with SP8NFO immediately on the first CQ.
Dan... I wasn't able to reproduce the other things you mentioned but I think that at least some may have been side-effects of the main issue.? ? * CW is still not usable on -004 now because the receiver is listening 1.5 kHz lower than it's transmitting. CW-R is now not doing anything that I notice: Both transmit and receive frequencies do not change (aside from that fixed 1.5 kHz offset) and the sideband does not change. All bands.
I was not able to reproduce this. I have a -60dBm calibration signal source that is on 7,003,513, measured on a frequency counter. I tune to 7.0035 and I hear it. LOUDLY. Then key (straight key mode) into dummy load with the frequency counter attached - it measured 7,0035 too...?
Remember that you can only use CW when the radio is in CW mode. And only use Digi when the radio is in Digi mode. And yes I need to think of the nicest, least intrusive way to use a little piece of the limited screen real estate to indicate to the operator which modulation mode he is in.? ? * The receive BPF does not switch to a newly selected band until you toggle the tuning but the LPF filter now seems to be recognizing the shift (Thanks!).
I couldn't reproduce this either. Actually on studying the code again I could not even see how such a thing as setting the LPF but not the BPF could even be possible. They both get setup in the same place in the code, the same function. So... ARE YOU SURE? ? * Attempting to change bands after a filter sweep if you start on 40m seems to lock it up still however now I can work it all fine if I start at 30m. (Getting there!)
Now this is quite possible - I need to check that filter sweeps put the radio back in the correct state (previous state) at the end of the sweep. Quite possibly something is a little off, here.?
73 Hans G0UPL
?
|
I made CW and FT8 contacts on 20m last night to confirm that firmware 005 is working for me.
I did hit the known issue (several times) where the rig stops responding to the straight key until you remove power and reboot.
Also, CW messages don't yet work (I admit I'm lazy and like to call CQ using the predefined messages).
FT8CN immediately recognized the rig as a QMX and worked as expected on my Android phone..
I'm seeing 2.5W out on 20m now (in previous sessions I was at 3.5W) -- I'll need to look into that.
One weird thing:? I tuned around on 40m just looking for CW signals (using 20m dipole) and the background noise rose and fell substantially as I tuned across the CW portion of the band (as though the passband of the BPF/LPF had notches) -- I'll need to look into that as well.
FYI.
73, Mike KK7ER
|
Yeah..."ME TOO" on asking that the CW messages/memories be the next item to "enable".
es vy 73 om de "baab" w9ya
|
On 7/20/2023 7:37 AM, Hans Summers wrote: I fixed it, see version 1_00_005 see Thank you! It's working great now. One question: I know that CW-R is not yet implemented. Is it's purpose to move CW to the opposite sideband? 73 de K9ID -- Ron Kritzman DB Masters Software Tel: 847-668-5779
|
Yes Ron. It is a great tool for dodging QRM or spurious signals ("birdies").
73, Don
|
Thought so. Actually I prefer it because my other radios, mostly older Icoms, have CW on the LSB. Turning the dial clockwise raises the pitch of the incoming station. On the QMX or any rig where CW is on the upper side, clockwise lowers the pitch. When I go to move frequency I instinctively turn the dial the wrong way. I suppose that people who are used to CW-U would complain if it was the other way, which is why a selection is nice.?
|
Joshua KJ7LVZ
VE3KCL
Don, ND6T