Re: QMX possible to replace dead MCU?
Gordon,
The boot loader is encrypted so you can’t simply replace the MCU. Jeff Moore, W1NC, can assist you with this if you can’t get a new programmed MCU from QRP Labs directly.
Tony? On Fri, Jul 19, 2024 at 5:03?PM Gordon Young via <gordon.j.young= [email protected]> wrote:
Is it even possible to replace a dead STM32/MCU myself on the QMX?
I read that the boot-loader has a firmware decryption key.
I believe my QMX has a dead MCU. I can handle SMD replacement.
Q: I'm not sure I can install the boot-loader: I suspect it's not available to me?
FWIW:? I do very much appreciate the fact that Hans is ensuring the legitimacy of QRP Labs devices (preventing Chinesium QMX's from hitting the market via firmware encryption)
73s!
~Gordon
|
QMX possible to replace dead MCU?
Is it even possible to replace a dead STM32/MCU myself on the QMX?
I read that the boot-loader has a firmware decryption key.
I believe my QMX has a dead MCU. I can handle SMD replacement.
Q: I'm not sure I can install the boot-loader: I suspect it's not available to me?
FWIW:? I do very much appreciate the fact that Hans is ensuring the legitimacy of QRP Labs devices (preventing Chinesium QMX's from hitting the market via firmware encryption)
73s!
~Gordon
|
Re: Multi-Banding 50W Amp
Sorry, the voltages are too high for the capacitors and the currents too high for the small toroids.
?
From the manual: ?There are two important differences, both are related to power handling. Here, the amplifier
produces over 50W. That requires:
? Powdered iron toroids which have enough iron in them that they will not saturate and
become lossy, at 50W
? NP0/C0G capacitors which have a high enough voltage rating to withstand the peak voltage
excursions seen at 50W.
50W is a 141 V peak-peak sinewave.
|
...and a computer is not much bigger than a keyboard
George, if you just want to send CW to QMX from a computer, I have a workaround. As Robert Giuliano KB8RCO mentioned in your original thread,
?
It seems to me that "the timing" is what the "keyer" is all about, and most radios today already have a keyer built in.?
Would it be any easier to use 2 pins and code the keyer outputs (left paddle/right paddle) for ASCII characters ?than time the actual dits and dahs?
I have an Arduino program does exactly that. Basically it's a keyer that emulates an iambic paddle (instead of a straight key). It takes an ASCII string from computer via the serial port and send the CW symbols via 2 pins - dit and dah, to any radio's paddle input. The only requirement is that the WPM setting of the radio's internal keyer must be consistent with this program. If you can't wait for Hans to implement the KY command for QMX, maybe you can try my approach. BTW you can use diodes to 'OR' the keyer's outputs with a normal paddle's tip and ring, so the radio can be keyed from both sides.
?
73 de Zhenxing N6HAN
|
Re: Digital mode audio to headphone jack
JTDX on the Mac is quite buggy and often will stop sending audio out to the USB
port.
Corey, what Mac are you using? I know WSJTx on some old Macs have similar issues. When I was using WSJTx on my 2010 Mac mini with IC705, the audio stream sometimes would be routed to the internal speaker. The workaround is surprisingly simple - just change the sample rate of the radio's virtual sound card from 48kHz to 44.1 kHz. It can be done in "Audio MIDI Setup". Don't know if this is relevant though.
|
I have a QRPLabs 50W amp kit in the process of construction. I would like to use it with my QMX+ but would like to be able to switch bands on it. I was thinking of just bypassing the LPF in the amp and feeding it into an external LPF that can be switched to different bands either manually or via a signal from the QMX+. Looking at the schematic, I would jump from the C8/transformer junction to C15, omitting the LPF components. I would mostly use 40M to 17M. probably never higher than 17M. I realize it's class C so only for CW.
?
Would this work or is it a Bad Idea?
--
73...? ? ? Mark? ? ?AA7TA
|
Re: Digital mode audio to headphone jack
Ludwig,
?
That's it! I didn't realize that the transmit indicator actually had 3 separate conditions.
?
Thank you and 73s
--Corey (WB0RXQ)
|
Re: CAT control and compatibility with Kenwood TS-480
Thanks?Hans,
Much appreciated.
RD/RU You are right, it is five digits, it was a typo. I know for a fact RD and RU decreases and increases the RIT value on the TS-480.? Only recently (TS-890 and TS-990) Kenwood?has implemented an absolute RIT control, then using the RF command.? But for me, as a developer, compatibility with TS-480 is unimportant as long as I know exactly what RD and RU do. Thank you for the clarification. It would?be great if it was included in the documentation.?
Given your special definition of RD/RU I agree RU0; and RD0; should clear RIT.? Since RD/RU is relative on "Kenwood?like" radios, they all have a special command to clear RIT: RC; But again, as long as the behavior is known, compatibility with other radios is unimportant.
FA/FB? Thank you for the clarification. Much appreciated.
IF Thank you. The documentation currently says something else.?
Bj?rn SM7IUN
On Fri, Jul 19, 2024 at 10:43?AM Hans Summers via <hans.summers= [email protected]> wrote: I'm on the DXLog development team and struggle with implementing the drivers for the? QMX and QCX models. If I had access to them I could do reverse engineering but unfortunately I don't.?
The CAT part of the QMX/QCX documentation is not very detailed and unfortunately also? somewhat inconsistent.
Just to explain some background. QCX was created in 2017 and there was no CAT port on it; in fact the 32K Flash program memory was 100% full. But CAT was requested and I wrote it. I managed to squeeze existing functionality, to make space for the program code for the CAT features. I chose TS480 because the protocol is well documented and widely supported. I believe (based on similarities) that the Elecraft radios were also based originally off TS480 protocol but I may be wrong on that, and certainly they were later extended and now have their own entries in the list. QRP Labs also has its own radio type in recent hamlibs.?
Because of the very limited program memory and also the fact that of course QCX and TS480 don't have the same capabilities, only a subset of the TS480 protocol was implemented, and it was implemented in a less strict way compared to TS480, in order to save program space. Nevertheless it was sufficient for correct operation with many logging programs. The hardware required a few extra components to be able to share some IO pins with the LCD; later those components were incorporated into PCB revisions so it became standard.?
Later I had requests from QRP Works who wanted to make their SideKar product which had originally been an Elecraft Accessory only.?
George Korper if you are reading this - SideKar is another option for sending keyboard Morse from a QRP Labs transceiver, as you were asking about Keyboards in a different thread.
Anyway - so I added some support for the KY; command to send and retrieve decoded CW. KY is an Elecraft command not a TS480 command so was another departure from pure TS480.? ? Now some of those CAT commands were inherited into QDX CAT commands, and later QDX CAT was inherited into QMX. Unfortunately QMX doesn't yet handle KY (and perhaps other CW details) because of the inheritance route of the code modules through QCX->QDX->QMX... however this is just an anomaly and will be resolved, in fact it's even quite high on my QMX ToDo list to make KY work properly on QMX. So won't be long I hope.?
QMX has plenty of code space, it isn't so restricted as QCX was so if there are any requests, they can be accommodated (in time).?
I had thought the QCX/QMX documentation is precise and sufficiently detailed but if there are any improvements necessary please let me know/? ? Since my help desk ticket remains unanswered I am turning to this reflector.?
Apologies for the delay handling that, I'm *still* behind on emails and helpdesk tickets... at the same time trying to carve out a little time for development. Pitying?myself that there is only one me whhaaaaaahhh...? ? The challenges I face are the following:
RD/RU Is the implementation of the RD and RU commands really the same as the TS-480?? The documentation suggests otherwise. With the TS-480, RD must be followed by exactly four digits? and decreases the RIT with a certain number of hertz. RU works?the same but increases the RIT.? The QMX/QCX documentation seems to suggest RD instead sets the RIT to an absolute value and? that a sign also must be included?in the command (like with Kenwood's RF command). How does RD/RU really work with the QMX/QCX and what is the format?
I'm not sure your description of the RU RD commands on a TS-480 is quite right either. The TS-480 document is here:
RD and RU commands are described on pages 17 and 18 respectively.?
The RU description is: "Moves the RIT offset frequency up. Slows down the scan speed in Scan mode". I have no idea what scan mode is but it isn't a current QCX or QMX thing anyway so ignore that. The Set command has format RU p1 p1 p1 p1 p1 ; where P1 is the offset frequency 00000-99999 in Hz. So there are five digits (not four as you state). But my interpretation of the TS480 specification is that this sets the RIT offset frequency in the up direction. Not that it applies a shift to the currently set RIT offset frequency. In other words I interpreted it as an absolute RIT setting, not a shift to the current setting. It seems more logical to me.?
But am I WRONG? Does someone have an actual TS-480 to check? And even if wrong - should I change QMX? (Note: QCX firmware is very stable with few if any known bugs and hasn't had an update for more than 2 years; I'd be reluctant to meddle in QCX firmware unless it is very terribly urgent).?
Attached is a screenshot of the RU command from the Kenwood document. RD is similar. Aside from the 5 digit thing which seems clear, I can't in my view, see a way to interpret this that RU increases the RIT offset. To me it seems it should specify an absolute, positive RIT offset. Since it declares that P1 is the offset frequency, and had previously referred in the description to the "RIT offset frequency". It doesn't say that P1 is a CHANGE to the RIT offset frequency.?
So this is why I implemented it this way. I did not implement the Read command because it appeared to be applicable only in Scan mode.?
To summarize the QMX implementation: - RUn; sets an absolute positive RIT of n Hz.?
- RDn; sets an absolute negative RIT of n Hz.?
- Neither RU nor RD have a Read implementation, in other words RD; is not supported.?
- Neither RU nor RD require a sign in front of them; in fact whether you put?+ or - in front of the number, it is ignored. Only the number is read, not the sign; the sign is determined by RU (+) or RD (-)
- RU and RD are not strict about requiring 5 digits. In other words you can type in RU50; and the RIT will be set to?+50 Hz. You don't need to send RU00050; which would be a strict TS480 compliance; but if you do send RU00050; it will also work fine.?
As a result of typing out this reply and my experiments on an actual QMX+ here I have found the following anomalies and added it to my Dev list to correct: - RU0; and RD0; don't set RIT to zero (switching OFF RIT), they do nothing. I think they should set RIT to zero.?
- RT (sets or reads the RIT function status) only has a Read function, RT; returns the RIT on/off status. If I were to implement the Set function then this would be a way to switch RIT off. Currently there is NO way to set RIT off or to set it to zero from the CAT interface.?
- I think setting a RU, FA etc command should also move the cursor on the QMX to the altered frequency field; that would make sense and be consistent with what happens if you enter RIT editing on the QMX by double tapping the left button.?
- Documentation issue: the QMX operating manual says on page 97 in its example of RD, "for example “RD-200;” sets the RIT to -200Hz"; in fact as I mentioned, the minus sign is not required and is ignored. So RD200; would be the correct command to set RIT to -200 Hz.?
FA/FB With?the?TS-480, the FA and FB commands?must be followed by exactly 11 digits.? The QMX/QCX seems to accept?a varying?length. Is this true?
Yes. As for RU and RD, with FA and FB QCX/QMX is not strict about the length of the frequency specification, it does not have to be 11 digits.?? ? IF With the TS-480, the mode field (character number 30) returns the same code as MD.? This does not seem to be the case with the QMX/QCX. Is this true?
No. On both QCX and QMX IF command character 30 is the same as the result of the MD command.
|
Re: Progrock2 for Atlas210x
Scott, in general, applications will be slightly different, but will fall into a handful of approaches.
?
For example, in my repeater application, I have oscillator modules using a FET. I will remove the crystal and the feedback in the oscillator circuit, replace the crystal with a progrock2, and use the former oscillator as a buffer. It's possible to replace both oscillators with a single progrock2, but in this case, I want to remain compatible with the original design, since the repeater can use eight selectable oscilator pairs. I have 10VDC available for power.
?
For my Atlas, I plan to use a singe progrock2 directly for three of the ten user selectable band segments, directly replacing three crystals. I have 7.5VDC power there.
?
Without seeing the inside of your radio and a schematic of the oscillator section, all I can say is your solution might resemble one of the above. Padding, filtering, and power are the things that can complicate basic replacement.
?
The datasheet for the SI5351A will show other possibilities. It's a good starting point.
?
73 Chuck AI7SY
|
Re: QMX+ Manual - BS170 - solder only 2 pins?
Very true Steve. In my organization, we were required to fill all empty vias with solder. I was never clear on why doing so improved reliability but maybe this relieved thermal/mechanical stress on the vias during the grueling temperature cycling, powered burn-in, and vibration table/drop tests we had to pass to qualify the equipment. If nothing else, it did make final visual inspection of each unit easier to perform.
On Fri, Jul 19, 2024 at 6:22?AM Steven Dick, K1RF via <sbdick= [email protected]> wrote: Often in mil applications, vias are fill with conductive or non-conductive material.? Most of the high density PCBs I worked with had vias that were filled with a conductive epoxy.? It enhanced reliability.? See link below.
-Steve K1RF
------ Original Message ------
Date 7/18/2024 8:17:50 PM
Subject Re: [QRPLabs] QMX+ Manual - BS170 - solder only 2 pins?
I grew up with MIL-STD 883B. We had to have large enough plated-through holes to guarantee that solder from one side of the board would flow smoothly through to the other side. Solder filets were visually inspected for every component on both sides of the board to ensure shiny connections. We found the hard-way that failure of the solder to flow through to the other side was usually due to the component leads being too large in comparison to the hole diameters, preventing the solder from flowing around the component leads properly and resulting in cold solder joints and/or solder blobs. I remember one project where a multilayer board with this issue was rejected by our QA people and the entire board had to be redesigned. Ouch! Many years ago the company I worked for had a contract that required all vias be completely filled with solder. This was in the 70’s and vias were a bit bigger.
Larry KB3CUF? Military standards have much to say about how much of the plated holes should be full of solder (usually 75% or more although NASA likes 100% every time )?
?
In hand assembly it’s hard to control this from the other side of the board with different heating curves etc but as I said if the joint on the reverse is a good one it should be at least part filled and fine. ?
Perfectionists might like to add a bit more solder from the transistor side then suck the new excess off the bottom side then repeat ?….. ? but you can go on like that all night ?and all the time the component’s getting cooked?
|
Re: QMX New build - no flash drive - how to troubleshoot?
Hi Ludwig,
Thank you for all the support!? No worries at all.
I'm going to further study the schematic & decide if I need to replace any components in the path of VDD or attempt to replace the STM32.
These radios are (in relative terms) inexpensive to purchase. I'm sure I'll snag another kit and do another build. Who knows maybe I can resurrect this one still!
On Fri, Jul 19, 2024 at 02:57 AM, @Ludwig_DH8WN wrote:
I don't know if an other broken element connected to VDD would affect the controller in this way or the controller have to produce this effect by itself.
|
Re: QMX+ Manual - BS170 - solder only 2 pins?
I should have added, we used via fill for via-in-pad applications for surface mount components.? In this technique, a via is connected directly to the pad.? This technique makes the signal pad much shorter compared to the conventional "dog bone" structure used with pad connection to a via for through hole applications or a conventional route to a via used in surface mount applications.? ?If the via is not placed on the pad, there is no need for via fill in surface mount applications, which adds manufacturing? complexity and additional cost.? Many of our pc board were very high speed (tens to hundreds of MHz and a few GHz) with lots of layers and we made extensive use of via-in-pads. Attached photo courtesy Sierra Circuits.
-Steve K1RF
------ Original Message ------
Date 7/19/2024 7:22:34 AM
Subject Re: [QRPLabs] QMX+ Manual - BS170 - solder only 2 pins?
Often in mil applications, vias are fill with conductive or non-conductive material.? Most of the high density PCBs I worked with had vias that were filled with a conductive epoxy.? It enhanced reliability.? See link below.
-Steve K1RF
------ Original Message ------
Date 7/18/2024 8:17:50 PM
Subject Re: [QRPLabs] QMX+ Manual - BS170 - solder only 2 pins?
I grew up with MIL-STD 883B. We had to have large enough plated-through holes to guarantee that solder from one side of the board would flow smoothly through to the other side. Solder filets were visually inspected for every component on both sides of the board to ensure shiny connections. We found the hard-way that failure of the solder to flow through to the other side was usually due to the component leads being too large in comparison to the hole diameters, preventing the solder from flowing around the component leads properly and resulting in cold solder joints and/or solder blobs. I remember one project where a multilayer board with this issue was rejected by our QA people and the entire board had to be redesigned. Ouch! Many years ago the company I worked for had a contract that required all vias be completely filled with solder. This was in the 70’s and vias were a bit bigger.
Larry KB3CUF? Military standards have much to say about how much of the plated holes should be full of solder (usually 75% or more although NASA likes 100% every time )?
?
In hand assembly it’s hard to control this from the other side of the board with different heating curves etc but as I said if the joint on the reverse is a good one it should be at least part filled and fine. ?
Perfectionists might like to add a bit more solder from the transistor side then suck the new excess off the bottom side then repeat ?….. ? but you can go on like that all night ?and all the time the component’s getting cooked?
|
Oops, I just saw that Hans answered my question on a different thread.? Great timing!
?
Hans wrote:
Now some of those CAT commands were inherited into QDX CAT commands, and later QDX CAT was inherited into QMX. Unfortunately QMX doesn't yet handle KY (and perhaps other CW details) because of the inheritance route of the code modules through QCX->QDX->QMX... however this is just an anomaly and will be resolved, in fact it's even quite high on my QMX ToDo list to make KY work properly on QMX. So won't be long I hope.?
?
Carry on.
?
73, Mike KK7ER
?
|
Hans,
?
In the meantime would you consider enabling the KY CAT command like on the QCX*?
?
73, Mike KK7ER
?
|
Re: Digital mode audio to headphone jack
On Fri, Jul 19, 2024 at 03:26 PM, William C Brown wrote:
When this happens, there is no indication that the QMX is not sending audio
out anymore,
Corey,? please check Operation manual section "Operating QMX on digital modes" subsection "Transmit status indication during QMX digital transmissions". I think it's what you are looking for.
?
73, Ludwig
|
Digital mode audio to headphone jack
Is it possible to get audio output to the headphone jack when using digital modes?
JTDX on the Mac is quite buggy and often will stop sending audio out to the USB
port. When this happens, there is no indication that the QMX is not sending audio
out anymore, even though the radio shows full power output. I used a second radio to?
verify that this was happening to me on FT8. Oh, I found that on JTDX, if I switched
the audio back to the internal MAC speakers and then back to the QMX, audio output
to the radio is restored.?
Any input here would be great appreciated.
?
--Corey (WB0RXQ)
|
One thing I always found interesting was that most of the computer interfaces output to the "straight key" port.
Winkeyer (as an example) makes a big deal out of saying the timing is taken away from the computer and in "their box" .
?
Just thinking out loud ...
It seems to me that "the timing" is what the "keyer" is all about, and most radios today already have a keyer built in.?
Would it be any easier to use 2 pins and code the keyer outputs (left paddle/right paddle) for ASCII characters? than time the actual dits and dahs?
?
I suppose in Hans' case where he is talking about making radio a keyboard "host", it wouldn't matter.?
But on an external keyboard interface ...
?
Maybe the pin timing would then be dependent on the internal radio keyer speed.
I might have to play with this a bit.
?
-------
Robert Giuliano
KB8RCO
|
|
|
Re: QMX+ finals, BS170 / TN0110 thoughts?
#qdx
#qmx
I have built up my 9V-QMX 60-15m with TN0110 - the rounded side was ground flat, the top side was provided with an additional cooling plate, LPF-coils optimized.?
The information comes from the terminal control display (current from the power supply, temperature sensor)
Operating voltage 9.6V, at the finals 8.2V
Current consumption 0.9A (17m) to 1.1A (20m)
Output between 5.2W (60m) and 5W (20m), 18m 4W, 15m 4.3W
Power loss at the finals 1.7W (60m) to 3.33W (15m) increasing. (= ((IinTX-IinRX)x8,2V)-Pout)
So far I have not seen any clear differences to the BS170
My temperature test for 50°C of the cooling surface (the permissible Id and the power dissipation drop to 80%) at 25°C room temperature:
In CW no 50°C are reached with a CQ loop.
In FT8 no 50°C are reached with CQ loop from 60-20m, but on 18MHz after about 14x CQ and on 21MHz after 7xCQ
A test with Olivia (Fldigi, passes up to 2min) on 17m was aborted immediately due to the rapid temperature increase caused by the long transmission cycles.
Test with TinySA: Lowest harmonic attenuation -48dB on 60m, otherwise at least -56dB, sensitivity between -118dB (21MHz) and -128dB (7MHz) at approx. 10dB S/N, CW
I am satisfied with my QMX, I use it mainly for CW and some FT8
73 Reiner
|
Thanks to everyone. I missed the date on those posts.?
?
————————-
Jim?
N7RCS
Palm Bay, FL
|
Gordon Young
William C Brown
Bj?rn SM7IUN
Chuck AI7SY
dl8lrz Reiner