Re: qmx with tiny lcd display
Thank you Barb, thats a really good information that i need.?
|
Re: What's the Weakest Signal You Can Get a Response From?
QRP (< 5W) is a challenge.
hamspots.net and pskreporter.info both show lots of low SNR for your transmissions.
Mike W9MDB
toggle quoted message
Show quoted text
On Sunday, September 22, 2024 at 01:00:17 PM CDT, pgramsey, N9LFF via groups.io <pgramsey@...> wrote:
When working WSJT-X, what is the weakest incoming signal you expect to get a response from?
?
I'm running a QDX and a QMX, each putting out 2-4 W depending on the band. I find that I rarely get a response from anyone unless their incoming signal is better than -10 dB. That sort of makes sense. If they're putting out 50 W I should be 11 dB or so below them, and -20 is the practical limit, so assuming reciprocal links the incoming signal should be -9 or better. Obviously there are people out there not running 50 W, but apparently not many. I've run numerous antennas and various bands and always seems about the same.
?
I've convinced myself that this is normal, but is that consistent with other people's experience using a 4 W transmitter?
|
What's the Weakest Signal You Can Get a Response From?
When working WSJT-X, what is the weakest incoming signal you expect to get a response from?
?
I'm running a QDX and a QMX, each putting out 2-4 W depending on the band. I find that I rarely get a response from anyone unless their incoming signal is better than -10 dB. That sort of makes sense. If they're putting out 50 W I should be 11 dB or so below them, and -20 is the practical limit, so assuming reciprocal links the incoming signal should be -9 or better. Obviously there are people out there not running 50 W, but apparently not many. I've run numerous antennas and various bands and always seems about the same.
?
I've convinced myself that this is normal, but is that consistent with other people's experience using a 4 W transmitter?
|
Re: QMX+ back up clock accuracy
Hello Don
700 / 1900 Hz, standard two tone frequencies. VFO on 7030 kHz (I know, CW segment nor SSB... It's only testing to the spectrum analyzer via 50 ohm attenuators).?
I'm using phase pre-distortion based on a measurement curve (not realtime feedback, which isn't available). See attached. Phase pre-distortion doesn't make much difference to IMD3 but does noticeably improve higher order intermodulation products.?
Amplitude modulation linearity is excellent and I cannot measure any distortion at all within the accuracy limitations of a simple measurement; except for the RF blowthrough from driver through the PA even when PA voltage is zero, which is about 1Vpp at the BNC antenna connector. I am able to zero that by switching off the driver completely but on/off is all the control I have over the driver (74ACT08). Switching off when the output should be below 1Vpp can be called amplitude pre-distortion of a sorr, and appears to make some improvement albeit small.?
It should be noted that the modulation method is quite different from a conventional SSB exciter and PA. A normal linear PA is most disturbed by the non-linearity distortion: a non-straight line of RF output vs RF input, caused normally by the inescapable fact that transistor characteristics are curves not straight lines, and straightening the situation out isn't easy. In the QMX PA the transistors are saturated on/off, there is no use of the transistors curvy characteristics at all. The amplitude modulation is at audio frequencies and the amplitude modulation circuit can be a lot more complex, in QMX it is made up of 4 transistors in a feedback arrangement driving a P-channel MOSFET. This is extremely linear. So the problems with a PA and SSB modulation using EER are quite different to conventional SSB and linear PAs.
In any case the QMX results so far compare very competitively with commercial black box transceivers with 20x the price tag. I reached the stage where as I tinker in ever more detail with DSP adjustments, I'm able to get a dB more of IMD3 at the expense of higher orders, or vice versa, small changes that are debatable which is more important. So it's time now to move on and connect up the microphone signals (digitally).
73 Hans G0UPL
On Sun, Sep 22, 2024, 18:59 Donald S Brant Jr via <dsbrantjr= [email protected]> wrote: Hans:? Brilliant work!? What tone frequencies are you using? The uneven IMs point to some AM->PM distortion.
Has any predistortion been applied?? Is there code space to do so?? Plans to do so???
?
73, Don N2VGU
|
Re: qmx with tiny lcd display
When going from an aluminum housing as provided by QRP Labs to a 3D printed one, I suspect that, due to the poorer thermal conductivity of the plastic/air matrix, high temperature issues may arise unless means of ventilation or heat conduction is provided.
I have seen this overheating occur in 3D printed balun housings, where the balun was properly sized but there was no/insufficient airflow and the heat buildup due to normal losses eventually caused the housing to melt and the core was visibly damaged.
Making one or more of the larger housing faces from sheet aluminum (or even thin heat sink stock) and using some thermal interface material may be useful in conducting heat out of the box.??
73, Don N2VGU
|
Re: 180pf (181) Capacitor Needed
I usually have a draft 'project' on file with Mouser so it's ready to go when I need parts...
I've not had any problem buying single quantity NP0/C0G caps from Mouser, but I usually put together a $100 CAD so I get the free shipping.
73 de va3rr
toggle quoted message
Show quoted text
On Sun, Sep 22, 2024 at 10:55 AM, Tisha Hayes, AA4HA wrote: They also do have some minimum purchase quantities on some components so you
need to watch for that as well. They are not going to sell you (1) C0G
capacitor at $0.19 each, that is not how their business model works.
|
Re: QMX+ back up clock accuracy
Hans:? Brilliant work!? What tone frequencies are you using? The uneven IMs point to some AM->PM distortion.
Has any predistortion been applied?? Is there code space to do so?? Plans to do so???
?
73, Don N2VGU
|
Re: 180pf (181) Capacitor Needed
Because of the shipping cost being way more than the component cost I have a business account with Mouser and create 'BOMs' using Mouser's web page and just drop a single order to get a few dozen spares in one fell swoop.?
?
(under [accounts and orders], [BOM's] )
?
They also do have some minimum purchase quantities on some components so you need to watch for that as well. They are not going to sell you (1) C0G capacitor at $0.19 each, that is not how their business model works. You may need to be prepared to buy quite a few things. They are not like the Radio Shack of components but are more inclined to businesses that spend hundreds or thousands of dollars on a single Purchase Order.
--
Tisha, AA4HA
|
Re: QMX+ back up clock accuracy
Tonight's build of Hamlib will have set_clock and get_clock for the QMX (does not work for QCX or QDX according to their manuals).
set_clock can take utc and local as arguments to set the current time.
rigctl -m 2052 -r com33 -vvv get_clock set_clock local get_clock
Opened rig model 2052, 'QCX/QDX/QMX'
0000-00-00T14:53:59.000+00:00
0000-00-00T09:54:00.000+00:00
rigctl -m 2052 -r com33 -vvv get_clock set_clock utc get_clock
Opened rig model 2052, 'QCX/QDX/QMX'
0000-00-00T09:54:06.000+00:00
0000-00-00T14:54:05.000+00:00
Mike W9MDB
toggle quoted message
Show quoted text
On Sunday, September 22, 2024 at 09:02:03 AM CDT, Hans Summers <hans.summers@...> wrote: Hi Mike? Yes there's a CAT command "TM" for get/set of the real time clock. I forget the format but it should be documented in the operating manual along with all the other CAT commands.? 73 Hans G0UPL On Sun, Sep 22, 2024, 15:01 Mike Black via groups.io <mdblack98@...> wrote: Is there a CAT command to set the clock?
I have added QRPLabs? to Hamlib already so can customize it if there are extra commands available.
rigctl -l | find "QMX"
? 2052? QRPLabs? ? ? ? ? ? ? ? QCX/QDX/QMX? ? ? ? ? ? ?20240919.2? ? ? Stable? ? ? RIG_MODEL_QRPLABS
Mike W9MDB
On Sunday, September 22, 2024 at 06:37:39 AM CDT, Hans Summers <hans.summers@...> wrote:
Hello Mark
Fixing the accuracy of the RTC is again something on my to-do list. The RTC runs from the 32.768 kHz crystal and like any uncompensated uncalibrated crystal they can be off. I will include some kind of software calibration to use the 25 MHz TCXO or GPS (if available) to fix this.?
Right now my priority is on SSB... I've got the SSB modulation working very well (see attached 700/1900 two tone test with excellent IMD3 performance) and now moving on to tackle the microphone audio processing, filtering and compression.?
But I will get to this RTC accuracy issue in due course. It's nothing wrong with your radio. So just wait a little please...
73 Hans G0UPL
On Sun, Sep 22, 2024, 14:07 Mark Palmer via groups.io <esperanto2023@...> wrote:
I'm using a QMX+ with the CR2032 installed and "QMX+ RTC" selected.
The problem is with the accuracy of the clock when the unit isn't switched on.
AFAICT, it's pretty accurate when the radio is operating in WSPR mode: it can run happily for several days at a time without any GPS input. However, if the unit's not powered up, the clock goes fast very quickly - I put the radio away for a few days and find the clock is several minutes fast, roughly 30 seconds a day. :-(
The 25MHz reference oscillator is virtually spot on: I've checked it using Hans's 20m WSPR tools (checked with both tx and rx), tweaked it a tiny bit (7Hz) and now everything's almost perfect.
Similarly, I've checked the CR2032's connections, cleaned the contacts and that seems well: the cell gives 3.15 volts measured at the PCB (healthy enough, surely?).
I find the on-screen time display helpful for logging purposes - I don't want to have to connect up the GPS or manually set it every time I leave the rig unused for a few days. What's up? How can I improve the accuracy?
73 Mark G?OIW
|
Re: Still no output on QMX
James,
You could use an LM-317 voltage regulator to knock-down the input voltage to any value you determine. It uses two external resistors in a voltage-divider circuit to set the output. They do recommend that the supply voltage be >3V over the desired output so the darlington pair inside of the LM317 has enough headroom to regulate.
?
I like the LM-317 regulators; They can supply up to 1.5 A and have thermal protection to shut it down if it runs too hot. A small heat sink (the size of a quarter) is all that is needed to keep it happy.
?
It also can function as a precision current regulator; There is an example on the Texas Instruments datasheet that shows how to hook that up. That might be good for folks who want a current limited supply for initial testing. It only requires a single resistor and the LM-317.
?
9.3.3 Precision Current-Limiter Circuit
This application limits the output current to the ILIMIT in the diagram.
?
https://www.ti.com/lit/ds/slvs044y/slvs044y.pdf?
--
Tisha, AA4HA
|
Re: qmx with tiny lcd display
Try this one:
?
I used this before instead of a standard 1602 LCD and it is compatible
?
?
--
?
73, Barb, WB2CBA
|
Re: QMX+ back up clock accuracy
Hi Mike?
Yes there's a CAT command "TM" for get/set of the real time clock. I forget the format but it should be documented in the operating manual along with all the other CAT commands.?
73 Hans G0UPL
Is there a CAT command to set the clock?
I have added QRPLabs? to Hamlib already so can customize it if there are extra commands available.
rigctl -l | find "QMX"
? 2052? QRPLabs? ? ? ? ? ? ? ? QCX/QDX/QMX? ? ? ? ? ? ?20240919.2? ? ? Stable? ? ? RIG_MODEL_QRPLABS
Mike W9MDB
On Sunday, September 22, 2024 at 06:37:39 AM CDT, Hans Summers <hans.summers@...> wrote:
Hello Mark
Fixing the accuracy of the RTC is again something on my to-do list. The RTC runs from the 32.768 kHz crystal and like any uncompensated uncalibrated crystal they can be off. I will include some kind of software calibration to use the 25 MHz TCXO or GPS (if available) to fix this.?
Right now my priority is on SSB... I've got the SSB modulation working very well (see attached 700/1900 two tone test with excellent IMD3 performance) and now moving on to tackle the microphone audio processing, filtering and compression.?
But I will get to this RTC accuracy issue in due course. It's nothing wrong with your radio. So just wait a little please...
73 Hans G0UPL
On Sun, Sep 22, 2024, 14:07 Mark Palmer via <esperanto2023=[email protected]> wrote:
> I'm using a QMX+ with the CR2032 installed and "QMX+ RTC" selected.
>
> The problem is with the accuracy of the clock when the unit isn't switched on.
>
> AFAICT, it's pretty accurate when the radio is operating in WSPR mode: it can run happily for several days at a time without any GPS input. However, if the unit's not powered up, the clock goes fast very quickly - I put the radio away for a few days and find the clock is several minutes fast, roughly 30 seconds a day. :-(
>
> The 25MHz reference oscillator is virtually spot on: I've checked it using Hans's 20m WSPR tools (checked with both tx and rx), tweaked it a tiny bit (7Hz) and now everything's almost perfect.
>
> Similarly, I've checked the CR2032's connections, cleaned the contacts and that seems well: the cell gives 3.15 volts measured at the PCB (healthy enough, surely?).
>
> I find the on-screen time display helpful for logging purposes - I don't want to have to connect up the GPS or manually set it every time I leave the rig unused for a few days. What's up? How can I improve the accuracy?
>
> 73 Mark G?OIW
>
>
>
>
>
>
|
Andrew, ?Thanks. Yes, CW only. I should have stated that in the question.?
|
Re: qmx with tiny lcd display
Yes, find a smaller display and rotate it 90deg is the first step i will do.?
Creating enclosure with a new arrangement of components and 3d printing the next step.
?
|
Re: QMX+ back up clock accuracy
Is there a CAT command to set the clock?
I have added QRPLabs? to Hamlib already so can customize it if there are extra commands available.
rigctl -l | find "QMX"
? 2052? QRPLabs? ? ? ? ? ? ? ? QCX/QDX/QMX? ? ? ? ? ? ?20240919.2? ? ? Stable? ? ? RIG_MODEL_QRPLABS
Mike W9MDB
toggle quoted message
Show quoted text
On Sunday, September 22, 2024 at 06:37:39 AM CDT, Hans Summers <hans.summers@...> wrote: Hello Mark Fixing the accuracy of the RTC is again something on my to-do list. The RTC runs from the 32.768 kHz crystal and like any uncompensated uncalibrated crystal they can be off. I will include some kind of software calibration to use the 25 MHz TCXO or GPS (if available) to fix this.? Right now my priority is on SSB... I've got the SSB modulation working very well (see attached 700/1900 two tone test with excellent IMD3 performance) and now moving on to tackle the microphone audio processing, filtering and compression.? But I will get to this RTC accuracy issue in due course. It's nothing wrong with your radio. So just wait a little please... 73 Hans G0UPL On Sun, Sep 22, 2024, 14:07 Mark Palmer via groups.io <esperanto2023@...> wrote: I'm using a QMX+ with the CR2032 installed and "QMX+ RTC" selected.
The problem is with the accuracy of the clock when the unit isn't switched on.
AFAICT, it's pretty accurate when the radio is operating in WSPR mode: it can run happily for several days at a time without any GPS input. However, if the unit's not powered up, the clock goes fast very quickly - I put the radio away for a few days and find the clock is several minutes fast, roughly 30 seconds a day. :-(
The 25MHz reference oscillator is virtually spot on: I've checked it using Hans's 20m WSPR tools (checked with both tx and rx), tweaked it a tiny bit (7Hz) and now everything's almost perfect.
Similarly, I've checked the CR2032's connections, cleaned the contacts and that seems well: the cell gives 3.15 volts measured at the PCB (healthy enough, surely?).
I find the on-screen time display helpful for logging purposes - I don't want to have to connect up the GPS or manually set it every time I leave the rig unused for a few days. What's up? How can I improve the accuracy?
73 Mark G?OIW
|
Re: QMX+ back up clock accuracy
Hello Mark
Fixing the accuracy of the RTC is again something on my to-do list. The RTC runs from the 32.768 kHz crystal and like any uncompensated uncalibrated crystal they can be off. I will include some kind of software calibration to use the 25 MHz TCXO or GPS (if available) to fix this.?
Right now my priority is on SSB... I've got the SSB modulation working very well (see attached 700/1900 two tone test with excellent IMD3 performance) and now moving on to tackle the microphone audio processing, filtering and compression.?
But I will get to this RTC accuracy issue in due course. It's nothing wrong with your radio. So just wait a little please...
73 Hans G0UPL
I'm using a QMX+ with the CR2032 installed and "QMX+ RTC" selected.
The problem is with the accuracy of the clock when the unit isn't switched on.
AFAICT, it's pretty accurate when the radio is operating in WSPR mode: it can run happily for several days at a time without any GPS input. However, if the unit's not powered up, the clock goes fast very quickly - I put the radio away for a few days and find the clock is several minutes fast, roughly 30 seconds a day. :-(
The 25MHz reference oscillator is virtually spot on: I've checked it using Hans's 20m WSPR tools (checked with both tx and rx), tweaked it a tiny bit (7Hz) and now everything's almost perfect.
Similarly, I've checked the CR2032's connections, cleaned the contacts and that seems well: the cell gives 3.15 volts measured at the PCB (healthy enough, surely?).
I find the on-screen time display helpful for logging purposes - I don't want to have to connect up the GPS or manually set it every time I leave the rig unused for a few days. What's up? How can I improve the accuracy?
73 Mark G?OIW
|
QMX+ back up clock accuracy
I'm using a QMX+ with the CR2032 installed and "QMX+ RTC" selected.
The problem is with the accuracy of the clock when the unit isn't switched on.
AFAICT, it's pretty accurate when the radio is operating in WSPR mode: it can run happily for several days at a time without any GPS input. However, if the unit's not powered up, the clock goes fast very quickly - I put the radio away for a few days and find the clock is several minutes fast, roughly 30 seconds a day. :-(
The 25MHz reference oscillator is virtually spot on: I've checked it using Hans's 20m WSPR tools (checked with both tx and rx), tweaked it a tiny bit (7Hz) and now everything's almost perfect.
Similarly, I've checked the CR2032's connections, cleaned the contacts and that seems well: the cell gives 3.15 volts measured at the PCB (healthy enough, surely?).
I find the on-screen time display helpful for logging purposes - I don't want to have to connect up the GPS or manually set it every time I leave the rig unused for a few days. What's up? How can I improve the accuracy?
73 Mark G?OIW
|
Re: Still no output on QMX
On Sun, Sep 22, 2024 at 05:25 AM, James wrote:
winding T501? 12V or 9V?
James, there are pros and cons for both. Searching through the forum you will find some discussion.
?
I wish you success for your work and hope to read good news from you.
?
73, Ludwig
?
|
Re: Still no output on QMX
On Sat, Sep 14, 2024 at 11:03 PM, @Ludwig_DH8WN wrote:
On Sat, Sep 14, 2024 at 04:26 AM, James wrote:
The Gate and the source voltage remains the same at 8.0V during all three "No Key", "P", "T".
If I read the tester display right, Q507 seems to be ok.
To be clear, it looks like Q507 is ok but I'm not sure because
But why is the voltage at G the same as at S?
So Q507 maybe not ok or something around Q507 is wrong. If voltage at G and at S are the same, across a well working Q507 no current is flowing and no voltage at the PA like for your QMX.
?
My suggestion:
?
Please check schematics and trace from gate of Q507 to R506 and collector of Q514. Check voltage of both sides of R506 and voltage at base of Q513 (signal DAC from the controller). Check also R507 for the right value and no short.
?
Normally a current should flow across R506 in each situation. When "no key" and "P" depressed this current is low and the gate voltage is a bit below the source voltage but not the same. When "T" depressed the signal DAC from the controller (B of Q513) is going from nearly 0 V to more than 1 V. This results in an increase of the current across R506 and the gate voltage of Q507 drops more. This lowers the source drain resistance of Q507 and a current is flowing to T501 and the BS170. This should be.
?
So you have to find why not for your QMX.
?
73 Ludwig
|
Re: Removing flux. Is it necessary? Yes/No
I use 99.9%? IPA (not the beer!) and an old medium toothbrush to dissolve the flux away. While the area is still wet I lay two layers of kitchen roll over it and rub over the top of that with the same toothbrush. It's a fairly gentle method, doesn't leave and debris and the boards come up minty-fresh every time.
?
Ronan
MM0IVR?
|
Donald S Brant Jr
va3rr
Tisha Hayes, AA4HA
Barb WB2CBA
Mark Palmer