Re: QDX - RF Filter Sweep graphs?
My sweeps were done into a little 4-resistor dummy load. They seem consistent enough, but I could try it on a shielded load.
I don't imagine I'd know there was a problem if I didn't have the sweeps. Given how weird the results are and how few things there to go wrong in the simple filter circuit vs how normal the radio acts, I'm wondering if the problem is on the sensing side of things. That definitely seems like a Hans question.
¨CBrent WT4U
|
Re: QDX Isues with sWindows10
#qdx
Disabling USB suspend seems to have solved it for me. I'm new with win 10 so not sure yet about all its quirks, but already had some issues with audio drivers being overwritten after an update. So keep that one in mind too.
|
Re: QDX - rig control error
This is the ModemManager issue. ?Shut down the ModemManager service and disable it.
Aug
AG5AT
|
Re: MSF or DCF clock module instead of GPS
I was using an Arduino Nano and RTC (DS3231) in a remote corner of my basement with the U3S.
I was MAINLY looking for a simple PPS signal, but this thread seems to want Time and Date.
?
My simplified solution was going to split the RX/TX of a software serial port between the GPS and U3S.
The concept was:
When the Arduino received NMEA sentences, it would store the last one for each of the 3 main sentences (RMC, GSA, GGA) in NVRAM.
Without a GPS signal, the Arduino would read time from the RTC, read / change the time in the last stored RMC message, calculate the new checksum, and send that message to the U3S.
? ?The RX line connected to a GPS.
? ?The PPS and TX line connected to the U3S.
Then I found I could adjust the frequency of the U3S clock to be accurate enough for WSPR to run for a week.?
I just let it run into a dummy load for a week and used the spreadsheet to adjust the crystal frequency.? My U3S ran for about 2 weeks with very little drift.
--
Rob KB8RCO
|
Re: MSF or DCF clock module instead of GPS
Hans,
I too think that this is an interesting problem.
I had a quick look at what might be possible.
I was thinking that the GPRMC NMEA sentence would do all that is required. That sentence is supposed to contain Time, date, position, course and speed data. Position information allows the U3S to determine the maidenhead locator but course and speed are not required here - perhaps just set them to 0.
I am thinking of using an Arduino (Uno or Nano - the original poster indicated he has a Uno).
Initialising the RTC is one issue that could solved by writing code to set up an interface using SerialMonitor over the Arduino's USB port to allow the user to enter time and date (and lat/long position) from the PC and write that to the RTC. Subsequent starts would have this initial data entry facility timeout if no new information was entered or the port could not be initialised (stand a lone setup) and the RTC would just use whatever values it was running with.
A separate SoftwareSerial port (with just the tx line, since the RTC does not need to receive anything from the U3S) can be created and the Arduino just packages up the date and time fields from the RTC variables and streams out the $GPRMC sentences to the U3S.
Does the U3S look for the time data in any particular NMEA sentence?
I was thinking that using the RTC SQW output at 1Hz to simulate the 1PPS signal that a GPS would provide, might have too much jitter. I am not sure of its usefulness in this regard. I would suspect that using it for frequency synchronisation would be out of the question. However if the requirement is just for the U3S? to be synchronised to within a second or so of the actual time then perhaps just having the U3S extract the current time from the $GPRMC sentence would suffice to meet the need.
I searched through the Operations Manual for some information as to which NMEA sentences were being parsed by the U3S but could not find anything.
73 Fred VK2WS
|
Re: MSF or DCF clock module instead of GPS
Hi Razvan
I'm curious as to why GPS is not an option. Whether this is a technical problem, or an ideological objection??
Building?a DCF receiver and decoder, and re-packaging that into NMEA strings that replicate a GPS is an interesting project... I am just not sure I understand the "why"... also I think the 1pps signal will have too much jitter from the noise on the receive path, to be useful for frequency discipline. The time data and 1pps can be used to discipline?the real time clock.?
73 Hans G0UPL
On Wed, Nov 17, 2021 at 12:16 PM Nicholas Booth < nicholas@...> wrote: Hi Razvan
If the isssue is that satellites aren't visible from a location, but you do have an internet connection then this device might be of some utility:
?
It's a NMEA sentence generator which references NTP.? it works well for its design use (beware rabbit holes) but ought to be adaptable for anything that needs serial time data.
Regards Nicholas 2E0UCD On Tue, 16 Nov 2021, 22:29 DL2ARL, < dl2arl@...> wrote: Ladies and Gents, entire group,
after vainly fiddling around with different possibilities of timekeeping, only to find out that none of them works for me :-(
GPS is no option for me because of various reasons
MSF or DCF Clock modules receiving the european Time Signals would be great, but after a very promising start several months, years ago, I did not find any means and ways of implementation
I come back to a solution proposed by somebody (Rob KB8RCO) some messages above in this thread.
I finally managed to find a module containing a genuine 3231 from a european shop. The modules offered by far-east salespersons are not very clear about stating which chip is used, so I paid what might not been worth the price, but went for sure to get what I was supposed to.
Now I sit and stare to the fancy module (see the picture included), I can even connect it to an Arduino Uno and might even be able to find a software in the internet to make a real clock out of this combo, using some makeshift display and a pair of knobs.
But this is not of very much help for me: what is needed is some stream-translation beteween what the module speaks and what the U3S is supposed to understand via it's GPS Channel. Forget the pps Signal, I am not relying on it, the frequency stability of my rig is good enough using an oven. If the pps signals would be needed just to satisfy the U3S input, I could provide some fake pps signals from a wristwatch and never rely on them to make the frequency calibration (via U3S Setting).
But what I am keen of having is some kind of automatic setting of the U3S clock. The internal clock is behaving erratically because it most probably hardly depends on the computing load the processor is subject to and the time accuracy goes up and down in a stochastic way and I am unable to find a parameter set to set it steady for more than 24 hours before running out of the wspr slice.
So every now and then, I desperately need a time setting.?
Me myself, by being a digital analphabete, am not able to write a sketch to convert the output of this little timekeeper gem into the GPS format U3S expects.? At this point I am totally dependent of someone in the group that is mercifull enough to write or at least to point towards a conversion routine using an Arduino of sorts. Without this, I will still have to reset the clock of the U3S at least once a day, which is a pain: the clock setting of the U3S, not being intended by Hans? to be done by humans with those two sticky knobs, has been conceived looong before user friendlyness was invented.
All suggestions wellcome!
?Yours friendly, Razvan DL2ARL
|
Re: MSF or DCF clock module instead of GPS
Hi Razvan
If the isssue is that satellites aren't visible from a location, but you do have an internet connection then this device might be of some utility:
?
It's a NMEA sentence generator which references NTP.? it works well for its design use (beware rabbit holes) but ought to be adaptable for anything that needs serial time data.
Regards Nicholas 2E0UCD
On Tue, 16 Nov 2021, 22:29 DL2ARL, < dl2arl@...> wrote: Ladies and Gents, entire group,
after vainly fiddling around with different possibilities of timekeeping, only to find out that none of them works for me :-(
GPS is no option for me because of various reasons
MSF or DCF Clock modules receiving the european Time Signals would be great, but after a very promising start several months, years ago, I did not find any means and ways of implementation
I come back to a solution proposed by somebody (Rob KB8RCO) some messages above in this thread.
I finally managed to find a module containing a genuine 3231 from a european shop. The modules offered by far-east salespersons are not very clear about stating which chip is used, so I paid what might not been worth the price, but went for sure to get what I was supposed to.
Now I sit and stare to the fancy module (see the picture included), I can even connect it to an Arduino Uno and might even be able to find a software in the internet to make a real clock out of this combo, using some makeshift display and a pair of knobs.
But this is not of very much help for me: what is needed is some stream-translation beteween what the module speaks and what the U3S is supposed to understand via it's GPS Channel. Forget the pps Signal, I am not relying on it, the frequency stability of my rig is good enough using an oven. If the pps signals would be needed just to satisfy the U3S input, I could provide some fake pps signals from a wristwatch and never rely on them to make the frequency calibration (via U3S Setting).
But what I am keen of having is some kind of automatic setting of the U3S clock. The internal clock is behaving erratically because it most probably hardly depends on the computing load the processor is subject to and the time accuracy goes up and down in a stochastic way and I am unable to find a parameter set to set it steady for more than 24 hours before running out of the wspr slice.
So every now and then, I desperately need a time setting.?
Me myself, by being a digital analphabete, am not able to write a sketch to convert the output of this little timekeeper gem into the GPS format U3S expects.? At this point I am totally dependent of someone in the group that is mercifull enough to write or at least to point towards a conversion routine using an Arduino of sorts. Without this, I will still have to reset the clock of the U3S at least once a day, which is a pain: the clock setting of the U3S, not being intended by Hans? to be done by humans with those two sticky knobs, has been conceived looong before user friendlyness was invented.
All suggestions wellcome!
?Yours friendly, Razvan DL2ARL
|
Re: is the 10W linear really linear?
Take a look half way down this page for some linearity measurements
Geoff GI0GDP
On Wednesday, 17 November 2021, 04:17:32 GMT, Shirley Dulcey KE1L <mark@...> wrote:
The ham radio standard for an RF linear is pretty loose. In-band distortion products (intermodulation) should be down at least -30 dB from PEP, and some linears have been put on the market that don't even manage that. Harmonic distortion doesn't matter at all; the low pass filter on the output takes care of that. A PA or amplifier that meets that standard produces acceptable results for voice and most digital modes.
The QRP Labs 10W linear is linear by that definition. An amplifier with specs like that would be considered ludicrously non-linear if it were a high fidelity audio amplifier.
It's possible to do better. The PA of the Collins S-line uses negative feedback. Some high end transceivers have an optional Class A mode for the PA that trades off lower output power and poor efficiency for better linearity. DSP rigs can use predistortion,?which is essentially a digitally?implemented version of feedback. All of these techniques can produce lower levels of harmonic and intermodulation distortion. But unless you're experimenting with making a cleaner amplifier rather than just building something to get on the air, hams rarely bother with any of those things because they add cost and complexity.
On Tue, Nov 16, 2021 at 9:50 PM Brent DeWitt < bdewitt@...> wrote: I've built two of these now and have been able to get 10W or more out, but I have a question for the other builders:
Can this be built as a "Linear" amplifier?? As a career RF/EMC/EMI engineer, I'm pretty simplistic, and maybe harsh, to say that I expect a "linear" amplifier to have an output that is nearly identical to its input if the output is attenuated by the gain.? Has anyone made this happen with this amplifier?? I hear often that a filter is needed on the output to meet harmonic output.? Why would that be the case if the input meets -40 dBc?? I must be missing something.? Thank you all for you thoughts!
--
Brent DeWitt, AB1LF
Milford, MA
|
Re: QDX - rig control error
Hi Roger Instead of WSJT-X, take a look at JTDX ( ).
Good tip! I just installed it (Linux). Works well... I think I prefer it to WSJT-X. The appearance is to my eyes, nicer,?a slicker?user interface. I like being able to use "Dark mode" too.?
I was somewhat surprised that I can run WSJT-X, JTDX and JS8Call all at the same time. They don't argue about access to the Virtual COM Port for CAT, which I thought that they would. I suppose that they are all using the hamlib library for the underlying CAT access, so it appears to the PC as one exclusive connection to the actual port.?
I also fell victim too, to no RF output... A reminder of how useful QDX firmware version 1.01c is. I had the LED flashing twice... which means what... it means audio is getting to QDX but the audio level is too low. Same thing on both WSJT-X and JTDX, though the?output power slider was correctly at maximum in both cases. So I had to go hunting around in Linux and found eventually the "output volume" which I suppose is equivalent to Master Volume I dimly recall from Windows... not at 100%. After correcting that I got three flashes... and QSOs.?
73 Hans G0UPL
|
Re: QDX - first contact
#qdx
Thanks Lynn
Glad to hear it worked well with Win 11, I will add that to the website.?
Yes, the four SMD inductors are replaced by through-hole on the Rev2 PCBs. And no 220uF cap needed.?
73 Hans G0UPL
On Wed, Nov 17, 2021 at 9:13 AM Lynn < lynn@...> wrote: I just received mine this afternoon and spent the evening putting it together. I've got to say, I'm impressed.
I built mine for 12 volts because it works better in my station. I did the basic continuity checks spelled out in the manual then took a leap of faith and just plugged it in. The magic smoke stayed inside where it belongs. It came right up on my Windows 11 machine (yes, another leap of faith). I set up WSJT-X for the correct serial port, enabled CAT PTT, set the Tx and Rx audio to the QDX (mono), and maxed out the Tx level control as spelled out in the manual. The second station I called on 40 meters came back. I made 5 more contacts in about 20 minutes on 40 and 80 meters. Not bad for QRP. For me it works best to call CQ or pick out the strongest CQs. 5 watts will get you out there (all over the US according to PSK Reporter) but you'll usually be on the lower end and can easily get swamped by a strong stations on a busy band.
Tomorrow I'll check out 30 and 20 meter operation.
For those of you that didn't get in on the feeding frenzy when Hans opened ordering on the QDX, don't feel too bad (I know, easy for me to say). The modifications that needed to be done (adding the 220 uF cap and replacing the self resonating SMT inductors with through-hole units tacked to the bottom of the board) were a pain to install and should (I hope) be corrected in future versions of the board.
All in all, another great kit from Hans and crew.
73s, Lynn, KU7Q
|
Re: QDX - first contact
#qdx
I just received mine this afternoon and spent the evening putting it together. I've got to say, I'm impressed.
I built mine for 12 volts because it works better in my station. I did the basic continuity checks spelled out in the manual then took a leap of faith and just plugged it in. The magic smoke stayed inside where it belongs. It came right up on my Windows 11 machine (yes, another leap of faith). I set up WSJT-X for the correct serial port, enabled CAT PTT, set the Tx and Rx audio to the QDX (mono), and maxed out the Tx level control as spelled out in the manual. The second station I called on 40 meters came back. I made 5 more contacts in about 20 minutes on 40 and 80 meters. Not bad for QRP. For me it works best to call CQ or pick out the strongest CQs. 5 watts will get you out there (all over the US according to PSK Reporter) but you'll usually be on the lower end and can easily get swamped by a strong stations on a busy band.
Tomorrow I'll check out 30 and 20 meter operation.
For those of you that didn't get in on the feeding frenzy when Hans opened ordering on the QDX, don't feel too bad (I know, easy for me to say). The modifications that needed to be done (adding the 220 uF cap and replacing the self resonating SMT inductors with through-hole units tacked to the bottom of the board) were a pain to install and should (I hope) be corrected in future versions of the board.
All in all, another great kit from Hans and crew.
73s, Lynn, KU7Q
|
Re: QDX - first contact
#qdx
Cool Stuff! Just started building mine this morning and am at the point where I need to decide 9v or 12v. 12v makes it a simpler setup but have some just perfect buck switchers that are really small that can easily be put inline with the power cable.
73
-Mike
|
Re: Terminal Diagnositc Use
#qdx
Hi Chuck
> As an aside, does anyone have a simple explanation for the ugly noise on the? > far-left side of my WSJT-X waterfall? I assume I've got some problems with? > my LPFs, but I may be wrong - this noise shows up on all bands.
Yes.?
On the "WSJT-X - Wide Graph" screen (waterfall), in the controls and buttons section at the bottom, is a checkbox named "Flatten". By default is is ticked ("Flatten" is switched ON).?
From the WSJT-X manual :
"When?Flatten?is checked, WSJT-X attempts to correct for slope or curvature in the receiver¡¯s passband shape."
BUT... QDX does not have any slope or curvature of the receiver's passband shape - it has a very flat response from 200Hz to 3.2kHz, and a sharp roll-off outside that.?
The ugly noise that you see over to the left of the waterfall. is the result of large artificial gain applied by WSJT-X at the lower and upper frequency extremities. It is what you get when you try to "flatten" something which is already almost perfectly flat (QDX).?
So simply un-check that "Flatten" checkbox... and admire the approach to perfection...
Note that on some Software Defined Radios that do baseband operations on I & Q channels, there would indeed be a lot of noise near the zero, because it is very hard to get rid of things like power line hum. QDX does not suffer these ailments because the SDR inside QDX actually digitally implements a sort of superhet?receiver with a 12kHz Intermediate Frequency, so all the work occurs at 12kHz which is far enough away from zero that those noise effects are gone. 12kHz is just because it is exactly one quarter of the 48ksps sampling rate, that makes the maths less difficult.?
73 Hans G0UPL
|
Re: is the 10W linear really linear?
The ham radio standard for an RF linear is pretty loose. In-band distortion products (intermodulation) should be down at least -30 dB from PEP, and some linears have been put on the market that don't even manage that. Harmonic distortion doesn't matter at all; the low pass filter on the output takes care of that. A PA or amplifier that meets that standard produces acceptable results for voice and most digital modes.
The QRP Labs 10W linear is linear by that definition. An amplifier with specs like that would be considered ludicrously non-linear if it were a high fidelity audio amplifier.
It's possible to do better. The PA of the Collins S-line uses negative feedback. Some high end transceivers have an optional Class A mode for the PA that trades off lower output power and poor efficiency for better linearity. DSP rigs can use predistortion,?which is essentially a digitally?implemented version of feedback. All of these techniques can produce lower levels of harmonic and intermodulation distortion. But unless you're experimenting with making a cleaner amplifier rather than just building something to get on the air, hams rarely bother with any of those things because they add cost and complexity.
On Tue, Nov 16, 2021 at 9:50 PM Brent DeWitt < bdewitt@...> wrote: I've built two of these now and have been able to get 10W or more out, but I have a question for the other builders:
Can this be built as a "Linear" amplifier?? As a career RF/EMC/EMI engineer, I'm pretty simplistic, and maybe harsh, to say that I expect a "linear" amplifier to have an output that is nearly identical to its input if the output is attenuated by the gain.? Has anyone made this happen with this amplifier?? I hear often that a filter is needed on the output to meet harmonic output.? Why would that be the case if the input meets -40 dBc?? I must be missing something.? Thank you all for you thoughts!
--
Brent DeWitt, AB1LF
Milford, MA
|
Re: MSF or DCF clock module instead of GPS
Hi Razvan,
You'll probably get overrun with offers, Razvan, but in the unlikely event you aren't, I could take this on for you. I'm about to start on the build of my own U3S, but above that, I'm about to unleash the design & code for a basic version of a modest digital sampling AGC module to the unsuspecting QCX community for incorporation into Classic and Plus versions of the QCX transceiver. With due acknowledgement of - and respect for - Hans' commitment to provide an analog version, which will, no doubt, be perfectly adequate, I am continuing with my project because a) I had already committed to do it; b) because I thought I would add other useful features, having committed to the hardware; and c) because somebody had kindly loaned me his QCX Plus as a development platform, and I owe him the reward of a working accessory.
As the Teensy I'm using will be bored out of its tiny mind doing just the job I'm giving to it (although v2 will do some filtering), it would be easy to write the code you ask for, both for a simple "328P-type" Arduino for U3S, or for the Teensy for QCX use.
I have a little more incentive:
I've mentioned the Classic and Plus QCXs, but it seems to me that in some situations, an accurate clock might be a slightly more convenient and faster alternative for field operation to a GPS unit if it can plug into a Mini in place of a QLG 1 or 2e, especially as a field unit could have a display. I can probably use one of Hans' standard case for that version...
--
Julian, N4JO.
|
Re: is the 10W linear really linear?
Brent
Good questions. I suggest reading about differences of class A versus other types of amplifiers. Second, note that amplifiers have limits to their linearity, and the best produce some harmonics. Producing 10 watts output with reasonable efficiency results in harmonics not meeting regulations. A low pass sufficient to attenuate second harmonics is necessary. Hint all your rigs likely include them.?
Curt wb8yyy?
|
Re: Terminal Diagnositc Use
#qdx
I've been looking for the answer to this question, also, and it's just like Hans said:
In order to use the input analysis terminal tool in QDX:
?
1. Set QDX to enable VOX in the configuration screen (when connected to QDX on PuTTY).?
?
2. Make sure WSJT-X radio is set to None (no CAT connection).
However, to see the tones in the serial terminal with the input analysis tool, you must enable TX from WSJT-X. I've attached pictures of my WSJT-X config for the radio (for the audio, it is still set to use QDX's audio interface), and the QDX configuration with VOX enabled. I just cloned the QDX config in WSJT-X and changed it in the clone, so that I can easily switch back and for between these configs.
(Forgive me if the pictures are wonky - this is actually my first post, though I've been a long-time lurker. Thanks Hans and team for all the great kits!)
Anyway, once you have those things configured - you can see that WSJT-X hears the signals from the QDX audio interface - the only thing left to do is actually *enable TX* on WSJT-X, which generates a tone that the QDX captures for the input analysis. It lets you see *your QDX's* tone, not analyze the input from the antenna.
I only just figured this out - been trying to get it working for a couple days.
Hans actually said all this before in his QDX video starting at about 7:57 () but I guess I just didn't pay enough attention to it. I'm a little dense...
Anyway... this is a headache I've finally gotten past! Enabling TX! O_o Hope it helps some others watch colorful blocks flit across their terminals, too!
As an aside, does anyone have a simple explanation for the ugly noise on the far-left side of my WSJT-X waterfall? I assume I've got some problems with my LPFs, but I may be wrong - this noise shows up on all bands.
Thanks,
Chuck, KI5JUB
?
|
Just finished building my QDX and have had a few contacts on 30 and 40m this morning. I did upgrade to the 1.01c firmware, and build with 2:2 for 12V power.
I was a little worried about the enamel getting roughed up making T2, so far it hasn't hurt. I had one bad solder job where I put in the power jack that caused
some intermittent operation until redone. My PC board and case fit smoothly with no issues. I see three rapid blinks while transmitting and seem to be getting
out, here is my 40 meter pic with spots. K9JRW
?
|
is the 10W linear really linear?
I've built two of these now and have been able to get 10W or more out, but I have a question for the other builders:
Can this be built as a "Linear" amplifier?? As a career RF/EMC/EMI engineer, I'm pretty simplistic, and maybe harsh, to say that I expect a "linear" amplifier to have an output that is nearly identical to its input if the output is attenuated by the gain.? Has anyone made this happen with this amplifier?? I hear often that a filter is needed on the output to meet harmonic output.? Why would that be the case if the input meets -40 dBc?? I must be missing something.? Thank you all for you thoughts!
--
Brent DeWitt, AB1LF
Milford, MA
|
Hi,
Any advice on troubleshooting my QDX PTT function.? Button turns red in Wsjt-x and stays there.
Thx,
Sam
AC4OW
|
Vojislav [YU4HAK]
Shirley Dulcey KE1L
Julian N4JO
Chuck AI5LC
Brent DeWitt
Sam Hanie