|
Hi Wei, yes it might well be from the WSPRnet database, the full locator is in there under my registration. ?That still doesn¡¯t explain why JD38 is being decoded by some and IO91 by others! it does seem to end up as JD38, ?I take this from the WSPR rocks website which shares WSPRnet and it¡¯s there to behold if you search the last 24hrs with my call sign as the tx. ?
I¡¯m running it into a 40M resonant 1/4wave with a return loss of >20db, 12V supply (real 12V!) and the output is circa 3.6W according to both my test set and vector power meter. ?At this time I can¡¯t really put my finger on anything outside of the QMX+ box as a suspect. ?
73
Clive, G4KCM
|
Hi Clive,
Have you tried to decode your own WSPR transmissions? You can do that if you have another receiver(let QMX+ transmit to a dummy load to avoid mishap), or you can decode from a nearby WebSDR. This will help eliminate some unknown from the equation. I do agree with you that QMX+ WSPR encoding is suspicious, maybe it hits a rare corner.
73, Wei
|
Hi, good idea, I have amongst many others a Flex 6500 that¡¯s lends itself very easily to doing that. ?Will try to make some time tomorrow to run the QMX+ ?into a dummy load rather than spoofing my location as the middle of the South Atlantic?
|
Hi Clive,
In your latest WSPR net reports, you got not only JD38, but KN53, PH08, etc. There is an easy way to isolate this issue: Take two videos of QMX+ screen, one when WSPR is normal, the other when it's weird, and compare 162 symbols. They should be the same if QMX+ is innocent,?otherwise,?Han has no way to deny :-)?
73, Wei
Hi Clive,
Have you tried to decode your own WSPR transmissions? You can do that if you have another receiver(let QMX+ transmit to a dummy load to avoid mishap), or you can decode from a nearby WebSDR. This will help eliminate some unknown from the equation. I do agree with you that QMX+ WSPR encoding is suspicious, maybe it hits a rare corner.
73, Wei
|
Sorry, I meant Hans has no way to deny.
|
Wei, I have just set it up now to monitor my own signal and will leave it overnight as it took abt 2.5hrs before it started to move me around the globe last night. I'm not uploading the spots! We shall see!!?
The only RF related thing I have adjusted was to trim out a +10Hz oscillator error.? Having spent much of my working life in industrial instrumentation and control systems I'm a bit OCD when it comes to metrology :-), my test equipment and radios where possible are therefore GPS disciplined.
I appreciate your thoughts!
Clive
|
There's no mention in the OP as to what GPS is being used.
I once had the same issues with my U2/U3S transmitters.
Using third party cheap GPS modules often gave a problem.
Changing to the official QRP Labs QLG1 or QLG2 GPS receivers
the problem went away.
--
- 73 de Andy -
|
On 09/06/2024 18:53, Wei AG6AQ via groups.io wrote: I do agree with you that QMX+ WSPR encoding is suspicious, It is not unusual to get false decodes. Look at WSWPRnet posts, so called "Hash collisions", an attempt to accommodate complex call signs, is often blamed. 73 Alan G4ZFQ
|
Andy it is the QRP Labs QLG3 module internal to the QMX+
73, Clive
|
I'd also be very intrigued to see a video of the screen when the decoded spots are correct, and one when they are not.?
If the sequence of 0, 1, 2, and 3's is the same in both videos?then categorically the QMX is not to blame. I think that the fact that the onscreen locator is correct, already substantially reduces the probability that the QMX is to blame, and certainly absolves the QLG3 or the QLG3 antenna or the parsing of serial data, from any blame.?
If the sequence of 0, 1, 2 and 3's is NOT the same in?both videos then I'd be both very surprised, and have to put on my thinking cap (since the call, locator and power settings input to the encoding function ARE the same...).
On Sun, Jun 9, 2024 at 10:34?PM Clive - G4KCM via <sandersclive= [email protected]> wrote: Wei, I have just set it up now to monitor my own signal and will leave it overnight as it took abt 2.5hrs before it started to move me around the globe last night. I'm not uploading the spots! We shall see!!?
The only RF related thing I have adjusted was to trim out a +10Hz oscillator error.? Having spent much of my working life in industrial instrumentation and control systems I'm a bit OCD when it comes to metrology :-), my test equipment and radios where possible are therefore GPS disciplined.
I appreciate your thoughts!
Clive
|
Ok, I understand what¡¯s needed and as it¡¯s likely due to the wx I will be sat on a camp site in the rain it would be a good project. ?What I tried this morning was to monitor my own signal locally and only had one QRA wrong which was decoded as IO90 instead of IO91. ?Encouraged with this I put it on air again and got probably about a half a dozen or so decodes and quite a few were incorrect it¡¯s in the database around 13:46 today. ?Changed to WSJT in WSPR mode and digimode on the QMX+ and got loads more spots and no incorrect QRA locators it¡¯s been running since ok.
I do find that initiating the internal WSPR beacon from the menu sometimes causes it to start but there are no decodes if I power cycle it comes up ok, I think this was a past issue? ?
Will try some more objective tests later in the week. ?I¡¯m intrigued as well!
73, ?Clive G4KCM
|
Hi Clive,
The results suggest QMX+ and QLG3 are OK. I suspect that it's caused by interference from another station with similar strength, accurate timing and frequency increase this possibility. You can try a different frequency, at least, we shouldn't see JD38.
Within that 200Hz window, it's more crowed in the middle 100Hz. I used to pick 137 as "my" frequency, until one day, I transmitted for a hour without any reports, I then found out a nearby station was transmitting exactly at 137, 13db strong than mine. I switched to "197" and so far so good.
At the end, maybe only the successfulness of WSPR is to blame.
73, Wei
|
On Mon, Jun 10, 2024 at 03:35 PM, Hans Summers wrote:
I'd also be very intrigued to see a video of the screen when the decoded spots are correct, and one when they are not.?
?
If the sequence of 0, 1, 2, and 3's is the same in both videos?then categorically the QMX is not to blame. I think that the fact that the onscreen locator is correct, already substantially reduces the probability that the QMX is to blame, and certainly absolves the QLG3 or the QLG3 antenna or the parsing of serial data, from any blame.?
?
If the sequence of 0, 1, 2 and 3's is NOT the same in?both videos then I'd be both very surprised, and have to put on my thinking cap (since the call, locator and power settings input to the encoding function ARE the same...).
On Sun, Jun 9, 2024 at 10:34?PM Clive - G4KCM via <sandersclive= [email protected]> wrote: Wei, I have just set it up now to monitor my own signal and will leave it overnight as it took abt 2.5hrs before it started to move me around the globe last night. I'm not uploading the spots! We shall see!!?
The only RF related thing I have adjusted was to trim out a +10Hz oscillator error.? Having spent much of my working life in industrial instrumentation and control systems I'm a bit OCD when it comes to metrology :-), my test equipment and radios where possible are therefore GPS disciplined.
I appreciate your thoughts!
Clive
?
?
Well the mystery is solved. ?
The more I thought about it and looking at the WSPR protocol I couldn¡¯t really see it being an issue with the QMX+ / QLG3. ?So I decided to have a close look at my signal. ?Plugged the QMX+ into a power attenuator -> Spectrum analyser and oscilloscope and it all became very obvious¡. Noise and mains hum from the nice little 12V Linksys wall wart I was using in the shack to power it! ?I had checked the output volts but nothing further. ?Plugged it into my analogue bench power supply, which I always use for initial tests as it¡¯s got current limiting, and ?all has been wonderful ever since. ?Sorry all, thank you for the comments and suggestions. At least it has made me learn more about WSPR than I would have had there been no issues! ?Head hung in shame!!
From a previous post I saw from Hans¡ ¡°Beware of Wallwarts¡±!?
73
Clive, G4KCM
|
I have two thoughts on this thread.
First, except for my Yaesu FT450D that I use for base QRP all of my other QRP endeavors are used on battery power, far less chance for voltage errors or additional unwanted noise. ?I use wallwarts for the device it was sold with or¡±maybe¡± to charge a battery, mostly I use solar for charging though.
Second, I have used WSPR for 7 or more years on a half dozen devices, and the U4B is my latest and favorite rig now. ?BUT, beware of the phony folks out there. ?I have noted to our powers (FCC) of some users using false Calls from unlikely locations. ?These devices are fully programmable to pretty much whatever excites you, so one needs to watch out for phony reporting as the WSPR system probably doesn¡¯t check for the validity of either side of the ¡°QSO¡±.
John
toggle quoted message
Show quoted text
On Jun 11, 2024, at 12:35, Clive - G4KCM < sandersclive@...> wrote: On Mon, Jun 10, 2024 at 03:35 PM, Hans Summers wrote:
I'd also be very intrigued to see a video of the screen when the decoded spots are correct, and one when they are not.?
?
If the sequence of 0, 1, 2, and 3's is the same in both videos?then categorically the QMX is not to blame. I think that the fact that the onscreen locator is correct, already substantially reduces the probability that the QMX is to blame, and certainly absolves the QLG3 or the QLG3 antenna or the parsing of serial data, from any blame.?
?
If the sequence of 0, 1, 2 and 3's is NOT the same in?both videos then I'd be both very surprised, and have to put on my thinking cap (since the call, locator and power settings input to the encoding function ARE the same...).
On Sun, Jun 9, 2024 at 10:34?PM Clive - G4KCM via <sandersclive= [email protected]> wrote: Wei, I have just set it up now to monitor my own signal and will leave it overnight as it took abt 2.5hrs before it started to move me around the globe last night. I'm not uploading the spots! We shall see!!?
The only RF related thing I have adjusted was to trim out a +10Hz oscillator error.? Having spent much of my working life in industrial instrumentation and control systems I'm a bit OCD when it comes to metrology :-), my test equipment and radios where possible are therefore GPS disciplined.
I appreciate your thoughts!
Clive
?
?
Well the mystery is solved. ?
The more I thought about it and looking at the WSPR protocol I couldn¡¯t really see it being an issue with the QMX+ / QLG3. ?So I decided to have a close look at my signal. ?Plugged the QMX+ into a power attenuator -> Spectrum analyser and oscilloscope and it all became very obvious¡. Noise and mains hum from the nice little 12V Linksys wall wart I was using in the shack to power it! ?I had checked the output volts but nothing further. ?Plugged it into my analogue bench power supply, which I always use for initial tests as it¡¯s got current limiting, and ?all has been wonderful ever since. ?Sorry all, thank you for the comments and suggestions. At least it has made me learn more about WSPR than I would have had there been no issues! ?Head hung in shame!!
From a previous post I saw from Hans¡ ¡°Beware of Wallwarts¡±!?
73
Clive, G4KCM
|
Beware of wallwarts, yes!
Nevertheless I don't find this a satisfactory explanation, as a resolution to the problem. I don't see how a low quality supply could cause a QMX to transmit WSPR with different locators, display the correct locator on the screen, and even show up in WSPRnet with different locators decoded in the SAME transmission. The WSPR protocol is heavily encoded with tons of sophisticated error correction logic. It works or it doesn't. There shouldn't be any case like this where you get different locators from the same transmission etc.?
A couple of weeks I tested QMX+ on QSPR, across all bands 80 to 10m. That was using WSJT-X for the encoding not the internal standalone beacon feature. Later when looking at the decode map there were some instances of reception reports of G0UPL from a Western US gridsquare, copied by other US stations. Additionally it was at a time when I would expect the band I was using to be completely closed, let alone from my QTH to US. So the only possibility must be either people intentionally or unintentionally pirating or, perhaps WSPRnet bugs.?
Weird stuff happens
73 Hans G0UPL
|
On Tue, Jun 11, 2024 at 09:35 AM, Clive - G4KCM wrote:
Well the mystery is solved. ?
The more I thought about it and looking at the WSPR protocol I couldn¡¯t really see it being an issue with the QMX+ / QLG3. ?So I decided to have a close look at my signal. ?Plugged the QMX+ into a power attenuator -> Spectrum analyser and oscilloscope and it all became very obvious¡. Noise and mains hum from the nice little 12V Linksys wall wart I was using in the shack to power it! ?I had checked the output volts but nothing further. ?Plugged it into my analogue bench power supply, which I always use for initial tests as it¡¯s got current limiting, and ?all has been wonderful ever since. ?Sorry all, thank you for the comments and suggestions. At least it has made me learn more about WSPR than I would have had there been no issues! ?Head hung in shame!!
From a previous post I saw from Hans¡ ¡°Beware of Wallwarts¡±!?
73
Clive, G4KCM
Clive Over the past 20yrs the worst wallwart I have ever come across was a 12VDC leftover from a Linksys router. Took a few weeks to locate it 2 floors away from the shack. It now resides at the recycler. cheers
|
On 12/06/2024 15:53, Hans Summers via groups.io wrote: So the only possibility must be either people intentionally or unintentionally pirating or, perhaps WSPRnet bugs. Hans, Part of the bad decode can be due to the WSPR protocol. The extended call hashtag protocol seems to be overwhelmed by so many users so more than one match becomes common. Other wrong locator problems may be due to bugs, maybe in the WSRPnet mapping and/or database. Sometimes misconfiguration, not submitting your 6 figure location, also pirates are a possibility, RX and TX.. 73 Alan G4ZFQ
|
Hello Alan
Yes, agreed, completely. Thank you for correcting my omission of this important consideration. I always felt the "extended" protocol with the 15-bit one-to-many hash code is an absolute menace that should only ever be used in very limited circumstances, if at all. It probably doesn't have enough warnings attached to it.?
Anyway let's not forget that for more than a decade, despite its quirks, WSPR has been tremendously successful, popular, educational and useful. All for no cost. You know what they say about looking into the mouth of a gift horse!
73 Hans G0UPL
On 12/06/2024 15:53, Hans Summers via wrote:
> So the only possibility must be either people intentionally or
> unintentionally pirating or, perhaps WSPRnet bugs.
Hans,
Part of the bad decode can be due to the WSPR protocol.? The extended
call hashtag protocol seems to be overwhelmed by so many users so more
than one match becomes common.
Other wrong locator problems may be due to bugs, maybe in the WSRPnet
mapping? and/or database.
Sometimes misconfiguration, not submitting your 6 figure location, also
pirates are a possibility, RX and TX..
73 Alan G4ZFQ
|
On 12/06/2024 16:24, Hans Summers via groups.io wrote: Anyway let's not forget that for more than a decade, despite its quirks, WSPR has been tremendously successful, In the last 15 years I've wasted loads of time on WSPR, still do. The most common complaint is wrong locators, mostly obvious, most other times a quick search confirms a false locator. Rarely a station's approximate location can not be confirmed, The chances of it being somewhere exotic and not being found by searching seem very unlikely to me. 73 Alan G4ZFQ
|
Curiously enough I am now experiencing this.?
?
I am running the unit off (what I believe to be) a fairly stable bench supply, set to exactly 12V. Not a wall wart, certainly.?
?
For many hours all reports were IO91. Now almost all are JD38. The screen of the radio reports IO91, and the GPS diagnostics on the radio all say IO91PI.?
?
Most odd!
?
It's interesting that we are in the same grid square Alan, and seeing the same mis-report too.?
?
Alex.
|
Clive - G4KCM
KK4ITX John