Date
1 - 9 of 9
Frequency window edge settings?
|
I am not an experienced wsprdaemon user so I hesitate to ask what is probably a dumb question, but here goes anyway, my curiosity compels me...
...I was testing a 10 mW WSPR beacon on 20m today and happened to notice that the frequency reported by KPH at 1436 UTC was >50 Hz offset from 2 of the other spotters at that time (WA2TP and my remote receiver KX4AZ/T). This then made me wonder if there is any provision in wsprdaemon to set a wider spotting bandwidth than 200 Hz, for cases where either the sender or receiver drifts off the mark. I imagine if the band edges are hard limited to the 200 Hz window that could impact the spot counts. |
|
|
|
Rob Robinett
Hi Bruce, No questions are dumb, and yours is particularly insightful. WD uses the 'wsprd' binary to decode spots and when is it executed with the ''-w" (wideband decode) command line flag it will search +-150 Hz for spots. You can instruct WD to add "-w" to WD's us of 'wsprd' by adding the line: WSPRD_CMD_FLAGS="-w" to your wsprdaemon.conf file. If you try that out, please report your results. Thanks very much for that alert about bad frequency reports from KPH. The GPS RF distribution system at KPH is currently down and so the KPG Kiwi's oscillators had drifted off by about 50 Hz. I have just manually corrected them so spots should now be recorded much more accurately. Tomorrow I will be on site at KPH to (among other things) replace the current passive GPS splitter with a custom active splitter supplied by Glen N6GN, and with good GPS signals the Kiwis will stay within 1 Hz accuracy. 73, Rob On Fri, Feb 11, 2022 at 10:41 AM Bruce KX4AZ <bruce@...> wrote: I am not an experienced wsprdaemon user so I hesitate to ask what is probably a dumb question, but here goes anyway, my curiosity compels me... --
|
|
|
|
Thanks Rob, I may indeed give the -w a try so see if it impacts the spot count any for my KX4AZ/T remote receiver, to catch some "strays" outside of the band. Right now I am only letting it run on 20 meters, since the antenna was damaged by a windstorm in December, and that band was still usable enough until I return in person to repair. But the full 0-30 MHz waterfall is just dreadful compared to what it had been...almost bad enough for me to de-list it from the KiwiSDR map. Ah the joys of a remote receiving site!
I just fired up the nanoWSPR transmitter on 20m and noticed that KPH spotted me at the correct frequency, about 42 Hz above the band center, confirming you corrected it. Speaking of band center, since 1500 Hz is the default setting in most WSPR transmission software (such as WSJT-x and the nanoWSPR I am trying), I imagine there are also lost spots where a new user leaves it set at the default, and overlaps with others in that center zone. Which also makes me wonder if the WSPR decoding software makes multiple passes and tries subtraction like the FT8 mode does. |
|
|
|
Glenn Elmore
Newer decoders do go through multiple passes, extracting energy from previously decoded spots to improve success and they seem to do a very good job of digging out spots on top of one another. But even without this, the fraction of stations that are actually even able to QRM given only their dial settings is pretty small, I think. Many years ago I scraped a fair number of spots from 20m using a <1ppb system and plotted the distribution of as-received transmitted frequencies. Even allowing for a +- 1 Hz window for the few that may have been Doppler spread when the MUF passed through and the ionospheric path length was changing, only a small percentage actually ended up within a 6 Hz WSPR BW of the default mid band value and atop one another in a manner that would increase QRM. You can do this again for yourself if you like by looking at N6GN/K data, which should deliver fractional ppb accuracy, taken from the central US. Perhaps things have changed somewhat in recent years, but I doubt it. The plot I made showed that the center of the distribution was 20 Hz or so high of band center and not sharp. This is perhaps because most stations using even "high stability" TCXO references do not have the accuracy and precision necessary to cause a problem, even on HF. On VHF and above, we have had to go to GPS reference to even get commercial multiband, multimode radios to work for WSPR. You can find (too much) detail on the 2m+ WSPR group site. I've attributed the frequency skew to the high side of band center to the fact that when non-disciplined quartz reference oscillators age they tend to slough off material and so drift higher. Glenn n6gn |
|
|
|
Bryan Klofas
Hey Guys--
toggle quoted message
Show quoted text
I also don't have GPS at my Kiwi location (I'll run another coax soon), so I gave the -w a try. I noticed that WSPRD_CMD_FLAGS="-w" overwrote the default wsprd command arguments, instead of appending -w to the default options. So the wsprd command at the end of 2 minutes was "bin/wsprd -c -w -f 7.03..." , instead of "bin/wsprd -c -C 500 -o 4 -d -w -f 7.03..." I used WSPRD_CMD_FLAGS="-C 500 -o 4 -d -w" and it seems to work fine now, not sure if I'm getting any more spots than before. I don't quite understand the -C wsprd option, though, why -C 500 instead of the default 10000? Thanks, have a great evening! -- Bryan Klofas KF6ZEO On 2/12/22 06:27, Bruce KX4AZ wrote:
Thanks Rob, I may indeed give the -w a try so see if it impacts the spot count any for my KX4AZ/T remote receiver, to catch some "strays" outside of the band. Right now I am only letting it run on 20 meters, since the antenna was damaged by a windstorm in December, and that band was still usable enough until I return in person to repair. But the full 0-30 MHz waterfall is just dreadful compared to what it had been...almost bad enough for me to de-list it from the KiwiSDR map. Ah the joys of a /remote/ receiving site! |
|
|
|
I've been running a WSPR beacon at -10 dBm output (0.1 milliwatts) for the past day, and purely by chance I happened to notice that the N2HQI receiver on 20m is more than 30 Hz off the mark. So the WSPR signals in the upper part of the band are being missed, lowering the total spot counts.
|
|
|
|
Rob Robinett
I believe that N2HQI uses a bank of RTL-SDRs feeding instances of WSJT-x for his decoding, so it is not surprising that his frequency reports are off on higher bands. RTLs can be fed by an external clock, so he might appreciate hearing about your observations. On Sun, Feb 20, 2022 at 11:01 AM Bruce KX4AZ <bruce@...> wrote: I've been running a WSPR beacon at -10 dBm output (0.1 milliwatts) for the past day, and purely by chance I happened to notice that the N2HQI receiver on 20m is more than 30 Hz off the mark. So the WSPR signals in the upper part of the band are being missed, lowering the total spot counts. --
|
|
|
|
Glenn Elmore
If you do mention it to him, he might e interested in https://groups.google.com/g/2-meter-wspr/c/_1ewnMTVlvU/m/Yck4SL7LAgAJ which has worked for VHF/UHF WSPR... Glenn n6gn
On 2/20/22 12:04, Rob Robinett wrote:
|
|
|
|
On Sun, Feb 20, 2022 at 03:02 PM, Glenn Elmore wrote:
Thanks for the interesting link to the work you did with the Bodnar GPSDO as an external RTL-SDR clock. I sent an email to N2HQI the other day about the frequency error on 20 meters; if I hear anything back I'll make sure to forward the info. If he does make the frequency correction it will be interesting to see how much that boosts the site in the spotter rankings versus the other "big guns" in the upstate NY area. This could become a new aspect to my WSPR "hobby", sending email alerts to spotters who have drifted enough to impact their spot counts. I have a renewed appreciation for the KiwiSDR's use of GPS to keep things accurate. I hadn't thought much about receive frequency errors until I started running the nanoWSPR transmitter, which doesn't hop randomly like the Zachtek unit does. Not only that, but since the nanoWSPR uses the Si5351a chip with a vanilla variety 25 MHz crystal, it has been interesting to observe the frequency drift as the room temperature changes. It would be a cinch to wire up my GPSDO to the 5351a chip...but currently more interesting to me to observe the temp. effect. Right now I have the output attenuated to -16 dBm (25 microwatts), and there are still a handful of spotters over the past day. |
|
|