Listening with HackRF + OCXO + SDR#
465 views
Skip to first unread message
Jon Suite
Jun 30, 2016, 2:41:15 AM6/30/16
to 2 Meter WSPR
I've been listening to 2m/70cm WSPR the past few days using a HackRF SDR clocked with an external OCXO. A few thoughts:
1. The HackRF's sensitivity is very poor compared to my 857D. I can't pick up 70cm at all and 2m is marginal. I may try using an external preamp to see if it helps.
2. SDR#'s tuning can't be trusted. You will always have a constant offset somewhere in the range of +/- 30Hz depending on what SDR# has decided to use as the HackRF's LO. This was driving me crazy when I was trying to figure out the error of the OXCO since sometimes it would appear a little slow and other times it would appear a little fast. If you look back at my reports for the last few days you'll see what I mean.
3. The OCXO itself seems to be stable within +/- 1Hz over 24 hours even though I've put it in the worst conditions I could find without sticking it outside. It's sitting on a windowsill where it catches the afternoon sun.
4. Thanks to all of you running a GPSDO so I can have several good references to test against :)
-Jon AI6AJ
1. The HackRF's sensitivity is very poor compared to my 857D. I can't pick up 70cm at all and 2m is marginal. I may try using an external preamp to see if it helps.
2. SDR#'s tuning can't be trusted. You will always have a constant offset somewhere in the range of +/- 30Hz depending on what SDR# has decided to use as the HackRF's LO. This was driving me crazy when I was trying to figure out the error of the OXCO since sometimes it would appear a little slow and other times it would appear a little fast. If you look back at my reports for the last few days you'll see what I mean.
3. The OCXO itself seems to be stable within +/- 1Hz over 24 hours even though I've put it in the worst conditions I could find without sticking it outside. It's sitting on a windowsill where it catches the afternoon sun.
4. Thanks to all of you running a GPSDO so I can have several good references to test against :)
-Jon AI6AJ
Glenn Elmore
Jun 30, 2016, 9:54:56 AM6/30/16
to 2-mete...@googlegroups.com
Jon,
The frequency accuracy problem also exists with RTL-dongle SDRs, even though these have pretty good sensitivity. WINSWL, in particular, has run one of these all the way to 23 cm. I think at this point he's only managed 2m and 70 cm WSPR spots but it does work. He also found that settability depended upon the configuration. In particular the driver used to steer the hardware may not have enough math precision to get sub-Hz accuracy. Tuning may not even be monotonic. I think he sees few-Hz stability at 70cm with the 'tcxo' version of the dongle, not bad for a free running system.
This problem also exists with Hans Summer's U3s QRSS/WSPR beacons. Though he finally got math libraries with adequate precision to calculate frequency settings for the on-board synthesizer, an Si5351a, he deliberately sacrificed center frequency accuracy for accurate tone spacing in order to use programming values that result in somewhat better phase noise, per Silicon Labs' data sheet.
Even with GPSDO precision, there are remaining problems. I think the FT8x7 Yaesu radio's actually only step in 10 Hz increments, no matter what you might ask for by CAT. Thus even with a perfect master 22.625 MHz oscillator, they may still show up to a 5 Hz error. Once identified this can be adjusted for with the WSPR dial settings. If the receive frequency set in WSPR is the *actual* frequency then the transmit can be set accurately as well.
Thanks for sharing your experiences.
Glenn n6gn
The frequency accuracy problem also exists with RTL-dongle SDRs, even though these have pretty good sensitivity. WINSWL, in particular, has run one of these all the way to 23 cm. I think at this point he's only managed 2m and 70 cm WSPR spots but it does work. He also found that settability depended upon the configuration. In particular the driver used to steer the hardware may not have enough math precision to get sub-Hz accuracy. Tuning may not even be monotonic. I think he sees few-Hz stability at 70cm with the 'tcxo' version of the dongle, not bad for a free running system.
This problem also exists with Hans Summer's U3s QRSS/WSPR beacons. Though he finally got math libraries with adequate precision to calculate frequency settings for the on-board synthesizer, an Si5351a, he deliberately sacrificed center frequency accuracy for accurate tone spacing in order to use programming values that result in somewhat better phase noise, per Silicon Labs' data sheet.
Even with GPSDO precision, there are remaining problems. I think the FT8x7 Yaesu radio's actually only step in 10 Hz increments, no matter what you might ask for by CAT. Thus even with a perfect master 22.625 MHz oscillator, they may still show up to a 5 Hz error. Once identified this can be adjusted for with the WSPR dial settings. If the receive frequency set in WSPR is the *actual* frequency then the transmit can be set accurately as well.
Thanks for sharing your experiences.
Glenn n6gn
--
You received this message because you are subscribed to the Google Groups "2 Meter WSPR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 2-meter-wspr...@googlegroups.com.
To post to this group, send email to 2-mete...@googlegroups.com.
Visit this group at https://groups.google.com/group/2-meter-wspr.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages