RX & TX Frequency reports


John
 

I was reading the QRP-Labs QDX operating manual today and found these
two WSPR frequency reporting URLs.
RX +- 1 is normal for the current software version & GPS.

Https://qrp-labs.com/images/wsprnet/rxerror.html

Https://qrp-labs.com/images/wsprnet/txfreq.html

John


Rob Robinett
 

Glenn N6GN has measured excellent frequency accuracy and stability for the QDX receiver. But the transmitter appears to heat up the crystal which introduces significant drift.   Also, for use as a FST4W beacon, the modulation technique differs significantly from the standard and thus limits its usefulness 

On Sat, Sep 24, 2022 at 9:53 AM John via groups.io <n0ure=yahoo.com@groups.io> wrote:
I was reading the QRP-Labs QDX operating manual today and found these
two WSPR frequency reporting URLs.
RX +- 1 is normal for the current software version & GPS.

Https://qrp-labs.com/images/wsprnet/rxerror.html

Https://qrp-labs.com/images/wsprnet/txfreq.html

John

--
Rob Robinett
AI6VN
mobile: +1 650 218 8896


Glenn Elmore
 

John,

Accurate rx and tx frequency reporting on HF, e.g. 20m, isn't simple.  There are several factors with radio HW being only one of them.  During active magnetosphere  and even during quieter times effective ionospheric height can vary greatly and rapidly appearing to be more dramatic sometimes at 40m than on 20m. Also, for Kiwis even those using GPSDOs there is a baseband error within  GPSDO referenced  Kiwis that amounts to >1 Hz of reporting error. You see that in N6GN/K reporting in your link. This is  because the GPSDO and Kiwi are set to 66.6 rather than 66.66 MHz.  STock GPS-aided Kiwis with recent FW actually are better in this regard and do very well. 

As Rob has mentioned, we've found that a thermally stable QDX which is NOT transmitting does quite well. It may have a frequency error but there is very little additional spreading. So it may at the same time report inaccurate frequency but also serve well for FST4W spots.  As soon as one starts cycling transmit though things go very bad in a hurry.  In addition, FST4W transmission encoding also contributes to spreading, perhaps approach .5 Hz, generally making it unsatisfactory for 20m FST4W (which doesn't remove drift in the way WSPR does)  even when externally phase locked with a good GPSDO.

Absolute frequency accuracy in many ways is not as vital as low spectral spreading, except for the case where spots are being used to measure absolute Doppler spread for ionospheric research - something that HamSci has been interested in studying to better understand the ionosphere, magnetosphere and effects of coronal mass ejections from the sun. 

So it's rather  complicated and just looking at a single number such as absolute frequency accuracy or short term stability and spectral spreading doesn't in itself tell the whole story. Radio HW, the ionosphere along with MUF and number of hops all can enter in.

This slide from the recent TAPR/DCC conference gives an idea of some of these issues and what we've found on 20m FST4W

On 9/24/22 13:53, John via groups.io wrote:

I was reading the QRP-Labs QDX operating manual today and found these
two WSPR frequency reporting URLs.
RX +- 1 is normal for the current software version & GPS.

Https://qrp-labs.com/images/wsprnet/rxerror.html

Https://qrp-labs.com/images/wsprnet/txfreq.html

John


Rolf Ekstrand
 

Have tried the QDX on 20 m FST4W.  However the newer ones are worse than the first rev 1 issue. The reason being is that the TCXO is located closer the the finals whereas on rev 1 is is at the edge of the board.  Then the enclosure is too small and tight.   I have been running both rev 1 and 3 at approx 2 w ( lower voltage) and out of the enclosure with additional cooling ( a copper U from the DIZ 5 W xcvr)  and  it still is not stable enough for 300 sec.  Even 120 shows significant spread with multiple decodes by the same rx/cycle 

73  Rolf K9DZT


Glenn Elmore
 

It's even worse than TCXO drift when tx is cycled. Even with a GPSDO injected as clock instead of the TCXO, additional spreading due to the algorithm that maps incoming audio to RF appears to add significant spreading.  I've written but not heard back from QRP Labs about this.

Meanwhile the receive side of QDX, with no transmissions thermally cycling the board is really very good WRT additional spreading. There may/will be additional absolute frequency error but the spreading contribution to FST4W is very small.

I hope Hans will be able to improve the transmitting/encoding algorithm. The QDX could then be a really capable and inexpensive FST4W platform when used with an external clock.


On 9/26/22 12:37, Rolf Ekstrand wrote:

Have tried the QDX on 20 m FST4W.  However the newer ones are worse than the first rev 1 issue. The reason being is that the TCXO is located closer the the finals whereas on rev 1 is is at the edge of the board.  Then the enclosure is too small and tight.   I have been running both rev 1 and 3 at approx 2 w ( lower voltage) and out of the enclosure with additional cooling ( a copper U from the DIZ 5 W xcvr)  and  it still is not stable enough for 300 sec.  Even 120 shows significant spread with multiple decodes by the same rx/cycle 

73  Rolf K9DZT