Date
1 - 6 of 6
FST4 Doppler Spread Algorithm in WSJT-X
Tolboom, Ryan
Hello,
At HamSCI this weekend I had a discussion with Rob (AI6VN) regarding whether or not Doppler spread could be calculated in WSPR transmissions. I decided to dig into the WSJT-X source and take a look at how it does it for FST4. Anyway, I wrote up what I found out and thought this group might be interested in it. You can find my post here: https://using.tech/posts/fst4/.
73,
Ryan Tolboom N2BP
Rob Robinett
Hi Ryan,
Thank you so much for investing your time and expertise in studying this subject.
It seem to me that obtaining spreading values from the 4000 WSPR-2 mode beacon signals currently on air might significantly enhance WSPR's contribution to the measurement of the ionosphere.
Since It would be simple for me to add spreading processing of WSPR-2 mode to the many top WSPR receive sites, either by running a companion program to "wsprd" or by running an enhanced version of "wsprd", would you consider working with me to implement such an enhancement? We already have in place both the network of receive sites and a publicly accessible database where the measurements would be stored.
73,
Rob
Thank you so much for investing your time and expertise in studying this subject.
It seem to me that obtaining spreading values from the 4000 WSPR-2 mode beacon signals currently on air might significantly enhance WSPR's contribution to the measurement of the ionosphere.
Since It would be simple for me to add spreading processing of WSPR-2 mode to the many top WSPR receive sites, either by running a companion program to "wsprd" or by running an enhanced version of "wsprd", would you consider working with me to implement such an enhancement? We already have in place both the network of receive sites and a publicly accessible database where the measurements would be stored.
73,
Rob
Glenn Elmore
Ryan
Echoing Rob, Thanks so much for doing this and documenting it. I think it may be truly helpful to what many of us are working to accomplish.
I really appreciate it.
Glenn n6gn
On 3/23/23 09:20, Tolboom, Ryan wrote:
Hello,
At HamSCI this weekend I had a discussion with Rob (AI6VN) regarding whether or not Doppler spread could be calculated in WSPR transmissions. I decided to dig into the WSJT-X source and take a look at how it does it for FST4. Anyway, I wrote up what I found out and thought this group might be interested in it. You can find my post here: https://using.tech/posts/fst4/.
73,
Ryan Tolboom N2BP
Thank you so much for your splendid commentary on the dopspread algorithm. Your excellent notes have given me a much better appreciation of how w50 is obtained.
I think there are several points to bear in mind about implementation for users of wsprd:
1. WSPR tolerates drift. The dopspread algorithm would need to operate on the de-drifted data. Else drift turns into spread.
2. The 'de-drifting' may not be perfect, and any residual would I suspect contribute to the spreading estimate.
3. The current small cadre of FST4W users on the HF bands have, I think, an appreciation of the hardware requirements to result in meaningful spread estimates. For example, we have shown that some HF band transceivers capable of 100 W, when operated at 5 W for FST4W, can have too much drift in the frequency from transmit/receive to be useful because of thermal effects. We have also shown that adding an external phase-locked GPSDO to a transceiver in this class reduces the equipment's spectral spreading to the minimum possible.
4. The Spreading 'thermometer' I showed in my presentation, and as the graphic for my talk at http://wsprdaemon.org/presentations.html serves to show what level of instrumental spectral spread needs to be achieved to avoid misinterpretation of ionospheric and path spread.
Of course, I'd be delighted to help assess a prototype implementation for wsprd.
best wishes
Gwyn G3ZIL
I think there are several points to bear in mind about implementation for users of wsprd:
1. WSPR tolerates drift. The dopspread algorithm would need to operate on the de-drifted data. Else drift turns into spread.
2. The 'de-drifting' may not be perfect, and any residual would I suspect contribute to the spreading estimate.
3. The current small cadre of FST4W users on the HF bands have, I think, an appreciation of the hardware requirements to result in meaningful spread estimates. For example, we have shown that some HF band transceivers capable of 100 W, when operated at 5 W for FST4W, can have too much drift in the frequency from transmit/receive to be useful because of thermal effects. We have also shown that adding an external phase-locked GPSDO to a transceiver in this class reduces the equipment's spectral spreading to the minimum possible.
4. The Spreading 'thermometer' I showed in my presentation, and as the graphic for my talk at http://wsprdaemon.org/presentations.html serves to show what level of instrumental spectral spread needs to be achieved to avoid misinterpretation of ionospheric and path spread.
Of course, I'd be delighted to help assess a prototype implementation for wsprd.
best wishes
Gwyn G3ZIL
Rob Robinett
Hi Gwyn,
Ryan has scheduled a Google Meeting for 7AM PDT Saturday morning (16:00 UTC) for us to discuss how to proceed. I don't know if your weekend schedule will allow you to participate, but your email gives us good insights.
Rob
On Fri, Mar 24, 2023 at 7:59 AM Gwyn Griffiths via groups.io <gxgriffiths=virginmedia.com@groups.io> wrote:
Thank you so much for your splendid commentary on the dopspread algorithm. Your excellent notes have given me a much better appreciation of how w50 is obtained.
I think there are several points to bear in mind about implementation for users of wsprd:
1. WSPR tolerates drift. The dopspread algorithm would need to operate on the de-drifted data. Else drift turns into spread.
2. The 'de-drifting' may not be perfect, and any residual would I suspect contribute to the spreading estimate.
3. The current small cadre of FST4W users on the HF bands have, I think, an appreciation of the hardware requirements to result in meaningful spread estimates. For example, we have shown that HF band transcievers capable of 100 W, when operated at 5 W for FST4W, have too much drift in the frequency from transmit/receive to be useful because of thermal effects.
4. The Spreading 'thermometer' I showed in my presentation, and as the graphic for my talk at http://wsprdaemon.org/presentations.html serves to show what level of instrumental spectral spread needs to be achieved to avoid misinterpretation of ionospheric and path spread.
Of course, I'd be delighted to help assess a prototype implementation for wsprd.
best wishes
Gwyn G3ZIL