Raspberry Pi-based WSPR project

[VK4ADC]

Hello Doug and the list

For comparison and inspiration here are some measurements on a Si5351A on the RFzero platform: http://www.rfzero.net/documentation/rf/

Bo, The Real OZ

Bo

Good info to explore.

A few comparison points :
Raspberry Pi Zero about $Aud20 shipped, 
RFZero about $Aud85 plus shipping - and I'm assuming that an Arduino  board is needed as well at $Aud???
(I don't play Arduino .. and don't plan to.)

... Then you need to add the band filters, power amps etc , cost much the same for either variant.
... Then you have to get familiar enough with the Arduino IDE and C to make it all work. The hardest bit.

My project started off as a low-cost WSPR generator. Sure the complexity and costs have risen with multi-band operation and power amps but those were my choice to undertake.

While I am not anything like a Linux/Raspbian guru, there are enough articles around that basically step-by-step take you through getting a Pi board going on WSPR using the commonly-available code.



The interesting bits (to me) about the RFZero board:

The RFzero board is an Arduino compatible GPS based RF clock generator built around the powerful SAMD21G18, an 48 MHz 32-bit ARM processor from Microchip (Atmel). The RFzero can generate frequencies from 2605 Hz and beyond 200 MHz. At the same time 28 I/O pins are available and eight of those are via an ULN2803A power driver. The programming and configuration is done via the USB port.

The typical use  of the RFzero is as a beacon (IBP, SPB, CW, FT8, JT9, PI4 and WSPR), signal generator, VFO, Qatari Oscar 100 dual LO or a low cost 10 MHz GPSDO. A dozen of example programs (sketches) are integrated into the Arduino IDE.

The RFzero has been tested generating WSPR signals on 200 MHz and receiving the fifth harmonics on 1 GHz with 100% decoding performance. On 1,3 GHz the decoding percentage was reduced to around 80%. However, this may be due to the performance of the receiver that was unlocked. The RFzero has also been tested generating PI4 signals on 108 MHz that were fed into a x96 multiplier resulting in a 100% decoded signal on 10 GHz. These tests were carried out with a foam covered oscillator.

The technical parameters (harmonic levels, phase noise etc) are all looking good - undoubtedly far better than the basic Raspberry Pi but the project complexity has risen many times.

I think I will stick with my basic Pi Zero... plus multi-band filters and PAs. And it's working well based on the signal reports out of EU on 10MHz WSPR of late.

Doug

[OZ2M]

Hello

There is already a MCU, similar to the one used on the Arduino Zero hence the name RFzero, on the RFzero. A GPS is also on the board and used for frequency calibration and timing purposes.

There is a dozen of example programs already made three of which are WSPR centric:
- install the Arduino IDE (Windows, Linux, Mac OS) [must be similar to RPi somehow]
- add the RFzero link to the Arduino IDE
- go the the menu and find the relevant RFzero WSPR program [no idea how this is done on Linux and RPi]
- upload it to the RFzero [must be similar to RPi]
- do the configuration via the USB port [hardcoded in RPi?]
- done

But great inspiration. I will do my first ever how-to-video which undoubtedly will take me longer time than performing the above steps.

There are no needs to learn any C at all. Not a single line of code has to be written. Connect the USB to a computer, a GPS antenna and an antenna/PA - done.

There is no need for a bandpass filter for the RFzero. A low pass filter may be added/mounted if at all relevant. I recommend investigating if the harmonics of, any, source are the culprit of the post PA harmonics. For non broadband PAs the harmonics are often generated in the PA itself instead of being an amplification of the input signal. I have tested a 6 m PA driven by 13 dBm from a R&S signal generator with really low second harmonic and compared this to a situation where a 100 MHz 13 dBm, seldom none of the harmonic are at the same level as the fundamental is, signal was also fed to the same input as the 50 MHz signal. The second harmonics from the PA was the same.

Performance costs.

Bo

[VK4ADC]

Bo

As I stated in my post, I simply don't "do" Arduino.  I have used PICAXEs for my projects for quite some years so I know quite a bit about them. Arduinos - I know nothing !!!!! (as Sgt Schulz used to say often in Hogan's Heroes). Don't know the variants at all - don't want to. Enough to do without that too.

Yes, performance costs - but the closest amateur to me (assuming he is currently active at all) is about 3-4KM (being semi-rural) so close-in noise emissions in-band are unlikely to annoy anyone but me.  The TV here is mainly satellite feed being quite some distance from the TV transmitters so no real issue there. The internet feed is FTTN then underground copper so no radio links to affect. 

The technical performance requirement here (at my home QTH) is a little looser than it might be in suburbia simply due to spacing to others who might be affected. Even so, my tests show > 50dBC has been achieved for harmonics etc, and close to - or exceeding - the ACMA requirement.

Others might choose the RFZero path, I'm not.

[VK2KRR]

I received your 6m WSPR signal yesterday Doug incase you missed it.
2019-05-04 01:02 VK4ADC 50.294514 -13 0 QG62lg 5 VK2KRR QF34mr 1008 km

[VK4ADC]

I received your 6m WSPR signal yesterday Doug incase you missed it.
2019-05-04 01:02 VK4ADC 50.294514 -13 0 QG62lg 5 VK2KRR QF34mr 1008 km

Leigh

Proves it still transmits on 6M - plus the other 4 bands !!

Doug

[VK2KRR]

01:32 VK4ADC 50.294542 -17 3 QG62lg 5 VK5PJ PF95mk 1525 km
Another nice one today to Peter.