10GHz - getting ready....

[radiogareth]

After successfully running up my Polish-sourced 10GHz CBBN 1W transverter on Friday I've progressed to sorting out a dish feed and spare dish for separate RX.
I picked up a skip-bound large SKY dish recently, unfortunately missing its LNB arm. A wanted request on this forum produced the required measurement (Thanks Steve G4HTZ ) and some 3-D printed spacers allowed stock 3/4 aluminium to be used for the boom. An LNB mount from unknown origins fitted, with slightly different spacers, the same being required for the actual LNB, a Bullseye devoid of its outer casing (previously removed to fit it perfectly onto a BSB mini-dish.
For a 10GHz source, I used a cheap eBay ADF4351 board with built-in LCD and keypad. Other than showing mHz, it works well. I set it to 3.460GHZ (x3=10.380GHz) and fed its output through a pipe-cap filter kindly tuned for me by Dave G8GKQ at last summers CAT21. This fed directly into an 'adapted to SMA' WG16 commercial feed.
Picture shows it down in the garden shed doorway.
For RX I used Langstone as the main task was to check the dish angles L+R and U+D. The signal was soon found at 630.160MHz (10.380-9.750 + a bit of reference difference) and the dish duly tweaked. At this point I had a huge signal, so wound the Langstone gain all the way down to -1dB at which point I had about the right level of signal for peaking.
What was interesting was seeing what I THINK is phase noise, unlike a clean CW carrier, I had that but additional little squiggles either side of a random nature - almost like ssb speech looking. I think I read somewhere that above 2GHz, this chip is prone to this. In the transverter, the LO will be running at 2.012GHz (x4 +TX & RX IF) so it may not be as much of a problem. Additional supply filtering does improve on this, low ESR capacitors ready to fit.
After establishing that I needed a few more degrees down-tilt capability and drilling alternative mounting holes, I fitted the dish to my 'TOM' and tested out the elevation/azimuth adjustment. Very nice, its going to be a very handy dish. Tomorrow I make a G3WPO offset dish feed and circular 22mm copper launcher from SMA, followed by more tests. The transverter has already done DATV 'across the shack' and the option of using a second dish and LNB might improve the RX capability. Testing will tell.....
Gareth

10GHZ_signal_source.jpg (218.27 KiB) Viewed 4033 times

Sky_dish_terrestial.jpg (308.26 KiB) Viewed 4033 times

Sky_TOM.jpg (285.39 KiB) Viewed 4033 times

[radiogareth]

Link to the possible 'phase noise' seen....
https://youtu.be/nQ9kx1Nr7-I
Comments welcome.
Gareth

[g4eml]

Hi Gareth,

Those squiggles are probably down to noise on the power supplies of the ADF board. Extra filtering should help improve that. The cheap Chinese boards are known for being poor in that area, especially the ones with the display. For your finished system I would recommend the SV1AFN board. They are more expensive but do perform better.

However with a bit of additional work the Chinese boards can be improved are probably good enough.


Colin.

[g8gtz]

To check if the phase noise is acceptable, try receiving your signal on an LNB into MiniTiouner.

If you get an MER above 15 dB the phase noise is "good enough".

Noel

[G8GKQ]

I'm not convinced that what can be seen on the video is phase noise Gareth.

Dave

[g4eml]

I agree, more likely oscillations or random noise on the regulator outputs. However the fix is probably the same, more supply smoothing.

That board must have a processor to drive the display and the Synthesiser. It’s quite likely the noise is coming from that, especially if they have used the same supply for everything.
For a fixed frequency it’s better to use a dedicated processor that just sets the frequency and then goes to sleep.

Colin.

[radiogareth]

More testing today, and looking at the chip pin-outs and where they go on the pcb (ADF4351) there are handy pads on the PCB where I fitted a couple of 100uF capacitors. Crude testing in similar circumstances to yesterday shows that whatever the wobbly bits were from, they have now gone MER across the shack is up to 19 and MUCH BETTER with a bullseye lnb than a cheap old Octagon that, as it does not say 'pll' on the label anywhere, is probably a DRO one and therefore poorer results to be expected.
Power output seems to have dropped though (I have been isolating the RF O/P via a circulator) but the pallet is now pretty warm so hopefully just very cautious temp related shutdown.
Will hopefully mount it all into some sort of usable fashion tomorrow such that its dish mounted/mountable with negligible feeders.
A G3PHO feedhorn was made this morning, as per data from here http://www.w1ghz.org/QEX/circular_wg.pdf specifically page 45 where I found my SMA sockets had a 0.050" pin diameter so I used the spacings specified. Some solder ingress around the end-stop was removed on my lathe. Wish a nanoVNA covered the band
Gareth

[g8gtz]

Good progress - 19dB MER is very respectable.

Next test - can you hear the Reigate 10GHz NB beacon on the Langstone?

Noel

[radiogareth]

Finished the G3WPO Modified horn feed for one of my dishes. Article here with the names of the copper fittings needed.
http://www.g3pho.free-online.co.uk/microwaves/horn.htm
Question is, with a 1/4 wave probe in circular waveguide, WHICH WAY ROUND should it be for horizontal polarisation? With the probe parallel to the ground, or vertical? I'm thinking vertical......
Gareth

G3WPO_feed.jpg (68.36 KiB) Viewed 3704 times

[G8GKQ]

Hi Gareth

Don't over-think it! Its just like a quarter wave whip. So the probe is horizontal parallel to the ground (SMA on the side of the feed).

Dave