Posts by iz2eeq

Hello David, thank you!

I’ve always loved GNU radio and have been using it for many years, starting with a USRP1 and two basic tx and rx boards... the software has improved a lot since then and the opportunity to embed it in a small device like a Raspberry makes it really interesting for experimentation.

I am using HackRF because I had an old unit in my box... I would say it works well but the required minimum sample rate (2M) is unnecessary high (and computationally expensive) for narrow bw applications, so I would recommend anything that can work with a lower bandwidth... I read somewhere that the Pluto supports as low as 312.5KHz, so that alone would make it a better device, hardware wise.

Pluto is also better when it comes to DAC/ADC, it uses 12 bits, while HackRF uses just 8.

On the other hand the support of HackRF on GNUradio, specifically with Osmocom and Soapy plug-ins is quite good and working well, especially the TX/RX switching that works out of the box.

I have not tested the Pluto on GNURadio so I wouldn’t know about it.

Ciao

matteo iz2eeq

Hi everyone, I've been finally able to (successfully) test an all digital, Gnuradio based implementation of a TX path on QO-100.

The idea was to build some Gnuradio Flows that could be run on a Raspberry PI and that implemented different transmission modes, using an UDP source as signal source so that the whole thing could be used remotely via wifi from another computer.

The HW used is a Raspberry PI 3 rev. B, connected to an HackRF One board cascaded with a 2.4GHz preamplifier and finally the ubiquitous Chinese 4W Wifi amplifier.

A build of Gnuradio 3.8 has been built on the Raspberry in order to achieve the max performance possibile on that architecture.

Everything worked nicely and I could work the TX remotely by sending characters via UDP and have them transmitted as CW, PSK31 and QPSK31. HackRF One is equipped with TCXO so frequency was very stable, which is fundamental and assured perfect reception.

Code for the tested flows is here : https://github.com/matteo-campanella/gnuradio-flows

More modes are coming - now testing voice single sideband modulation using gsm vocoders, looks promising!!!

This is the picture of the hardware assembly:

73 de Matteo iz2eeq

Successfully used method 1a on a Octagon Green, and it worked great! Very good hint, Remco!

Just built a CNC machined version of this patch feed, going to test it this weekend (hopefully) with a modified Octagon Green LNB for external osc. Killed two chinese tips in the making of, but finally achieved to cut the 1mm thick brass.

Has anyone tested one of these Green on the satellite? I’ve just got one from Amazon, the single model, and it’s quite chirpy if compared to the good old Octagon one. Now going for modification to inject frequency ref, but I’m a bit sceptic... if it’s already chirpy with quartz, could it get any better with an external ref, no matter how clean it is? Unless the problem is with a bad quartz obviously...

73 de iz2eeq

Hello Markro, do you think there would be any chance for this to run on a last generation raspberry from a computational power point of view? Obviously there would be mods needed because of moving to Arm, but do you think there would be enough power in an Arm processor to work?

73

Hello Harry, thanks for your comments. I had read the NEO7 was much less jittery when divided by powers of two but, as it turns out from your experience, it is not jittery free enough to be used on its own. Yet I wanted to go for a GPS disciplined solution, so I will end up using NEO as a reference together with a controlled oscillator and see where it gets!

73 iz2eeq

Hello guys, I’ve started experimenting with a cheap starcom, which have a 3566 pll and a 25MHz xtal , does anyone know if this chip can handle a 24Mhz clock (which would make a Neo7 module a good candidate for gps clock reference)? I could not find a data sheet...

73 iz2eeq

Quote from om0aao

Congrats, It will works! RG-58 on 2400MHz is great. HI. CU on the bird!

It is indeed, it makes loss calculation very easy, as 1dB per meter HI HI

I'm finally on the air with QO-100! The receiving part was already active, I was just waiting for the 2.4GHz transceiver and PA to arrive from SG Labs, and they finally did.

The RX part is composed by the LNB voltage generator, a raspberry pi 3 and a RTL based SDR; the LNB voltage generator is controlled by the raspberry, as well as the power to the TX part and to the rotator of the TX dish (controlled by a k3ng rotator controller on an Arduino Nano). Antenna is 80cm offset dish and the feed an Octagon PLL LNB, no clock injection.

The TX part is made of SG Labs transverter and a buck step down converter from 24V (from the RX box) to 13.8V. Dish is prime focus 1m and feed is an old AO40 self made patch antenna. In the end PA has not been used as the bare output of the transverter (about 1w at the patch antenna after 2 meters of RG58) is enough to bring me on the bird. A 20w PE1RKI isolator has been used at the output of the transverter to make sure nothing gets fried in case of bad matching with the antenna.

Here there are a few pics of the TX and RX chain; now that the work is done I can relax on an armchair and have a little fun on this beautiful satellite. See you on QO100!

73 de iz2eeq

Quote from Remco

I made other patches from FR4 PCB but the small spacing between the plates makes it difficult to solder the inner surfaces to the copper pipe.

But there is hope! M0EYT makes his patches from 36mm OD copper pipe. He anneals it, cools it red hot in water, sledges it with a hammer and CNC's it. He already send out about 50 (or may be even mode) plate pairs.

Oh my, that's where blacksmith meets ham

Hello everyone, i have a question about the material that can be used to assemble patches and helices reflectors... since I'm having a hard time finding a copper or brass foil supplier, would it be possible to use pcb to build those parts? What would be the effect of a double sided pcb FR4 at these frequencies (supposing that the patch and the reflector are made of pcb for example with the two sides connected together). Maybe it's a newbie question but that's what I am after all 😊. Thank you!

Quote from G3WDG

One issue that I have been thinking about with helix feeds is blockage from the helix structure in front of the feed pipe.

I have a simple feed here with 22mm water pipe protruding a few mm from the reflector of a 2.25turn G3RUH feed. The feeds works great on my big dish, but have been testing it today on a 76cm prime focus dish with approx 0.4 f/D.

Two tests so far, measuring the spectrum and MER of the WB beacon. Both tests show significant degradation when the helix is present (I unsoldered it from the feed while the feed was mounted in the dish, changing nothing else). The helix's reflector plate was left in place.

For marginal systems, particularly on the WB transponder, this would suggest that while simple to construct, a helix feed may be inferior to a patch. For the NB transponder I doubt if there is any discernible difference (from blockage) as the transponder noise floor is the limiting factor for most of us.

Upper screenshots are with helix present, lower two are with it removed. Spectrum is at 5dB/div.

There wasn't any measurable difference in beacon MER with the helix terminated with 50 ohms, or left open circuit.

There could be advantage in moving the open end of the feed further into the helix to reduce blockage, but as I understand it that would move the phase centres of the 2.4 and 10GHz feeds apart, leading to loss in gain for one of them. So some tradeoff may be possible.

Lots to experiment with these days!

73

Charlie

Display More

Hello Charlie, which software are you using to make measurements? Does it work with RTL based SDRs? I am looking for such a software to get repeatable measurements...

thank you

73 de iz2eeq matteo

PA3FYM either I found a way of filling the gap or the shiny discs will get back to sleep in the junk box.

PA3FYM yes, correct... should be an almost perfect fit for the copper pipe with an internal diameter of 22

Quote from Remco

Founding father of this feed is G0MJW and I assisted with finding the right feed point to let it produce CP. It is modelled in CST Studio and a MP4 movie of the simulations canbe seen here . Will try to find the picture of the dimensions.

In the meanwhile you can look at these ones here or here or here or here or here : -)

Edit: below the dimensions of the feed. Spacing between patch and reflector is 3mm. I have the feed point at X,Y = 8,28 and a cropped square reflector.

2300 - 2500 MHz simulation done by G0MJW is below. It confirms the right position of the feed point is critical for CP, i.e. balloon perfectly circling around Z = 50 + j0 Ω ! :

Display More

Would it change alot if the reflector had a diameter of 95mm?

I have a number of old hard disk platters, but they are smaller. Yet my karma would benefit from the reuse some stuff from my junk box.

That's a clever approach! And it would save one dish, too. I will start to recycle what was working on AO-40, and eventually explore the dual feed combo. thank you.

Hello everyone, after setting up rx with 95cm prime focus dish and octagon lnb to SDR, I am now in the process to setup the tx section - I still have the old patch antenna I used for AO-40 RX, design from PE0SAT - http://www.pe0sat.vgnet.nl/home-made/patch-antenna/, and I was considering to reuse it with another offset 90cm dish.

Nevertheless, I have noticed many are choosing the helix approach and I was wondering what are the advantages of helix with respect to patch, as the latter is in my opinion easier to build... not to mention I have it already ready

Thank you all and hope to talk to you soon on Es'hail 2!!!

73 de iz2eeq