Generator for receiving QO-100 on MAX2870

Not so long ago I became interested in communication through the Qatari satellite QO-100. Actually, it is such a unique geostationary amateur radio, which makes it completely unique. It is strange that nothing was written about him on Habré, technically this is a completely new level of amateur radio communication. It operates in the microwave range, so specific requirements are imposed on the equipment. In this note, I would like to tell you how I solved the problem of creating a high-quality oscillator on the MAX2870, which was unexpected for me, using Bluepill and Rust.

For those who hear about the QO-100 satellite for the first time, I will explain that in order to communicate with it, it is enough to have quite affordable equipment. For work, a conventional satellite television dish is used, reception is carried out in the 10 GHz band using a conventional television converter with vertical polarization. A simple, cheap and widely available RTL SDR, or any other SDR capable of receiving at a frequency of 740 MHz, can be used as a receiver – this is the intermediate frequency that the satellite converter outputs. The receiver is connected to the converter through a simple power injector. We launch the free SDRConsole program on the computer, tune in there to a frequency in the region of 740 MHz and you can listen to ham radio broadcasts from the satellite.

To tune the dish to the satellite, three beacons are used, which the satellite constantly transmits. We orient the dish according to the maximum received signal of the beacons.

For transmission into space, the 2.4 GHz band is used, which makes it possible to use ordinary household WiFi signal amplifiers and modules available on aliexpress as amplifiers. For transmission, however, not an ordinary WiFi antenna is used, you need a dish feed with circular polarization. Its design is primitive, a description of various options for self-production is available on the net. I use ADALM PLUTO SDR as a universal receiver and transmitter and an antenna, known on the net as POTY, made from a piece of copper water pipe and a pair of brass plates. As far as I understand, this is the simplest way to provide transmission to the satellite. A couple of watts of power and a plate with a diameter of 80 cm is enough for the universe to hear you.

I quickly assembled and set up the receiving part, but immediately ran into a problem that many wrote about – the reception of transmissions from the satellite goes like in an old tube radio. Until the lamps warm up, the sound of received stations “floats”. So in the case of satellite reception, it takes half an hour to “warm up” the equipment until the stations stop running away from the tuned frequency. But even after warming up, you still have to manually adjust.

The cause of the problem lies in the design of the satellite converter. In modern models of converters for the reference oscillator, a conventional 25 MHz quartz resonator is used, the frequency of which is multiplied by 390 times somewhere in the bowels of the converter, mixed with the signal from the satellite, after which we get 740 MHz at the output of the converter. I have always considered crystal oscillators to be very stable, but when used in the microwave, they no longer look like a good solution. Their frequency stability is no longer enough for comfortable reception of narrow-band radio signals that are used by radio amateurs.

The cunning SDRConsole program contains a possible software solution to the problem. Since the frequencies of the beacons transmitted by the satellite are known exactly, you can tune the reception to one of them and tell the program to monitor the received beacon frequency. If the frequency runs away, then the program will independently adjust the reception so that the beacon frequency remains constant.

For some time I used this method of reception, it is quite working, but still not entirely comfortable. The inner perfectionist in me wanted the best. And for this it was necessary to make and connect a highly stable generator to the converter. The task did not seem difficult to me, since the network has a lot of examples of how to do this, ranging from self-improvement of the converter and ending with all kinds of generator circuits. It is worth noting that already modified converters, and complete sets, can be bought, including in the Russian Federation, for example, from RA3APW. I also wanted to try everything that is possible to do with my own hands and experiment.

This is what the LNB board looks like after revision
This is what the LNB board looks like after revision

The refinement of the converter comes down to removing the quartz resonator from the board and connecting the resonator to one of the outputs of the resonator through a simple matching circuit to bring a signal from an external stable generator. For the convenience of refinement, they usually take converters with outputs to two TVs. One of the outputs is convenient to use as a connector for connecting a generator. I tried to remake different models of converters and I must say that it was somewhat more difficult to remake more modern models – they use very small SMD parts and, for all the simplicity of rework, you need good soldering skills.

But the most important thing is the generator. I immediately indicated that I wanted to use the so-called GPS DOCXO as a support, which in Russian means a thermally stabilized quartz oscillator, automatically adjusted to the clock signal from GPS satellites. Relatively speaking, satellites transmit accurate time signals, which give out on-board atomic clocks of very high accuracy and stability. These signals are used to constantly adjust the frequency of the crystal oscillator. On the same Aliexpress, there are enough offers of such generators. I bought one of these, but its output frequency turned out to be unregulated, it is always 10 MHz. Now it was necessary to somehow remake it in 25 MHz.

The first option I tried was to use the handy ICS512 chip. Inside it contains a PLL synthesizer, and this microcircuit is just capable of multiplying the frequency by 2.5 times. A very convenient and inexpensive solution to the problem, I thought, until I learned what phase noise is. I quickly made a multiplier on this chip and connected it to the converter. But along with the expected stable reception, I got a strange effect – the received signals became weak and less legible. I read on the forums that this may be due to high phase noise. With simple measuring equipment that is available to me, phase noise is difficult to measure, and when the frequency is multiplied by 390 times, the noise increases many times and spoils the reception. When dividing the frequency, the phase noise seems to decrease. The most unpleasant thing is that it is difficult to reduce this type of noise by simple means, such as installing filters.

However, I decided to try the filter option. I made a frequency divider 10 MHz by 2 at 155TM2. The 5 MHz meander received at the output has just the 5th harmonic of the frequency I need 25 MHz. I bought a handful of quartz crystals for this frequency, measured the parameters, selected similar ones in order to make a narrow-band filter in order to somehow reduce the phase noise. But with analog technology, I always have problems with tuning – with a narrow band it is difficult to get so that 25 MHz is in the bandwidth, I had to expand the band. As a result, I made a 4 crystal filter, I also installed an amplifier after it, then the converter worked. I must say that the result is much better than with ICS512. However, I did not like the complexity of the settings, I decided to try something else.

The search for possible options on Aliexpress led me to the generator module on the MAX2870 chip. It is not as cheap as the ICS512, but it has much more features. Most importantly, inside it has a main microwave generator from 3 GHz to 6 GHz, the frequency division of which produces an output signal. As a result, the microcircuit has a low level of phase noise. The selected module also has an input for an external reference oscillator, which can be used instead of the one soldered on the module. That’s what I need. As it turned out when examining the finished module, this input is in fact the output of the installed crystal oscillator. So switching to an external one will require the removal of jumpers – this is not very convenient. The board itself is of average quality, you can see the manual soldering of everything, and the wiring of high-frequency circuits leaves questions. There is no screening at all. But for the output frequency I need of 25 MHz, this is not so important.

The only significant drawback for me is the need for a control microcontroller, which must program the desired output frequency via the SPI interface. All this complicates the design and energy consumption.

In the storerooms I had a bluepill module on stm32, I decided to use it to set up the generator. The capabilities of bluepill, of course, are clearly redundant for such a task, but its price is junk, why not? In addition, it has just 3 volt logic, it does not need level converters that would be needed when using some kind of Arduino. Blupill connects directly to the generator module, no other parts needed. You only need firmware to configure the generator.

Here I must say that I am an enthusiast of using the Rust language to create programs for MK, so I immediately went to crates.io in search of a driver for the MAX2870. Unfortunately, I didn’t find anything, but a web search turned up a couple of libraries for Arduino, which I used as a cheat sheet. In addition, the Chinese seller of the module sent me some configuration code written in C. As it turned out, to run the generator, you need to send a couple of dozen bytes to it, so the program turned out to be quite simple. The only question is how to determine the values ​​of these bytes. Judging by the datasheet on the microcircuit, with the same given output frequency of the generator, there are several options for settings. It seems to me that the possibility of choice is superfluous, so I can’t call the MAX2870 convenient to use. In the end, I manually calculated all the values ​​and hardcoded them into the program.

The program can be downloaded here https://github.com/lesha108/max2870gen

Surprisingly, the program worked the first time and the generator gave the expected frequency. And here it was not without a fly in the ointment – the output signal turned out to be not at all sinusoidal and with a palisade of harmonics. So without a low-pass filter at the output, it is impossible to work with this generator. I applied the filter according to this scheme, with it the converter started up and worked quite well.

While I’m using a module with a built-in reference oscillator, the next step is to connect to the GPS DO. As it turned out, it turned out to have a 10 MHz sinusoidal output, so I have yet to make a square-wave converter with 3 volt logic levels, and such that the jitter is small. So the field for creativity and experimentation is still large. If someone has already solved a similar problem, share in the comments.

In conclusion, I wish all satellite communication enthusiasts good reception on 59!

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *