2400MHz Amplifier Test.

Finally, after the winter solstice, I had a bit more energy.

I purchased the 2400MHz amplifier from SG Lab in Bulgaria. This has a good reputation for reliability, and I wanted a bit more output than the 2.5W I already had.

This is the v3, needing less than 40mW to deliver a typical output of 20W @28V power supply and 17W @24V.

I did have a signal generator using a computer control Windows program, but no usable Windows machine with a CD-ROM drive. Oops! What do we do now?

Well, well. I do have my IC-905 with variable power output up to 2W nominal.

The RF unit is in my "shack" on the ground floor, and the lab with all my test equipment is in the attic, so I picked up the controller unit and a piece of cat8 network cable, so I could control the thing from the upstairs lab. Now to get the RF power upstairs. As I only needed maximum 40mW, I could use some cheap SMA cables, RG174, two lengths of about 5m. 

This makes for a neat cheap attenuator. With the 2W output from the TRX I could measure about 30mW, using the peak indicator of the spectrum analyzer. That is good, as I can't overload the amplifier.

Because the output of the PA is 10-20W max, I needed to bring that down the level. I used a directional coupler (800-2500MHz), with 40 dB coupling. Bingo! 10W becomes 1mW. Of course a dummy load capable of dissipating the output power was connected to the "through" output.

Connecting the system I could measure about 10W out with a drive from the TRX of about 33% (650mW minus the cable attenuation) Result: 0dBm out of the directional coupler 10W from the amplifier.

Increasing the drive gave about one more dB.

This looks like to low output, but I would consider it within the tolerances of the test equipment. There is some extra loss in the connectors, adapters and cables after the output of the amplifier, and the spectrum analyzer, as well as tolerance of the input power measurement of the spectrum analyzer. so measuring 12W maximum output is within the tolerances of the minimum 15W specified from the amplifier, by about 1.5dB. That is good enough for me.

The amplifier test setup was the module attached to a cooling fin, with Kapton tape, and using some good thermal compound. 

It did not get hot to the touch by normal operation, even using full carrier FM.

I intend to use it for a while indoors, and if it looks stable, I will likely mount it outdoors. this will require a better mechanical construction, but also provide more power available to the antenna.

This setup is intended for QO-100 uplink use.

If this works well, the next step is some more power for 1297MHz FM.

Up-link Up Converter and System for QO-100, and Some Success.

Finally, I got some test equipment for 2.4GHz up and running. It is not yet complete, but I could start testing the up converter and power amplifiers in the lab.

A few tests and measurements were done, with the following results.

For temperature stability reasons I decided to place the up-converter per se indoors. The unit can deliver between 1.5 and 2W. The outdoor unit contains a Chinese WiFi booster, a so-called 8W unit, the Edup AB-003. Tests show the saturated output power just under 4W, and with 3W it would probably have sufficient linearity for SSB on the satellite.

The test was done in the lab, with a cable length approximately the length that will be used from the indoor to the outdoor unit, 15 - 20m. I decided to go low cost and use low cost, i.e higher loss, cable, since a really low loss cable would be overly expensive and un-flexible.Loss is about 15dB.

With this setup it was impossible to get more than 2W out of the Edup (actually a bit less). Ach! too little drive for the Edup. What to do?

Next test, simulated indoor unit with an extra Edup amplifier, only a little less than 3W out. Hmmm! Now what?

One more test: (Simulated) outdoor unit with 2 Edup amplifiers in series. That helps. Just under 4W is now easily possible, the system gain is good. 

Next "problem": With lower gain/output of the indoor up-converter the local oscillator and image rejection is reduced. I want as clean a signal as I can reasonably get. OK, a 2.4GHz PCB "hairpin" filter mounted between the two Edup amplifiers in the outdoor unit should do it. The filter should improve the LO rejection about 30dB, easily compensating for the reduced rejection. On top of that the output of the up-converter has to be increased due to 5-6dB loss in the filter. Still sufficient system gain, as expected. The image rejection for a 432MHz IF is improved about 45dB. All looks good. 

Testing this setup results in just under 4W saturated out of the Edup amplifiers, and we are ready for an initial live test. The amplifiers and the filter are mounted in an (electrician's) distribution box A 2.4GHz patch antenna is connected to an extended cable. At this point, the first test will provide 2W saturated at the patch antenna that will be mounted on a 60cm dish. This should deliver a good CW signal over the transponder. SSB, however will sound ugly because of the threshold for the RF sense in the amplifiers.

A set of attenuators in series have been connected, and should provide sufficient attenuation of the 35W from the transceiver, not over-driving or destroying the up-converter mixer.

The system has now been set up, and is tested. It is far from ideal, many improvements and optimizations are possible.

The test with CW is a qualified success. The signals are not very strong, and the first evening there were no replies to my CQ calls, even though I could receive my own signal (too late for much activity). The afternoon/evening after (today, Saturday) 12 QSOs have been made, and most of the stations could receive my faint signals.

I was warned that the patch antenna is not ideal, so at some stage I will make a helical feed - and yes, I will have to make sure to get it wound the correct way.

A shorter cable from the outdoor PA, combined with the helical, should provide at least 5dB better signal-to-noise ratio, and a bigger dish (the present one is 60cm) will provide even more gain. It looks like I can find a 110cm dish from a local amateur.

A modification of the Edup amplifiers for constant TX mode will have to be done, so SSB transmission will be possible. About 3W to the antenna should be possible. A higher power amplifier may be convenient to have.

All in all much to do to improve the system, a part of it (indoor activities) can be done anytime, but a bit of the outdoor stuff should be done soon, before winter.

 Other improvements will be increased frequency stability and precision. At the moment I have completely separate transmission and receiving systems, and the frequency offsets are different, so tracking transmit and receive frequencies is cumbersome. I will likely end up with a GPS locked system, but slowly, slowly, not too hasty.

Microwave Parts and Modules Part 2: Surplus PCBs and Modules.

Some weeks ago I went to a small ham gathering at the site of a  local repeater.

I had a talk with local hams about uplink equipment to the QO100 satellite. Many people are planning getting active on the satellite, so equipment for the 2400MHz band is in higher demand.
I had been planning to use some cheap Chinese modules for the local oscillator, but one suggested to find some better modules with Franco, alias RF-Microwaves.

Looking at the site I found that they have a lot of interesting stuff, especially in their surplus dept. Here are a few examples:
- A local oscillator module running at 2009MHz with the standard 8MHz crystal
- Some LNB modules without feed horn, but with input connectors
- PCBs with 4 HEMT FETs, where the PCB is arranged, so that it is possible to cut out two low noise amplifiers
- WLAN PCBs with several good components, including some 2.4GHz band pass filters
- 15GHz oscillator modules with several 10s of mW output.

I ordered a few of each, and some microwave absorbing foam. The service from Franco is excellent, I ordered one afternoon, and 2 days later, in the morning, the goods were delivered. The exception would if he is at a ham gathering, such as the Friedrichshafen Ham Meeting, probably the biggest in Europe.

Now for the goods, some of which are the last few left.

The 2GHz local oscillator module (about €20):
This is a complete module with casing and mounted with vibration absorption. Looks really good. On the website there is some modification information. The VCO can be used from about 1960 - 2040MHz, and the modification consists of making an external reference oscillator, replacing the internal 8MHz crystal oscillator. The external reference can be made with a well filtered DDS module.
For the IF of 432MHz the LO frequency should be 1968MHz, which is in the VCO range

As s second use of the LO module, it should be feasible to add a multiplier x5, for a local oscillator for a linear 10GHz transverter/converter.

A brief description of the LO module is on Franco's website.

LNB module (about €3):
This is a DRO controlled down converter module for satellite TV, with the mixer/oscillator in a single chip.
F6CXO has proposed the use of this module as a down converter for a (lower frequency) spectrum analyzer. The modification involves removing the 11/12GHz amplifiers and the filter. This cannot remove the image frequencies, so one needs to be aware of the frequencies that can be generated by the circuit you are testing. For amateur use, it does present a possibility to see what is going on in the 10GHz band, without having a more expensive spectrum analyzer. Mine covers up to 1.7GHz, so with a LO tuned to 11GHz the frequency range of 9.3 - 12.7GHz can be monitored.
Max RF input for the converter is about 0dBm (1mW). The RF input connector should be changed to a SMA, for the original ones it is probably very expensive or difficult to find a source for them.
The modification, if done carefully and with antistatic precautions, should provide you with 3 working GaasFETs, apart from the intended converter. I would probably make 2 converters, the second one using the lower LO frequency of about 9.7GHz, thus covering 8.0 - 12.7GHz with my analyzer.
A description of this modification is posted at Franco's website.
I would propose a few uses for the module, apart from sourcing components:
- Using the converter as it is, as a WBFM down converter to the 600 - 800MHz range.
- Adding a connector to the output of the 10/12GHz  preamplifier (before or after the (modified) band pass filter, providing a simple 10GHz amplifier with a gain of 15 - 20dB, likely with 20-30mW output.
- It might be possible to add injection locking to one of the DROs, if so the converter could be used as a simple narrow band down converter for 10GHz.
- Using the LNB as it is, for rather unstable down converter for wideband FM.
- Using the casing (moulded aluminium?) as it is for other modules I should build
I am sure other uses, apart from being a source of components, will turn up ...
Not bad for a €3 module.

PCBs with 4 HEMT FET amplifiers:
There is a document on this PCB, too, on Franco's website.
Here are some uses the document suggests:
- A low noise 10GHz LNA from a cut-out piece of the PCB. Each PCB could provide two of those in a relatively easy way, also salvaging two HEMT FETs. Yes, the board has 4 amplifiers, it looks like it has come from a dual head LNB. A negative supply for the gate of the FETs is necessary to add to the simple amplifier.
- Using a piece of the 50ohm stripline as a piece of test equipment, e.g. a component tester or for a bias-tee. The PCB material is good for up to 20GHz.
- Making 1 12/24GHZ doubler.
- The FETs are useable on 24GHz with up to +13dBm output and a gain of about 7dB.
This looks like an excellent source for building microwave stuff.
Like with the "connector-LNB", you get a lot for €3.

The WLAN PCB looks like a good source of components for building some 2.4GHz transverters/converters.
I was buying them, mostly for the 2.4GHz filters, but there is a PLL IC, a prescaler, and other good components there.

The last one, the LO PCB:
This contains two LO chains for around 14/15GHz:
- A straightforward receiver LO, starting on around 2.5GHz, adding multipliers. A tripler for 7.5GHz and a doubler for 15GHz
- A TX LO which also includes a modulator, meaning that it looks useable as a (WB)FM modulated transmitter, with a suitable PLL oscillator modification of the 2.5GHz part.
The output transistors, (two) coupled with a hybrid, are supposed to be capable of delivering about 100mW on 14GHz. I am now wondering if the hybrid is broadbanded enough to cover 10GHz, but that is something that needs to be tested. Also the 14/15GHz filters should probably be bypassed and replaced with some 5 and 10GHz filtering. I am hoping to be able to use a modified version of this PCB as a 10GHz power amplifier with 100mW output, if the hybrids have sufficient bandwidth. It may, however be necessary to design a separate PCB or set of PCBs, using the output FETs. With that, I will have to ask someone with **much** more experience than I have. I know, my general knowledge will probably not be sufficient for such a design ... yet.
- Also, an excellent source of components, FETs, MMICs etc.

Oh, well, so many ideas to try out, and lots of other activities. It will take time, and learning some new skills, but, as a start, I have a few of those modules

Es'hail 2 Sattellite, aka Qatar OSCAR 100

The Es'hail 2 satellite has been launched and has entered geostationary orbit. This satellite has TV transponders, and something new for radio amateurs : a 2400MHz to 10GHz transponder in geostationary orbit.
The coverage should be all of Europe and Africa, to the East into India, and to the West a tiny bit of Brazil. Even Antarctica should be within range, if some of the research stations are within the footprint of the satellite I assume that some testing will be going on before the transponders go online, but this is an exciting time for radio amateurs interested in satellite communication.

First update:
The sat is in orbit and the amateur radio transponders are now operational.
They work very well. An SSB uplink with 60cm dish antenna and linear polarization produces a readable signal, a around 10-20 dB above the noise, as I have heard on the WebSDR.
On the spectrum of the WebSDR it is even possible to see the effect of running too much uplink power. This creates dark bands on the waterfall spectrum, and sometimes audible attenuation of received signals.
Frequency stability is the main problem with the uplink and downlink, for obvious reasons.
As a first attempt at working through the satellite I was thinking of purchasing a relatively low cost up converter from SG-Labs in Bulgaria for the uplink. This has about 500mW output, and with a 1W power amplifier at the feed point it should be possible to be heard via the transponder with SSB, and certainly with CW.
A fun experiment would be trying FT8 or JT65 with a much lower ERP.
Receiver side will be with a low noise satellite PLL LNB, possibly with a later modification of the reference oscillator.
I have decided to take time to attempt a simple up converter made of Chinese modules available on eBay.

Second update :
10GHz reception converter (arriving shortly) for downlink:
LNB suitable for OSCAR 100 reception has arrived. Octagon OSLO (PLL controlled Local oscillator.
Initial test will be done with the unmodified LNB.
It will need a modification for better suitability: External reference oscillator, preferably a TCXO, at a stable temperature, fed through the cable (disconnect the internal XO, info on the Internet).
Alternatively, for even better stability, a GPS locked oscillator could be used.

2.4GHz uplink:
Modules from China have been ordered :
1) Synthesizer : ADF4350, covers 137MHz - 4.4GHz Needs to be programmed. There seems to be a library for the Arduino microcontroller. Price per unit about 10 GBP
It turns out, looking through my stock that I have a single AD4350 and an Arduino UNO, so it is possible to start the process of building the local oscillator for the transmit converter. Now for learning some Arduino (variant of C) coding.I ordered a few mor ADF4350s as spares in case I have an accident of destroying the one I have.
Learning the coding could easily take the time needed for the stuff from China to arrive, unless I find a suitable program that just needs a simple modification. No need to invent everything all over again.
2) Passive up (and down) converter modules (also to have some spares). Price per unit about 8 GBP
3) 1W amplifier modules. Price about 10 GBP
I ordered a few, because I have a long run of cable, and would like to have one as output for the indoor unit, and one at the antenna feed point.
The official uplink power for the satellite is regarded to be about 10W into a 60 - 90cm dish. However, an output of 1W appears to be sufficient to make SSB contacts via the satellite, as I have heard several stations with similar set-ups via the WebSDR.
4) Double balanced RF mixer (down conversion) Price about 5 GBP.
Why a down converter ? Well, it is a simple addition adding receive functionality, creating not just a satellite uplink transmit converter, but getting a 2400MHz transverter, making simple two-way contacts possible on that band, albeit with low power. It will most likely be possible to make some local contacts, and possibly a bit more during tropo conditions.

I already have some amplifiers capable of an output of about 100mW on the band, more than sufficient drive for the 1W amplifier mentioned above. Those were priced about 5 GBP.
One more thing is needed for this to work without making interference. A band pass filter.
I need to look into the best solution for that.

If (when) I succeed in getting a signal through the satellite, the adventure begins. Optimization of all aspects of reception and transmission, and experiments with very low power. I was told by OZ2OE that he had heard a Dutch station running a 75mW CW signal into a 1m dish, and it was audible at 539.
Also, running 1W into a much smaller antenna is an idea I want to play with. I have a tiny Wifi yagi, 40cm boom length and 17 elements. That will be an interesting comparison with the 80cm dish I have, and a good comparison. Constant gain from the yagi, and seeing the different improvements to the dish and feed.

I see quite a bit of experiments and lots of optimization in my future.