Demo station for AMSAT-UK colloquium

  • A couple of months ago I volunteered to provide a narrowband transponder demo station for use at the AMSAT-UK colloquium in October. Now I have to convert this collection of parts into something suitable.

    I have chosen to use a single 0.8m dish with POTY dual band feed for the antenna. I had previously purchased a 20 / 25 Watt Spectrian amplifier from 'Pyro Joe' via e-bay. The final major component is an upconverting mixer to take 432MHz in from an IC9700 and to output a signal on S Band. This will be designed around some available parts and built on a PCB to be ordered from China. - What can go wrong !!

    Hopefully this short series of posts may help people trying to assemble a station of their own.

    73 David

  • Hi Mike. For a demo I thought it's important to not 'scare people away' by using a lot of high tech solutions, e.g. GPS locking, which athough very good, may be a little intimidating for the HF contesters etc who will be attending the RSGB event that runs alongside the AMSAT-UK colloquium. I have fitted the Octagon LNB with a (new) TCXO and thought it could run into an SDR for display on a 32in monitor while also having the 9700 receiving an output on 145MHz via my 739MHz to VHF converter. So 2 methods of RX.
    I didn't know you had a board for locking. I look forward to reading about it here when you have one built and operational.

    In the meantime....see next upconverter PCBs are finally on their way from China. However, that turned out to be a time consuming process due to me using outdated software. Thanks for the recommendation. They got the job.


  • TCXOs....

    - the board for locking was designed last year but I assumed the concept flawed with my D75F oscillator being so bad once I had built one. I have started again now I know the units supplied were faulty, but it's made to take other makes of TCXOs. It also combines the biasT and diplexer.

    I have a fair number of microwave directional couplers - 20 dB, 1-2 GHz. I discovered you can use them as an emergency diplexer for the 25MHz if you don't mind losing 20dB of LNB gain. Prevents overloading the spectrum analyser in the above video.


  • On with the story.... After reading many datasheets, considering options and getting valuable advice from this forum, I settled on the block diagram below.

    The upconverter uses an external local oscillator source at 1967.50 MHz, generated from an ADF4351 evaluation board. These are available on e-bay. It is the same board found in the RSGB project "Easy 100" by Giles Read, published in the June 2019 edition of Rad Com.
    Programming data for the PLL is held in an 8 pin PIC12F629 and is transferred when power is applied.
    The LO signal is amplified by a Mini-Circuits ERA1-SM which drives the LO port of the mixer with 7 to 10mW.
    RF input on 70cms passes through an attenuator and then to the input side of the mixer.
    The 2.4GHz output from the RMS30 is applied to an attenuator (not shown) to terminate it at the LO and image frequencies.
    A SF2124E bandpass filter removes unwanted products, then the signal is amplified by another ERA1.

    A second bandpass filter is included to make sure the signal is clean, - This is important if the board is used from 144MHz to 2400MHz

    The final amplifier is a 10pin surface mount IC from Qorvo. The RFPA5208 has high gain (40dB) and an output power of at least 500mW. This device was selected because it can be used with the popular 8W (really 3W) output Wifi amplifiers, which need about 170mW for full power. Or, it could be used with the SG Labs 20W amplifier module which requires 500mW of drive.
    The PA has a DC Voltage output from an RF detector indicating the amount of RF being produced. The final block is a low pass filter to remove harmonics. The TDK product selected has a maximum power rating of 3W so is comfortably within its limits.
    The board requires 12V for the ERA1 MMICs and 5V at about 500mA for the PA. A 3V regulator is included on the PCB for the 12F629 and for the device enable pin on the PA.

  • Alas, I am an analogue man in a digital world ;(

    With a circuit diagram completed, I set out to design a PCB for the upconverter. The size was determined by a handy looking BGA heatsink - 43 x 41mm
    After several days of spare time, I sent the design to China. Unfortunately 'Betty' from said they were unable to process my Gerber files as my rather ancient copy of Proteus used an old format of Gerber output. A few e-mails later and I realised the whole design had to be thrown away and I really needed some upto date CAD software. At the suggestion of Dr Chris Bridges at Surrey Space Centre, I downloaded a copy of 'DesignSpark'. After the usual frustration of learning how to use new software, several more late nights were spent doing the same work I thought I had already completed. There probably are mistakes... but at least 'Betty' was happy and my RS-274-X formatted gerber files were accepted.

    Now, while I am waiting for DHL, It's time to start work on the Spectrian amplifier and find out if it is functional.



  • G0MJW ...

    i know why you prefer this solution.... :)

    IT WORKS ON NARROW AND WIDEBAND ... (look at FUNKAMATEUR MAGAZINE 12/2018) ... an article was made about ...but for GNURADIO and LINUX...

    And additiontional dont forget the LIMESDRs ...SAME SETUP other SDR MODULE... and DONT FORGET THE FILTERS...

    IN DUBUS 04/2018 you can see another solution by W7FU with ETTUS B210

    73 de Michael dg0opk

  • Yes. DesignSpark integrates the schematic and layout pages so you know if there are connections missing.
    One feature I really appreciated was the parts library. If a part was missing, as most RF parts are, you can download a file from RS Components or Mouser with the PCB pad layout, the schematic symbol and a 3D model. If the file does not exist, then you can request one be produced within 24hrs. I requested one for the low pass filter and it arrived back in 3 hours on a bank holiday weekend.
    Chris also suggested KiCad.

  • At last, a weekend and free time.
    I decided to have a look at the Spectrian amplifier I bought a couple of years ago from 'Pyro Joe' on e-bay, while QO-100 was still being assembled. The board comes mounted on a 1/4 inch (6.25mm) thick Aluminium heat spreader, so my first job was to mount it onto a decent sized heat sink. Being of US design, M4 screws were just a little too large to fit, but 4BA were perfect. A little drilling and tapping later and the the amplifier was ready for a DC test. The amplifier has a 14 pin connector with 4 pins for V+ (12V) 5 connections for ground or 0V , one ground to transmit pin, another which needs 24V input at 25mA and a pin which has a DC output voltage from a PA temperature sensor.
    With 12.2V and 24V @25mA applied, I connected a 30dB attenuator to the output and 50 ohms to the input. When the enable pin was grounded the supply current increased from a few milliamaps to 6.6A. - If this device is Class A, that should be fine. Now it was time for some for some RF input


    RF testing used a signal on 2400.50MHz at levels from -6dBm to +11dBm.
    With 12.2V and 6.6A standing current, I hoped that a class A amplifier would not need much more current across the range of input powers. Unfortunately testing shows the amplifier draws higher DC currents as you increase power. Clearly the PA stage is not class A after all. I wonder if efficiency can be improved. Overdriving this amp could be very expensive as it does not appear to go into compression...It just consumes more DC power and it can't do that forever. :(

    RF Input dBm
    RF Output Watts
    Apx. DC Current
    -6 0.625 6.6A
    -3 1.2
    0 2.25
    +3 4.2
    +4 5.2 6.8
    +5 6.6 7
    +6 8.2
    +7 10.9 8
    +8 13.5 9
    +9 17.5 10
    +10 21.5 11
    +11 26 12

    73 David G0MRF

  • Hi Mike.

    I had a similar reply on Twitter from M1SJE. He recommended a piece foil on the output line as it turns 90 degrees along the edge of the enclosure. I think there is a photo of that in one of the other topics on the forum. I had better check that NEC transistor and find out what the maximum limits are before I start exceeding them.

    So, I'll cut a couple of rectangles of foil and glue them to the end of a cotton bud, then I can prod it onto the output lines and see if I can make an improvement or two.............

    Well, the tuning aid worked. After checking all the way from the PA drain coupling capacitor to the isolator, the best place is as suggested by M1SJE and others on this forum. A fairly large piece of foil is needed at the 90 degree turn near the edge of the enclosure.

    Without foil......14.5W with foil 22W. There is an increase in PA current as well, but overall I'm saving a little over 1 Amp for the same power output. I suspect if I changed the PA voltage from 12.2 to 13V it would be close to 25 - 30W at 10A supply.

    However, I will keep it at just over 20W.....leaving a little in hand for the odd mistake.

    Tuning aid with foil glued in place pictured left. Added Copper foil right. - Note PA bias pot is very sensitive. Now set to 5.5A

  • TIme for a little work on the antenna system. POTY feed kit / Octagon OTLSO with TCXO / Mix Digital 0.8m dish and finally a nice tripod designed to hold a loudspeaker with a max weight of 50kg. ( Care needed, probably Chinese kilos)
    The POTY kit went together very easily. Only the dielectric lens needed a little sanding before it would fit into the 22mm Copper waveguide.

    The Mix Digital dish came with a LNB holder designed for a standard 40mm diameter LNB. However, it now needed to hold the 22mm pipe. There are many ways of making a suitable spacer....40mm diameter with a 22mm hole in the middle, but finding a short length of 40mm nylon or other waterproof material was difficult. Even 40mm diameter wood used for handrails came in 1.8m lengths, but I only needed 50mm or so.

    Eventually I found a craft supplier on e-bay selling 40mm diameter hardwood, each piece 40mm long. A few Euro bought a pack of 4 pieces with free delivery.
    A friend drilled a 22mm hole and then carefully cut vertically along one side with a band saw. The small amount of material removed with the saw allowed the LNB holder to clamp the wood to the pipe....perfect. A couple of coats of varnish and I was ready to go.


    A demo at the Cray Valley Radio Club gave me the opportunity to test the feed. What a pleasure to not have to align two separate dishes. Signals with 3W were not as good as my home system, but good enough for a QSO to impress the 28 members. Looking at the feed, I'm probably a couple of centimetres off the correct focal point. Time to optimise that later. Next step :- The upconverter boards have arrived..........

  • @G0MRF David..I can highly recommend using the lens of an venton lnb. In comparison to the lens you use the mer will be much better.

    I had a mer of 2.9 with the lens you have and with the venton one I got 5.2 !!!

    If you do not get a Venton one I have here one left and can send it to you.:)

  • DD0KP Hello Heiner.

    I have not glued the lens in position, so it can easily be replaced. With my 0.5m RX dish I can see the transponder noise and I hoped that the extra loss in the Nylon POTY lens would not be very noticable as I have extra gain from 0.8m.
    However, your MER going from 2.9 to 5.2 suggests otherwise. I'll move the focal point and measure the change. If there is a significant difference, perhaps I can arrange to buy the lens from you. Does it fit 22mm tube without much machining?

    73 David

  • no the total lens are two parts and it the inner one fits into the 22mm waveguide.. I fixed the inner one just with scotchtape and the outer one with superglue to my 110mm waste endcap, which you can buy for less than 1€

    It looks like this:

    So I highly recommend this lens. The improvement is dramaticly.

  • The upconverter PCBs have been here for a couple of weeks but finally it's time to see if they work and find out if I have made any errors in my rather rushed layout.

    The design uses a PIC12F629 to control an external PLL source that is based around the Analog Devices ADF4351. The same board used in the RSGB Easy-100 project.

    The 1967.5MHz local oscillator is amplified with an ERA1 which feeds +7 to +10dBm to a RMS30 mixer.

    Input to the mixer is at 70cms and the mixer output is 2400MHz.

    The board has 2 bandpass filters to remove any local oscillator leakage and the image frequency around 1535MHz.

    The wanted signal is normally amplified by an ERA2 then a RFPA5208 from Qorvo up to a level of 500mW.

    In this particular build I only need 10mW to drive the Spectrian amplifier to 25W, so the Qorvo PA will be replaced by a second ERA2

    Pictured below is the blank 43 x 41mm PCB with a matching BGA heatsink.