What TX amplifier do you plan to use?

  • Now, that makes sense.

    i saw that it is really important WHERE you place the copper foil. ;)

    And because of that you must try and move the foil till you are satisfied with the results. Changes from one PA to another.

    vy73 DB8TF

    • Official Post

    The term here is 'snowflaking' an amplifier (german term: Fähnchenschieben) and I consider it an art (and those who master it 'witches' .. hi!).

    Any amplifier has a complex impedance it wants to see on both input and output to produce max. output power. The striplines on both sides transform the complex port impedance to our known and loved 50 Ohms system impedance.

    Of course, real life is much harder. Matching the input influences the output matching and vice versa.. and matching for max. output is sometimes not identical for matching for max. gain.

  • I have always believed that equipment debugging beyond GHz is a metaphysical art.Seemingly random stacks of solder joints and burrs are part of the tuning loop.But there must be a theory that can explain it. I just don't know.

  • Hi,

    I'm using a TeliaSonera cellular amp which I stripped from an old base station system. It utilizes a couple of Ampleon BLF8G22LS-160BV with very poor efficiency. It is almost impossible to de-solder anything because the PCB is very thin and the 6 mm thick baseplate absorbs heat too well. Replacing the output SMP (P like Pee) to SMA was unsuccesful. X(I Just damaged the pcb.

    Now I'm taking power out from the circulator directly with a teflon coax.

    The pair of fets have very different biases. The other one warms up and draws a lot of current, about 50 watts of idle power warms up the shack.

    Nevertheless the system delivers 70 watts output (perhaps more).

  • Maybe you can post some photos?

    vy73 DB8TF

  • Sure! In the picture you'll see my homemade "connectors" using dual row, 2mm spaced pin array. Covered with hot melt glue. Original SMP output was located on the right of the circulator, ugly replacement SMA soldered instead. :) Could not find a SMP mating connector so I let it go.

  • Maybe you get better results when you remove the bandpass(es) after&before the driver as well as the circulator after the finals?

    I doubt they are designed for 2,4GHz.

    Better use a good and strong bandpass after the PA.

    Check also the datasheet of the Anaren? Xinger in front of the finals, if it is usable on 2,4GHz.

    On the Input pads of the finals you have enough space for "flaking" to squeeze out the last dB of that beauty ;)

    vy73 DB8TF

  • While this is true for small signal amplifiers, power amplifiers don't quite behave the same as far as large signal performance is concerned. There will generally be an optimum load impedance (ie output tuning) that sets the optimum power and this is not affected (at least at the fundamental frequency) by variations in input tuning.

    The procedure that I find works pretty well is to start with the output tuning as is, then tune up the input with a small signal (eg 10dB below what you would need for max output) for maximum gain. Having done that, move on to the output tuning with more drive, say about 3dB below that to give saturated output. Aim for max power or max efficiency as you want (max efficiency is often not far away from the maximum small signal gain tuning point). Once happy with the output tuning, increase drive further and repeat (small improvements now only likely).

    Once the output tuning is optimised, move back to the input with drive backed off about 3dB from max saturation. Small improvement in gain may be possible.


    Charlie G3WDG

  • Just did the very first tests into the dualband patch antenna (G0MJW, PA3FYM, M0EYT design). A horrible but funny configuration:

    Since I'm waiting for the upconverter, my only source is an R&S 2,4 GHz generator which does +13dBm.

    This is followed by an old MT2,3Z1W PA (1W) which I used years ago for the camera of a model plane. This amp work on 2,3 GHz, but not on 2,4 GHZ, but still gives me +7 dBm.

    This is followed by a spectrian amp from ebay. This is optimized for 2,33 GHz, but on 2,4 GHz is has only -1dBm, so quite usable. After setting the quiescent current to 600mA for each Mosfet, it has +15 dBm. (And yes, the other commenters are right, this amp is a heater, but works).

    This gives me 13 + 7 + 15 dBm = +35 dBm

    This is fed into 70m Aircell-7 cable going to my 1.2m dish. This cable has 25dB attenuation.

    Overall I get 35 - 25 = +10dBm / 10mW.

    The signal is visible in the waterfall and loud enough for a CW qso. It's really impressive how little power is needed for QO100.

    The final configuration will have the Amsat upconverter +20dBm followed by the spectrian makes in total +35dBm directly on the dish ,which should be enough for SSB.

    That test proves also the very good performance of the G0MJW, PA3FYM, M0EYT - Patch Antenna.