Just playing, trying to generate a clean 2.4GHz signal, advice needed please. Long post.

  • Do not worry! I am only in to a dummy load and sniffing into my Spectrum analyzer at present, still some time from going to air.


    I have been playing with my Hackrf generating a 1000K SR DATV test signal.

    This is my line up at present

    Hackrf, to RF2126 amp to MHL21336 Amp through 2.4GHz filter to 30 dB coupler and dummy load with the sample port to the spectrum analyzer.

    The line up:-

    Everything for testing is from a 12 Volt PSU so the the RF2126 amp has a DC converter down to 5 Volts and the MHL21336 has an DC buck to 26 Volts.


    This is a spectrum display at present with 10 MHz per division Horizontal and 10 dB per div vertical, so I am looking at present with around 50 dB down on any spurious signals.


    This is the same view but now at 2MHz per division.

    I think I may be able to reduce the spikes there as they are the sample rate used on the Hackrf so I believe increasing it from 10 to 12Mhz may reduce the spikes to an even lower level.


    I was trying to use this small amp stage after the Hackrf,


    but, the frequency response is so wide that it amplifies any and all spurious outputs from the Hackrf and helps pass them to later stages so unless I build another ID filter, (Which I may well do!) then I will not use it in the line up.


    But I still do not have anywhere near enough signal going this way to actually drive the main PA into any sort of power level.


    I put the bisonelectronics amp on the output as it was for a test and was disappointed. The shoulders increased to at best -40 dB at worst -35 dB below peak signal.

    A 1000SR test signal



    A simple test shows that the amp has around 20dB increase


    But we are still taking less than 1 Watt!;(


    That is a far as I got tonight playing with it all across the bench I need advice on how to proceed and what are realistic figures I should be expecting on shoulders, yes i have seen some saying -35 dB down are OK and some saying that is not good enough?


    Is anyone using a bisonelectronics amp as the final stage and how does it perform and what drive level do you need to get to?


    Adrian

  • More prep work going into things


    I made another ID filter for 2.4GHz and tried to neaten things up by putting some of the gear into a box. This is the present stage.

    The box layout


    The Bison amp has now got a cowl over the heat-sink fins to help cooling for a fan to be directed in at one end. Only using one of the two amps, the other is there in case of mishaps!



    Next is to get an outside enclosure to house it all in and fit the PSU's and control.


    Adrian

  • Well sat some point during the testing the RF2126 1 Watt amp fried it self, so possibly why so little output. Back to playing and waiting with some replacement chips on order.

    There are some issues with the chinese RF2126 design:

    1. Voltage divider for Vpc is inadequate, especially @6V; use poti and set 180mA idle current carefully!

    2. Capacitors and layout: See attached snap, I placed capacitors with values given in the data sheet. Removed/replaced this 0 ohm resistor in the stripline with a capacitor with the datasheet value (3p3) and put a small copper strip to connect output.


    I achieved some 900mW @5V @2400MHz


    Frank


    Update: Due to some reason three other RF2126 fried. I give up with this amp ... :(