Other method to inject external REF into LNB!

  • DF1QE Yes, I also built these push pull oscillators ('Gegentakt') at very young age for FM with ECC85 etc. Output was ca. 5W.
    Here (i.e. Hilversum region) called that oscillator a 'Kalitron' and we used 'Butterfly' capacitors from Philips army equipment.

    About the injection of the LO. A friend of mine, who designed LNB's for Philips in the 80 - 90's told me about the symmetrical injection and its positive effect on phase noise in NXP (former Philips) (TFF) chips.

    Based on this information and the fact I was donated a LNB with a NXP chipset I tried to convert my friends hint into a practical solution.

    Glad that it's of use for/to you : -)

  • Sorry Remco, I cound not resist this wordgame in the subject :)

    So Folks, here's a new (?) idea for injecting an external TCXO, which I successfully tested.

    Just leave the x-tal where it ist, disconnect the LO out pin of the PLL-Chip, and inject the external LO at the x-tal.

    I tested succsessfully this method with 2 Diavolo Twin V2 and one Diavolo Quad V2.

    This method has some advantages :

    you don't have to remove the x-tal

    you can easily go back to internal LO (original)

    the external LO signal will be filtered by the x-tal

    easy to implement

    so many advantages :)

    The value of R1 in my shematic determins the sensivity and optimal level of the external LO.

    With 820 Ohm at a RT320M chip it is optimized for 1Vss LO input.

    For other levels you should make your own experiments.

    The value of R2 in my shematic depends on which impedance and coax cable you use.

    68 - 82 Ohm is fine for 75 Ohm systems.

    47 to 56 Ohm is fine for 50 Ohm systems.

    But this value is actually less important.


    This method is not useful for (older) NXP chipsets.

    Therefor you better take a look at Remco's Subject : other-method-to-inject-external-ref-into-lnb


    Armin DF1QE

    I still know at which end to grip the soldering iron :)

    Edited 2 times, last by DF1QE ().

  • DL5MGD Hi Andreas !

    Uhhh, not a good idea ...

    The x-tal needs this capacitive load to work on the right frequency.

    And actually the coax input needs a termination, at least a nearby one ...

    Ok, if you didn't implement the resistors, you have the high capacity of the coax on one side of x-tal,

    but then you need at least half the capacity (6p) on the in pin to get the x-tal in a good resonance.


    Armin DF1QE

  • PA3FYM : Hi Remco !

    That's why I wrote "new (?)"

    But ok, let's say : you won :-)

    Anyhow, there is some difference.

    As I saw in your Twitter picture, you just disconnected one side of the x-tal and directly connected it to the F-socket.

    I don't think the x-tal works optimal this way.

    I did some experiments until the x-tal was in optimal resonance.

    In the Diavolo quad I modified, I even mounted a variable C to "tune" the x-tal.

    But to all of you who did that before : Why did noone post it here ?


    Armin DF1QE