Octagon Oslo PLL LNB with external Reference

  • Good morning,

    I am using here an Octagon Oslo Optima LNB and a sdr usb stick. As reference frequency source for both devices i made a small circuit from parts lying around which delivers 25MHz for the lnb and 28.8MHz for the sdr stick disciplined via gps

    Unless you have special requirements it isn't really required to discipline the SDR stick. Those "1 ppm TCXO" sticks e.g. from RTL-SDR.COM are stable enough for typical amateur communications use. Also because the stick is indoors at room temperature, vs the LNB that is out in the weather.

    But of course it never hurts to tinker and experiment :-)

  • Thanks for your replies. The second frequency (28,8MHz) for the sdr is only for fun or if one have no "good" sdr stick. My primarily goal was to disciplin the lnb frequency with off the shelf components and programming the gps receivers 1pps output to 1MHz for a faster and more easily to construct loopfilter.

  • DL5MGD : I also experimented with a Ublox, but it was a 7M module. As PFD frequency I used several values, also 1 MHz and fed that to an ADF4351 PLL. On the frequency counter here everything looks nice, but when I listened in SSB to a carrier from a signal generator (and ADF4351 as LO for an RX-converter) there was still jitter audible, despite I used an integer value for the output reference frequency.

    This was confirmed when I saw the signal on the oscilloscope.

    You use the M8N, and I haven't digged into the datasheet yet, but does this module generate a frequency without jitter and/or did you 'dejitterize' that signal by using the OCXO in GPDSO mode?

    Further, it would be nice if there is a version with 1PPS (1Hz) because that is what most GPS receivers are 'learned to produce' ; -)

  • Hi Remco,

    there is some jitter. The jitter only disappears if you use even divider values, best in power of 2 (eg 2,4,8,16 and so on). Neo 6, 7 and 8 use a 48 MHz oscillator to generate the output frequencies at the 1pps pin. The pll loop together with the ocxo removes this jitter.

    A version with 1pps will result in an extreme long pll filter constant. A pll would need to have a bandwidth considerably less than 1Hz, meaning that the loop would only lock up after many tens of minutes or hours. So the 1pps approach can not be used in a simple design like this. More information can be found eg here http://ve2zaz.net/GPS_Std/GPS_Std.htm or here http://www.ik0otg.net/index.ph…e&id=53&Itemid=59&lang=en

    vy 73


  • In commercial GPSDOs a "digital PLL" is used where the VCXO is driven from the output of a D/A converter which is set by a processor which performs the PLL function and loop filtering.

    With this setup one can remember the last D/A value used across powercycles, so the lock is obtained a lot quicker than with a fully analog design that has to find the correct control voltage from scratch after powercycle.

    Even with that, those units often require 30-60 minutes to fully stabilize.

  • Exciting times we are in.. I wish I would be young again.. :P

    ... rings Peter in 1983

    "Peter, do you want a receiver that will receive from 25 to 1700 MHz, all-mode, will decode sat telemetry, and give a panorama overview of the entire transponder? It will be 30DM. and, a 3cm converter to go with it for 15DM"

    I think you would have considered that two zeroes were missing :-)

  • For those interested in making the G0MRF 739MHz to 144MHz downconverter, 739MHz to SDR and LNBF driver, I still have pcbs at US$10 each plus US$3 to cover shipping fees.

    Do not hesitate to contact me at lgnachif@gmail.com


    Luciano PT9KK

  • Probably now most of those RTL2832 devices are used for generic SDR purposes instead of the original purpose of DVB receiving with FM/DAB as side function.

    It is a bit sad that all the production volume now goes into these devices instead of a device optimized for SDR (some more bits in the A/D, no DVB decoder) that we probably could have got for the same price when produced in this volume...

    But anyway, it is incredible technology for the radio enthousiast.

  • Hello,

    now I tried to reduce the Quartz Oscillator Frequency of the Octagon PLL LNB to get an IF of 9200 MHz. (Mix of the 10.5 GHz band to 23cm). Unfortunately, the frequency reduction only works down to about 25.9 MHz. At lower frequencies of 25.9 MHz the PLL does not lock anymore. :-(

    The external frequency reference is fed in via 50 Ohms and 1 nF. Quartz is removed. See image.

    Any idea what else I could do to get the PLL locked at a lower frequency?

  • There are three versions of the Octagon in the market. You have the second generation with the RDA3565 this version is limeted and your result is correct. The first generation has a RDA3560 and the third one uses R820T from RAFAEL Micro. Both lock down. Hope this help you. Heiner..

  • Ok, that's the reason. Thank´s for the info! This is now indeed difficult to get the right one from the LNBs on offer. Nobody points to the used chip ....

    73's de Robert

  • Hi Herbert,

    LO and High band is switched using a 22 kHz tone, but that's not required. Without tone the LO = 9750 MHz.

    The voltage level is required for the Polarization:

    11...14 Volt = (V)ertical --> NB Transponder

    16...20 Volt = (H)orizontal --> WB Transponder

    So for the NB transponder you just apply 13V via Bias-T and that's is it...

    73s Peter

    Peter Gülzow | DB2OS | AMSAT-DL member since 1983 | JO42VG