My LNB modification nightmare...

  • Hi,


    it never happens as you expect....


    In the last week of August i got my "DuoBand-Feed 2,4 / 10 GHz - LNB" from BaMaTech (https://www.bamatech.net/anten…-feed-2-4-10-ghz-lnb.html). A real great piece of hardware.

    Before buying, i asked Markus for the LNB type to know how to modify for a TCXO and he told me, he uses a Opticum Red LSP-02G.


    So i was happy so far, the LSP-02G seems to be easy to modify. All parts on top, everything accessible and well known (see http://www.pabr.org/radio/lnbl…lnblineup.en.html#LSP-02G).

    Then the parcel arrived and i opend the LNB housing which was labeled "LSP-02G". But something was wrong, the LNB doesn't look like a "LSP-02G"....Damn! (BTW: i don't blame Markus for that, just my bad luck!)


    I opened the LNB housing and removed this flimsy LNB-PCB which was sealed with silicon rubber and....teared off the pad where the F-connector was soldered. So this PCB couldn't be modified anymore.

    I analyzed the PCB and it seems, that Opticum now sells the Opticum Red "Robust" as "LSP-02G" (see http://www.pabr.org/radio/lnblineup/lnblineup.en.html#Robust).

    The differences of the both boards was the solder mask (green at the LSP-02G instead of black of the Robust) and the code marking of the transistors. The block diagram can be found at the attachments.


    The LNB is a "two chip" design with an integrated power supply chip D4202 from Silicore for all LNA stages including a 22kHz detector and a NXP TFF1015 PLL/mixer. Unfortunately there is no datasheet for the D4202 available. Hopefully the external clock can be fed through XO1 of the TFF1015.


    The modification of this type of LNB seems to be really hard. The is only a little cavity on the bottom side where the crystal is soldered, where modification parts can be placed.


    So i decided to order two Robusts and give it another try to do a TCXO modification.

    The silicon seal can be removed very good with "SOUDAL Silicon Remover". If everything has been cleaned (PCB, housing, ...) the crystal has to be removed carefully.


    I put some kapton tape at the area of the crystal and placed a STM L78L33ABU there. The supply can be connected to a via near this area.


    Next i will place the TCXO on top of the regulator and connect it via a 1nF capacitor to the crystal pad which leads to XO1 (the block diagram of the modification isn't that detailed!). In progress, stay tuned...


    The crystal cavity of the LNB housing is isolated with kapton tape to avoid short circuits of the extra parts with GND.


    As an intermediate result i would say, that it would be possible to modify this LNB for TCXO usage, but it is nothing for the feint-hearted. You need a good microscope, good tools and a steady hand.

    It would be easier to insert a series resonant circuit for external clock sourcing (which I do not prefer) or to use an external correction.


    The modification hasn't finished yet, so I am curious, if the LNB will work again.


    Just for your information...be warned :-D


    73, Thorsten

  • DD0KP

    Thank you Heiner. Unfortunately the datasheet gives no information about the maximum output current of the 5V PLL supply. If the 4202 could deliver additional 50mA, a much smaller 5V/3V3 low droupout regulator could be used instead of the fat SOT89...

  • I would not put the TCXO on top of the regulator. That is asking for trouble as the regulator temperature variation will be high. Much better to keep the TCXO indoors in a thermally stable environment. Temperature compensation is good, but not perfect.


    Mike

  • Hallo,

    I just modified two Rocket LNBs. Carefully cut the Silicon to remove the board. The only pin to resolder is the output on the right side. Because of leadfree soldering it is not easy. Don't worry losing the crystal pads. They are not needed furthermore...

    Remove the crystal on the backside and place the TCXO 25 MHz. The new Chip is upside down with double sided adhesive tape and connected twice to GND. The output is connectet to one former crystal-via. Remove the capacities from Pins 11 and 12 of the TFF1017 Mixer/PLL. Connect Vcc somewhere to the +5V Bus (Pin 16 of the Mixer). The +5V is generated by the detector chip ZXNB4202. Take care not to squeeze the +5V red wire when closing the case.

    Enjoy the better accuracy. Don't mix it with the drift. This is done with SDR Console.





    73s

    Andreas

  • G0MJW

    I agree. This will be an option for me, when the new AMSAT downconverter is able to deliver a reference clock for the LNB. In this case i will modify a second LNB-PCB with that resonant clock feed which will suit perfect to the existing cavity of the LNB.

    In the meanwhile i will use the more or less good solution with an internal TCXO. Currently this gives me the most flexibility with a suitable result :-)

  • DL9SEC  
    The output of the TCXO is connected to Pin 11 of the TFF1017. Is there any recommendation, 11 or 12? I didn't care, it works.

    Interesting, both TCXOs, I assembled in two LNBs give an endfrequency shift of around 3 KHz. Probably it is useful to put an adjusting voltage of 1,5 Volts to AFC-Pin 1 of the TCXO.
    My goal was to test the modification of the LNBs. At least, a modified LNB is ok to play around with, doing demonstrations using scanners, little dishes and so on...

    Because of the internal 5V, there is no need to operate the LNBs with 12V. If vertical polarisation is ok, reduce supply to avoid dissipation and drift.

    73s

    Andreas

  • "When the hurlyburly's done..."


    So finally the TCXO is stacked and in place (see the attached images). The PCB fits snugly into the housing again. There should be a millimeter or so between the highest point of the stack and the cavity's edge.


    Inspired by DL5CN i omitted a coupling capacity between the TCXO and the PLL. I don't know the internal architecture of the oscillator driver circuit, so i used XO1. Good to hear, if XO2 will work too, so there is a plan B :-)


    Attached you'll find the final modification block diagram with some details around my TCXO.


    Next, I'll check if my "FrankenLNB" could be revitalised. Keep your fingers crossed :-D


    73, Thorsten

  • Ok, after finishing the modification...how to test if it was successful without climbing on a roof or buy another dish?


    I discussed this with my former student colleague Jochen DG1SFJ and he told me about some 24GHz experiments and a similar problem. He found a very simple but very good idea at https://ea4eoz.blogspot.com/2012/10/24ghz-old-way.html?m=1.

    A common 1N4148 as a multiplier to generate a 10GHz signal. So the 23th harmonic of 437,069792MHz should generate a signal at 10489,656MHz and at the LNB IF output at 739,675MHz.


    So i took my good old FT-817 at 1W output, a 20dB attenuator and a LL4148. The attenuator matches the RF output of the FT-817, and the diode gets about 10mW which should be enough to make it scream at 10GHz (see the pictures).


    And....TADAAAAA....the expected signal only a few kHz aside.


    So i assume, that the modification was successful so far. I did the same test with an unmodified LNB and the signals are compareable.


    Now it can go on :-D


    Edit: Should be the 23th harmonic...

  • A common 1N4148 as a multiplier to generate a 10GHz signal. So the 24th harmonic of 437,069792MHz should generate a signal at 10489,656MHz and at the LNB IF output at 739,675MHz.

    I did the same without Diode on 437.0700 MHz to test my RX-converter and LNB. Tadaaa - I heard the 24th harmonic on 10489.680 MHz. ;)

  • I'm not sure what the issue is here ?


    To drop the supply to the TXCO just use a resistor, nothing more.


    Changing the feed is either luck or a test gear solution. I measured mine, but then I have a calibrated HP8970B.

    Dave Cawley | Ex G8EAO & G6ANG/T |
    (when you had to have a separate TV license !)

  • G4IUG No issue, Dave.

    Just want to be sure not to blow the D4202 until i know if the TFF1015H PLL runs with that TCXO.

    I couldn't find anything about a modification of that "new" LSP-02G or a TCXO modification with a TFF1015H.


    DL5CN gave some hints with a similar two-chip solution and it seems to work and the D4202 can supply another 50mA. So i can go ahead with a better, less power dissipating modification as you suggested...


    73, Thorsten

  • Hallo,

    I don't see the need to feed LNBs with 12V when only one polarisation is used. It can be reduced until the drop voltage of the regulator is reached. This helps to avoid unnecessary dissipation.

    73s
    Andreas

  • I don't see the need to feed LNBs with 12V when only one polarisation is used. It can be reduced until the drop voltage of the regulator is reached.

    Very true, however I don't get the virtue of low dissipation ?


    Dave

    Dave Cawley | Ex G8EAO & G6ANG/T |
    (when you had to have a separate TV license !)