Posts by dl4yhf

Greetings all,

After playing with a Bullseye LNB for a few days, and discovering that it can be quite easily through a single-wire serial protocol (details at https://raw.githubusercontent.…Alignment%20Procedure.pdf), I wondered how wide the reference oscillator's tuning range is - because that's not specifified in the PDF. And, as they say (quoted from the manual):

>>

Quote

There are obviously commands to overwrite and setup the user calibration and issuing these will invalidate any warranty to spec. There is no reason to ever change these constants as the internal algorithms will be impacted corrupting stability and temperature dependance and once changed cannot be recovered, so please don’t fire random commands at the LNB in an event to play around with them you are likely only going to brick the device. We record all values referenced to serial numbers so if units are returned we will know they have been overwritten and the LNB will be recycled and not returned so be warned...

<<

So... "there are obviously commands to overwrite this and that" .. umm .. does anyone here know a bit more than the PDF linked above reveals ?

There are commands to increment / decrement the LNB's LO frequency, but they don't reveal if that controls the 25 MHz TCXO itself, or if it controls a synthesizer between the TCXO and the LO itself. Wild guess: If that's really a programmable synthesizer inside, how far can its frequency be modified ?

An "adjustable TXCO" itself certainly cannot be pulled that far.

Dreaming about reprogramming the Bullseye from the usual 739.x MHz IF to something in the 432 MHz range for /p operation without a PC ... but of course I don't want to "brick" it, despite the low price.

73,

Wolf DL4YHF .

Another item to add to the "tx antenna challenge" :

Successfull QSO with IZ2CPS (thanks Roby !) with -for transmission- a "double quad" antenna made from copper wire bent into an "8" figure, with copper-clad PCB (150 * 100 mm) as the reflector. No dish- just the double-quad as explained in Rothammel's antenna book. We had used this for a WLAN link spanning a few hundred meters years ago.

Now (with relatively slow, but human-readable CW) and 900 mW from a BU-500 tx converter running "barefoot" same old antenna spanned almost 40 Mm (mega-meters). No dish (for transmit), just the tiny double-quad antenna. Makes a nice 'reference gain' antenna for comparisons with various other TX-antenna tests.

Of course this little gadget (see attachment) is not sufficient for SSB, and I doubt the Chinese "8 Watt" (in reality 3..4 Watt) WLAN booster will make "dish-less" SSB QSOs possible. But I will see what the old homebrew 25 turn "WLAN DX helix" can do (without a dish, of course), and how it compares with a dual feed in a 40- or 60 cm offset dish. But that's another subject...

Cheers,

Wolf DL4YHF .

Not sure if 'jumpy behaviour' is the correct technical term but yes, I have experienced this myself with a Vivanco STL US3K-N "Universal Single LNB", wrapped with foam rubber and aluminium tape to slow down thermal drift and reflect sunlight (the sun didn't illuminate the LNB radome, only the side of the LNB).

I don't have a screenshot of the spectrogram, but the initial drift looked like a slow exponential decay, as expected. Then, for a couple of times, the frequency jumped by several hundred Hz within a second, i.e. not a wobble (*) but unpredictable steps, up or down in frequency, at random intervals. Similar effect when gently knocking on the LNB with a single finger, after which the frequency doesn't always (but usually) return to the old value.

I had posted a photo of the opened LNB - see "receiving system not based on octagon lnb" . The crystal is a through-hole device, but unfortunately on the opposite side of the PCB (not visible in the photo). Otherwise I had replaced it with another 25 MHz crystal already, to see if it makes a difference. But it will be modified by a 10 MHz OCXO anyway (multiplied by five then divided by two, filtered signal sent to the LNB via frequency splitter).

Cheers,

Wolf .

Excuses for 'warming up' this old thread, but maybe this helps someone with similar issues:

I have experienced this 'jumpy' behaviour of crystals in oscillators when their allowed drive level (in terms of RF power) was exceeded in homebrew QRP equipment. If the drift isn't slow and sinusoidal, the crystal may be overloaded because the two "capacitors to ground" in the Colpitts configuration are not properly selected, or (possibility two) the supply voltage of the oscillator (in this case the synthesizer) changes by a few millivolts, causing a change of the semiconductor's input capacitance by a few femto-farads (like a parasitic varicap diode).

Hoewever, so far none of the cheap PLL LNBs I have has exhibited this behaviour, so no there was no opportunity to see if modifying the Colpitts-C's near the crystal (to reduce the crystal drive level by a few microwatts) would cure the wobbling. I remember datasheets for 32 kHz 'clock' resonators even suggest a series resistor between the oscillator output and the crystal, to reduce the power dissipated in the crystal (optimum resistance depends on the crystal's individual Q factor).

There will be some tradeoff between phase noise (less noise with higher crystal drive power) and drift (less drift with less drive, less stress on the crystal).

Has anyone on the group unsoldered a crystal from an wobbling/jumping LNB and tried to use it in a *low power* oscillator configuration (like the one with two antiparallel Schottky diodes to limit the voltage across the crystal, as shown in older ARRL handbooks) ? It would be interesting to see if they don't wobble in such a configuration, e.g. with a UHF receiver tuned in CW to an overtone of the crystal to hear it drifting / wobbling / jumping around in frequency.

Well, some food for thought.

Hope to meet you on QO-100 with the BU-500 carried home from Friedrichshafen soon. But that's OT here.

Cheers,

Wolf DL4YHF .

Hope the 400 MByte attachment gets through..

here is a photo of the Vivanco "STL US3K-N 44719".

The 25 MHz crystal is, as already mentioned, on the lower side of the PCB, but after opening the lid (secured with 5 screws), it's easy to see which traces lead to the crystal, and instead of removing the crystal, those traces could easily be cut.

73,

Wolf .

Greetings all,

I just picked up an LNB for testing, and completely disassembled + reassembled it.
This one is currently sold at "Expert" (and possibly many other stores) in Germany:

Vivanco "STL US3K-N 44179 Universal Single LNB". Cost less than 10 Euro.

Contains a 25 MHz PLL (the crystal is on the BACKSIDE of the PCB, towards the waveguide). There is NO frontend-selection before the PHEMTs - only a few millimeters between the gates and the two probes (for H+V) coupling into the waveguide.

Even without adding a controlled heater for the crystal, it's even better (less drift) than my older "HD-Profi Single LNB Gold" (which had a 27 MHz crystal, but doesn't seem to be available anymore).

In contrast to a "Schwaiger Single LNB" / "LNB 1" (from "Marktkauf") which also has a 25 MHz PLL and appears "useable", the Vivanco is much easier to disassemble (just in case someone wants to replace the crystal with an OXCO). The "Schwaiger LNB 1" has the F-connector mounted on the cover, and when disassembling it, the first thing that happened was that I ripped the IF center pin out of the board. Duh. No problem with the Vivanco LNB (STL US3K-N), because the F-connector is on the opposite side of the cover.

I can post a few photos from the interior if anyone is interested (not sure about size limits in attachments on this forum)

73,

Wolfgang "Wolf" DL4YHF

- hope to meet you in Bochum again, or in Friedrichshafen this year -