I went ahead and wrote-up my initial prototyping and included some measurements. Unfortunately, I didn't have the correct parts on hand, so measurements are a bit skewed. Nonetheless, this is a great starting point and I'll post updates on the next revision.
Link: Diplexer with integrated Bias-T (Triplexer) for QO-100 and X-band microwave reception using LNBs
Here are some pictures of my prototype
Why would it be misleading? Triplexer refers to a circuit that splits signals on a single port into 3 different ports depending on the signal's frequency. Or combines 3 signals of different frequencies into one port. That is precisely what this circuit does.
There's a Port for signals with frequencies higher than 120 MHz, a Port for a signal with a frequency below 30 MHz except DC, and a Port passing DC to 7,23 kHz (though only DC is applied). Therefore, this circuit is a Triplexer within the globally accepted meaning of the word. A quick Google search for the apllication of the word to this kind of circuit for QO100 use cases confirms the widespread use of the term. Not trying to refute your argument but I am haveing a very hard time to spot the misleading nature of the term triplexer.
That said, Super Bias-T sounds a lot cooler, which I would have thought of that before I sent the PCB of for manufacturing 
it is a Bias-T with an integrated Diplexer and 50 to 75 Ohm transformation. So since it has 3 frequency selective ports, I thought Triplexer was an appropriate name. Plus I thought it was commonly used for this kind of setup.
Hi everyone! I tried to research what kind of triplexer designs are currently used by others. However, I was unable to find much useful information. I am working on my own design and naturally try to avoid possible mistakes that others may already have made before me.
Here's my first prototype PCB. It uses a Mini Circuits 50 to 75 Ohm transformer for proper impedance match. Simple T networks with corner frequencies of around 30 MHz and 120 MHz provide Ref-Clock and IF isolation. After transforming to 75 Ohms system impedance, the DC voltage is injected trhough a RFC.
Here's a better design I implemented in DipTrace that I might send off to a PCB manufacturer for initial tests. The IF traces are kept a lot shorter in comparison to the REF IN traces. Can't wait to check its performance on a VNA.
And this is the schematic of revision 0.1:
Den Artikel habe ich jetzt auch noch auf Deutsch übersetzt: https://baltic-labor.de/2023/0…-band-und-qo-100-empfang/
Ich arbeite gerade an weiteren Artikeln, die den Umbau eines Goobay Universal Twin LNB erklären.
I continued on my quest to find affordable LNBs that allow easy modification for an external clock reference. The Goobay 67270 Universal Twin LNB also seems to be a good candidate. Under the same product number there appear to be 2 variants: One has a SMD crystal, the other has a through-hole crystal. Both can easily be replaced and a 2.2 uH / 18 pF series LC circuit connected to the first crystal pad will complete the mod. The needed drive level was about 1 Vpp.
I finally found the time to play around with the 5P49V6965 devices. They function as they're supposed to and happily translate frequencies. They accept a pretty broad range of input voltage levels way beyond the manufacturer's specifications. Therefore, I'll go ahead and design a little board that'll translate a 10 or 25 MHz reference signal to 4 other frequencies. It is possible to include an on-board TCXO and allow a switch-over to an external reference as the device has two different input clock pins.
If restricted to 4 fixed output frequencies, what would your opinions be on what the 4 most useful output frequencies would be?
On my initial design I would just use the 5P49V6965 as a stand-alone device, so the output clock properties will have to be burned statically into the device. If we can determine the 4 most useful frequencies, I can implement them so that other members of this forum can test the prototype boards as well. Later on I can include an MCU, or USB programming port, to make the board more versatile. Looking forward to your input on this one.
Here's a quick write-up on my blog about the (few) experiments I have conducted so far: https://baltic-lab.com/2023/07…as-versaclock-6e-devices/
The ICS512 and similar devices have never been great performers in regards of phase noise. To my knowledge, these ICS-chips are also no longer manufactured. Renesas offers much better clock generators / jitter attenuators, like the 5P49V6975 or RC22504A. I actually just ordered a bunch of both devices a few weeks ago. Since they have multiple, independent outputs, I'd like to experiment with a versatile 10 MHz to some other frequencies translator prototype. It might be possible to use these devices to generate all the different frequencies one might need for a QO-100 set-up with just one device.
While there already seems to be a thread covering the modification of the Goobay 67269 LNB and a great article by DH1DA, it seems that the manufacturer has changed the PCB layout completely. Since I ordered a larger amount of these, I ended up modifying them from scratch. I documented the modification as well as I could. I will follow up with measurements regarding frequency response and gain figures. The PDF can be downloaded from the link provided below.
PDF: https://doi.org/10.5281/zenodo.8102234
Blog (shortened version): https://baltic-lab.com/2023/07…-100-satellite-reception/
