Hi all,
I’d like to split an external antenna, connected via an SMA connector on the PCB to multiple RX-only components (versions of the board include 2, 4, 8, 16 and 32 components). I know that each split will incur about 3db loss in the signal.
What component should I use (if any) to split the signal?
Does anyone have any reference design for such a board?
Any pointers are welcome.
Thanks!
The frequency? PCB based Wilkinson’s are only practical in the GHz range
Sounds like a question I was asked to investigate a year or two ago. I found several pages on the 'net that helped you design Wilkinson splitters. My impression was that mere mortals (like us) could do such designs if we need only moderate levels of performance and keep to the UHF (and low-GHz) frequency range. If impedance match, or balance between the split signals, is critical then you should consider hiring somebody with RF experience and access to a network analyzer.
It looked like one advantage of designing a custom splitter is that impedance matching - for example, a 50-ohm source to a 75-ohm antenna - could be integrated into the splitter design. And, one limitation is the vendor-to-vendor variations of so-called “FR-4” board material at those frequencies.
I had a few samples of connectorized, commercial, splitters in the shop so I popped the covers off and passed them under a magnifier. I saw a few places where it looked like traces had been shaved or trimmed with a draftsman’s knife, probably to improve the match or balance. My conclusion was that, if the project went forward (it didn’t), we should consider purchasing the commercial units, removing them from their enclosures, and using them as modules on our board.
Dale
The main question is what freq range? You need to maintain the impedance.
You can do either a resistive divider as noted by others, or you can buy a “power divider” from minicircuits.com They make nice stuff, a bit $$ and can be hard to buy in small quantities.
I’d like to use it for the 2.4Ghz band. Thanks for the answers, I found a cheap Wilkinson RF Splitter that I’ll try to use, it has 50-ohm input/output and very low insertion loss (about 0.3dB).
Theoretically, it seems like it’s possible to split the antenna to 32 components in a cascade of splitters with a loss of about 1.2dB… still looking for real-world, working examples for such a circuit.
You need to go look at splitters again. If you split 32 ways, each port gets, at the VERY most, 1/32 of the original signal level. I’m not sure what your spec sheet says, but -1.2db is not the signal reduction at each port. More like sig level = (1/32) -1.2db and that is optimistic.
A 1:2 splitter is down 3db on each output (likely more). 1:4 is -6db
Also, at 2GHz, you are going to have to manage trace impedance and line lengths. See if you can make a 1:4 work before you do a 1:32. Study some microwave design. Good luck!
The specified “loss” of a splitter is customarily given as the loss in addition to the mathematical reduction in power level due to the split. Therefore if a 2-way splitter has a loss specification of, say, 0.3 dB, then each output signal will be approximately 3.3 dB less than the input signal.
Seed a search engine with " wilkinson divider circuit PCB " to find examples similar to HERE and HERE and HERE and HERE TOO.
Dale
Dale
You can get away with cheap FR-4 PCB at 2.4GHz, the losses are not too high.
BUT, the Er value is a bit lower than the low frequency value and how much lower depends on your vendor.
Ok, so the insertion loss is cumulative with the loss due to the split, makes sense. So 32-way may not be practical without amplification.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.