ActiveSplitter

ActiveSplitter Multicoupler

Hardware design for an active RF splitter for use with RTL-SDR receivers in the 100 MHz to 1 GHz band. Not intended to be a piece of precision test equipment, but something inexpensive and "good enough" to use for general radio monitoring and scanning activities in a smaller and more convenient form factor than cable TV splitters or conventional higher end multicouplers.

Uses amplifier circuit from this LNA design.

Design files are in KiCad format.

Estimated cost to build is $40 to $50 in low quantities with U.S. sourced parts. Could be as low as $10 BOM cost with overseas sourcing and higher quantities.

NOTE: I've built one and it seems to function about as intended, but I haven't measured performance using real test equipment yet.

• Designed to connect four RTL-SDR USB receivers to a single antenna without insertion loss.
• Amplifier circuit provides about 18 dB to 20 dB of gain. Four way resistive divider results in 12 dB of loss per output, so total is about 6 dB of gain to each port.
• Amplifier is designed to be powered from an SDR providing 3 to 5 volts DC power to the J1 RF output port via a bias tee or from regulated external power.
• Spacing of the output SMA connectors is designed to accomodate plugging the SDR sticks directly into this four-port powered USB hub.
• Main board and cover are PCBs. "Sides" of the box are formed by .1" header pins soldered to each board. The images show all pins populated, but only every fourth or so is probably need to provide sufficient RF shielding.
• Images below show using one SMA jack for the input and four SMA plugs for outputs connected directly to the RTL-SDR dongles. However, these can all be substituted for varying SMA genders and orientations (90 degree vs. straight) to accomodate different configurations and cables.
• PCB footprints are included for 50 ohm terminating resistors for any unused outputs.

Questions for reviewers:

• Nearly all RF splitter/combiner circuits that I reviewed seemed to be carefuly about maintaining symmetry in the length of all PCB traces in each "branch" of the split. How important is this, and why? Is there an "acceptable" amount of assymetery relative to wavelength? Giving up symmetry would likely allow the PCB to shrink in overall size a bit. For example, could use a single 4-port splitter instead of two 2-port splits. That should also increase port to port isolation across adjacent ports.
• I had originally planned to include a Pi attenuator between the output of the amplifier and the first split to further knock down excessive gain from the amplifier so that insertion loss would cancel out gain. As it stands, there will probably be about 6 to 8 dB of gain from input to each output. Is there any reason that could cause problems? Should I try to add the attenuator back in, even if it makes the overall size larger?
• Will the .1" pin headers create a reasonable Faraday cage instead of a solid housing?
• Is there any advantage or reason to combine R6 with R8 and R5 with R7?
• Should the splitter that starts with R6 be closer to R8 (same with R5/R7)?
• Is the circuit providing DC power to the amplifier (L1, L2, C2) reasonable?
• Should the values of C1 and C3 be changed to support the target range of 100 MHz to 1 GHz?

Assembled View:

Schematic PDF:

Main PCB View 1:

Main PCB View 2:

Cover PCB: