Impedance termination & measurement w/ a power splitter

[berserkid]

I'm sourcing 500M & 6Ghz RF signal into a 0 degree 4way power splitter (uses sma/bnc cables)s port w/ the 4 output ports terminated separately to a 3dB attenutator/pad. There is a 50ohms pcb trace that individually connects the output side of the port and 3dB pad.

With the 4ports terminated to the 50ohm trace and the pad and then I measure the resitance (using the ohmmeter)from the pad or trace to gnd I'm getting ~39ohms and on the s port to gnd, it is ~39ohms as well. Is this the terminating impedance of my setup? If not how can I then measure the impedance? And how can I terminate my setup correctly as well?

 

[VEBill]

Many such power splitters will have a frequency range that does not include DC. A 50-ohm trace is a transmission line with a characteristic Z of 50 ohms (not a resistance of 50 ohms). 3dB 50-ohm pads will not measure 50 ohms at DC.

 

[pstuckey]

The best way to terminate your unused ports is with a 50 ohm load. A 3 dB pad is better than nothing, but is not good. If you don't have any loads, shoot for at least 10 dB worth of pads on the unused ports.

BNC is fine for 500 MHz, but won't work well at 6 GHz.

The usual way to measure impedance is with a network analyzer. These can be pretty expensive though, and if you are primarily interested in magnitude (i.e insertion loss, port to port isolation, S11 (mag)), these can all be measured with a signal generator, spectrum analyzer and directional coupler (for reflected power).

Peter