How Can They Make An SMA so Bad?
By
KØCQ
A few years ago I was hooking up a 10G transverter to a small amplifier and it made a neat connection to use several right angle male/female SMA for that connection. The insertion loss was more than I could measure.
I just grabbed a couple of those and ran some tests. My TDR didn't have fine enough resolution to detect anything. Using an HP423A detector with an HP415E as a relative power meter probably in the square law range, I found this pair inserted between the detector and the N to SMA adapters showed an insertion loss of 7 dB at 10.37 GHz. DC resistance was 0.09 ohm. Comparing the return loss to that of a pretty good termination inserting this pair showed the same return loss as the termination, with or without the termination. Loss City?
1
I attacked one of them with the milling machine taking the side off carefully.
What we have is a T, not a right angle in the center conductor. Best I can measure under the Stereozoom, the center conductor on the male side is 0.094” diameter and the outer conductor is 0.200” diameter. Giving a characteristic impedance of about 45 ohms. On the female side it looks like the center is 0.085” diameter and the outer is 0.175” giving an impedance of 43 ohms. The stub is about 0.093” past the center of the side branch, about 0.1 wavelength. The end of the stub is about .06” from the inside of the housing. If I compute that capacitance as half a pair of concentric spheres I get 0.13 pf. 0.1 wave of 45 ohm coax is +j33 ohms and that capacitance is -j118 giving a net reactance at the corner of about -j85 ohms. From here it looks like the insertion loss is from mismatch. The parts appear to be bare brass, but this style of SMA elbow made on the west side of the Pacific appears to be better used at ¼ GHz and down, not at 10 GHz.
73, Jerry, KØCQ
2