A hearty thanks go out for the early feedback on the six-port document! I've got some good ideas for improvements and am looking for more. If you have not gotten a chance to read http://www.delmarnorth.com/microwave/newsletters/sixportmodel.pdf then please do. I need help with making this document work, and am genuinely interested in making it better.
There is indeed a calibration issue with six-ports. I read an academic paper that described what is the equivalent of a two-tone test for six-port calibration. This paper describes a procedure for calibrating six-port devices without having to have an IPO's worth of test equipment, and instead using (essentially) an over-the-air or continual calibration.
I know these things (six ports) may seem like RF black magic and all, but my aim is to plow through the basics of the devices and see what works for amateur radio. On the face of it, we get large bandwidth and good performance from passive devices and simple layouts, with direct conversion from microwave to I and Q. Where we "pay" for this performance is in terms of calibrating the device, and in the signal processing requirements of the FPGA that receives this (jumbled mess?) of I and Q. I believe we're up to the task, though.
This document's scope is the ideal case, or the model of the six-port structure. The real six-port device is going to have different parameters and settings and calibrations, because things like non-ideal quadrature hybrid couplers and non-ideal 3dB dividers are reality, and we build reality, not models. However, the model will help define what we're after, and I'm still working out what the ideal case is.
Any feedback you have on the document is appreciated, and please feel free to share it with the list.
More soon!-Michelle W5NYV