Mixed-mode port mapping¶

Wrong port mapping produces numbers that look fine and are physically wrong. Confirm mapping before any mixed-mode arithmetic.

What must be true¶

You need a consistent answer to three questions:

Port roles — which physical terminal is on each VNA port (P_in, N_in, P_out, N_out)?
File identity — which dataset/file is which pair (p13, p14, p23, p24, or polarity PP/PN/NP/NN)?
Label order — when a filename says PN, is that [output][input] or [input][output]?

PN is the usual mix-up:

Convention	`PN` means
`[output][input]`	N_input → P_output
`[input][output]`	P_input → N_output

Those are opposite paths.

Using filenames that encode both¶

Labs often name files like G6_Probe_p14_PN.s2p. Cross-check the two encodings:

p13_NN → port 1 → port 3 labeled NN → port 1 = N_in, port 3 = N_out
p14_PN → port 1 → port 4 labeled PN → port 1 = N_in, port 4 = P_out
p23_NP → port 2 → port 3 labeled NP → port 2 = P_in, port 3 = N_out
p24_PP → port 2 → port 4 labeled PP → port 2 = P_in, port 4 = P_out

If all four agree, present that mapping and ask for one confirmation. If they contradict, or nothing is encoded, ask only what is still unclear — do not guess.

Example map (confirmed)¶

Port 1 = N_in, 2 = P_in, 3 = N_out, 4 = P_out, labels [output][input]:

File	VNA path	Polarity	Variable in the formulas
p13	1 → 3	NN	S_NN
p14	1 → 4	PN	S_PN
p23	2 → 3	NP	S_NP
p24	2 → 4	PP	S_PP

This is not the Bockelman default (which puts P on ports 1 and 3). Either map is fine; mixing them in the formulas is not.

CloudSprite parameters¶

Record the pair on each dataset (mixed_mode=p13 or port_pair=PN) and a shared serial so the four siblings can be found later. Filename is not a catalog.

Do not proceed without confirmed port assignments.