Dear Claes,
cable influences are not an easy topic and how to minimise them are discussed in detail during our VNA Tools training course.
How to characterise and define a test port cable is described here: https://groups.google.com/g/vnatools/c/bMacNaEPJfY/m/gtMEKotcAgAJ
The
predefined cable specification stored in the database of VNA Tools are based on
the one-way manufacturer specifications for a 90 deg bend effect. This is
typically a rather conservative assumption and may be much smaller if the
cables are moved less and if they are not twisted during the mating process by
using a clamp system.
To keep one side fixed with a clamp and moving the cable only on the other port
side is always a good recommendation to minimize cable effects with two port
measurements.
Each cable movement
entry in Measurement Journal has the following information:
Time stamp, Cable definition, Port number, Position number, Movement number
Position:
is incremented from 0 by default (could be influenced by the user). Equal
positions get the same uncertainty ID in VNA Tools and are therefore treated as
correlated. The original idea behind this: if the cable is moved again to the
same physical position (same index number), it has exactly the same behavior
as before.
However, as we have learned, this is not the case. The problem is that any movement
of the cable to make a new connection can have effects on the cable that vary
from movement to movement, even if the cable returns to the same physical
position.
Therefore, each cable movement requires an increase in the Position number.
Each time the Position number is changed, the cable uncertainties specified in
the database are added to the measurement data.
Movement: is incremented from 1 (cannot be influenced by the user).
Hope this helps,