Record of known problems in SACAMOS

21/10/2019: A problem has been reported when using overshields. If the overshield is 
            not the last cable to be defined then the numbering of reference conductors
            (shields) in the bundle can be incorrect. Fixed 22/10/2019 CJS.
	    
23/10/2023: An error occurs if we try to include a transfer impedance which is purely resistive.
            The resistive part of the transfer impedance is lumped at the ends if the shield conductor
	    in the SACAMOS model and is always included so a resistive transfer impedance does not need
	    to be explicitly included as it is always there, however the error should be trapped and a 
	    warning shown.   

31/10/2023: The internal impedance of rectangular conductors is assumed to be Zint= Rdc+B*sqrt(jw) over the whole
            frequency range i.e. the internal inductance varies as 1/sqrt(w) at low frequency. It may be best to
            use Zint= Rdc below a certain transition frequency see Paul, section 3.6.2.2 -> 3.6.2.4. This will 
            probably have a minor effect as the imaginary part of the internal impedance will be dominated by the resistive
            part at low frequency. 

31/10/2023: In the twisted pair model, the value of cable%conductor_impedance(2)%Resistance_multiplication_factor
            should be 0.5 not 1.5 for conductor 2. This is then consistent with the shielded twisted pair model. 
            i.e. for the differential mode, conductor 2 is the reference and conductor 1 is the conductor whose
            voltage/current we calculate. The total resistance in the MTL equations is the reference+conductor 
            resistance=1.5 + 0.5 R =2R as required for twoo conductors in series. 
            For the common mode the conductor is assumed to be conductor 2 (the reference of the differential mode domain)
            whose resistance contribution is 0.5R i.e the contribution of two conductors in parallel. 
            
15/11/2023  When the twisted pair model is used in a bundle in isolation i.e. no other conductors, then the 
            domain decomposition matrices appear to cause a problem due to the reference conductor being conductor 2
            and there being no common mode/ differential mode decomposition required. The effect of this is that the 
            characteristic impedance is effectively halved. A work around is to use two single wires to create this cable
            within the model.