[sdiy] A little OT, Begging a hand

[Ian Fritz]

At 03:36 PM 2/8/2011, Tim Stinchcombe wrote:

>OK, it's been 20+ years since I dealt with anything like this, but I'll
>throw this out there: since we have 5 unknowns, but can easily set 3
>variables, V1, V2 & V3, to each of at least 2 values (+ve, gnd and perhaps
>o/c) and then measure Vout, can't we turn it into an *over*determined
>system, and get say 8 equations (=2^3), and then seek some sort of 'least
>squares' fit solution?? (I have all sorts of (m x n) x (m x n)^T matrix
>images in my mind, but I don't know that I ever fully understood it back
>then anyway...)

You are assuming that all of the 2^3 conditions lead to meaningful and 
independent equations. They don't.  For example, if  V3 =0 then you get the 
same equations whether you set V1 = Ve, V2 = 0 or  V1 = 0 and V2 = Ve. All 
the voltages and currents are merely reversed.

Ian