[Elecraft] voltmeter offset and accuracy in K2

[Jim Giammanco]

Please forgive me for putting on my "physics teacher hat" and doubly
forgive the algebra which is to come  :^))

I'd argue against trying to recalibrate the K2 internal voltmeter to
compensate for the diode drop in series with the supply. How come?
Because changing the resistor values in the voltage divider is a
multiplier correction, while the problem is an additive offset. 

The algebra:

if the supply voltage is Vs and the metered voltage is Vm then

Vm = Vs - Vf		where Vf is the forward drop of a diode

if you fiddle with the voltage divider you can get

Vm = k (Vs - Vf)		where k is the factor for the "new"
divider.

Now, for any particular Vs, and knowing that Vf is pretty much constant,
you can choose a k that will make the meter read perfectly, that is Vm =
Vs      if      k = Vs/(Vs-Vf)

But if Vs changes, you'll need a new k.  Granted, since Vs will probably
vary over only a small range, say 11 V to 15 V, the difference won't be
large. But ... the diference now won't be known, whereas before there was
a known, predictable offset. The k value will only need to change by 2%
or so, but 2% of 14V is a couple of tenths.

You went from *knowing* the reading was about a half volt too low, to
having a reading that could be a couple of tenths high or low depending
on where in the range you are.

What you really needed to do was add Vf to both sides:

Vm (new)  =  Vm + Vf  = (Vs - Vf) +Vf

And that you can't do by changing the divider ratio. Could it be done in
the firmware? Sure, just add in the requisite counts to put back the
diode offset. But I'll betcha Wayne and Eric figured it was better to
know the actual voltage the rig is seeing (after the protection diode AND
any voltage drop in the power cord).

OK, physics teacher hat off, QRPer hat back on....  If it ain't broke....

73
Jim N5IB