Here's how I'd probably do it:

1) Maybe I am too much of a mathematician, but I would drop all use of angles in degrees and perform all angle calculations in radians. Not angle conversions to become confused with. Along those same lines, I would think about my speedometer in terms of a unit circle (http://en.wikipedia.org/wiki/Unit_circle and/or http://www.purplemath.com/modules/unitcirc.htm if you have forgotten this concept). I find that so many trig problems (like this one) become a lot easier when viewed on the unit circle.

2) It appears that your intention is use linear interpolation to interpolate between the min and max values and the min and max angles. Putting things on the unit circle, this means you want max% at theta=0 and min% at theta=pi and linearly interpolate between. This can easily be done using the TREND() function. 
   --  min -- max
values -- 0.5 -- 1
angles -- =pi() -- 0

=TREND(angles,values,needle%)
http://office.microsoft.com/en-us/ex...in=HP010342656

With the angle theta thus calculated, plotting the end of the speedometer needle is as simple as x=cos(theta) and y=sin(theta).

Now I understand Jon Peltier means when he says that one of the reasons he doesn't like speedometer charts is that he spends so much effort explaining the trig behind them.