Re: Hacking the AWG WIRE TABLE (fwd)

To: tesla-at-pupman-dot-com
Subject: Re: Hacking the AWG WIRE TABLE (fwd)
From: Tesla List <mod1-at-pupman-dot-com>
Date: Tue, 5 May 1998 10:44:03 -0600 (MDT)
Approved: mod1-at-pupman-dot-com

---------- Forwarded message ----------
Date: Tue, 5 May 1998 08:47:54 -0700
From: Jim Lux <jimlux-at-earthlink-dot-net>
To: Tesla List <tesla-at-pupman-dot-com>
Subject: Re: Hacking the AWG WIRE TABLE (fwd)

----------
> From: Tesla List <tesla-at-pupman-dot-com>
> To: tesla-at-pupman-dot-com
> Subject: Re: Hacking the AWG WIRE TABLE (fwd)
> Date: Tuesday, May 05, 1998 7:26 AM
> 
> 
> 
> ---------- Forwarded message ----------
> Date: Tue, 05 May 1998 08:27:57 -0500
> From: David Huffman <huffman-at-fnal.gov>
> To: Tesla List <tesla-at-pupman-dot-com>
> Subject: Re: Hacking the AWG WIRE TABLE (fwd)
> 
> This is how I hacked the wire gauge table.
> Entering the table information on a spreadsheet (I used the CRC handbook
> 53rd addition) I made a plot then added a trendline of an exponential
type.
> The equation can be displayed for the fit. I got this equation which was
> close enough for me.
> dWireDia[in] = 0.32485*exp(-0.11595*iGauge)
> Dave Huffman
> 
> 

Wow, all these techniques for finding out the proportionality constant for
a geometric series. I'll throw mine into the ring.

the series is = ConstantA * ConstantB^N

The problem is to find ConstantB.  ConstantB = Nth root of (
HighValue/LowValue)

So, taking the values for #10 and #30 (for example), ConstantB = 20th root
of (.1019/.100).

Most handheld calculators do this in two keystrokes. Or, for an Excel/Basic
equation:

constantB = (highvalue/lowvalue)^(1/N)

Internally, of course, the calculator or compiler probably generates code
to use the identity:

A^B = EXP( B * LN(A))

although, there are other series to calculate powers and roots which are
computationally more efficient.  For instance, using logs to base 2 and
powers of 2 is substantially faster than log base e or 10.