Better:
9 SIN COS TAN ATAN ACOS ASIN 9 - EEX 6 *
The answer is in ppm error (as somebody once said)
Chasing guard digits doesn't further refine the value above, which is available immediately via the SHOW feature.
I'd already reported this to Mike Sebastian.
This means that they are using another algorithm, not only a
Saturn CPU emulation on a newer CPU.
There is any way to measure its speed?
Try this on "Solve" (Algebrical TVM formula, please modify it where necessary):
(1+i)^n*pv+((1-(1+i)^n)/(1-(1+i)))*pmt+fv=0This works on my HP-95LX with 16 digits precision:
n=31,536,000calculate for fv.
i=1/315,360,000
pv=0
pmt=-0.01
Some results:
HP-95LX : 331,667.0056360354Internal TVM (take same i mult. by 100):
HP-17BII: 331,661.800473
HP-95LX : 331,667.0066907769(Example taken from MATHEMATICS WRITTEN IN SAND by W. Kahan)
HP-17BII: 331,667.006691
HP-10BII: 331,667,006691
HP-12C : 331,667,0067
Best regards,
Nelson
The 49g+ equation solver gets the same result as the 17BII solver, 331,661.800473
The 49g+ financial solver is more accurate, returning 331,667.0067.
32sii Equation editor:
Entered (1+i)^n*pv+((1-(1+i)^n)/(1-(1+i)))*pmt+fv as an expression, (checksum 41D6, 57.0 bytes)
and solved for fv: got 331,661.800473 (Same as 17bii and 95LX solvers)
I think that the HP-95LX/200LX (and other HP palmtops) solvers in the "HP Calc" program is the most accurate solver that I knof of! It uses 16 digits (I don't know how much it uses internally) and the answer is even better than the HP-49G+...
BTW The 95LX "forensics" result is
9.000000000029575
Best regards,
Nelson
Edited: 20 Feb 2004, 3:31 p.m.