ROOT bug? HP 48S/48G

[Eddie W. Shore]

'X^3+5*X^2-2*X+7'

'X'

0

ROOT

Returns .189254744132. (not a root, f(x) returns about 7)

But using the poly solver...

[1,5,-2,7] gives the correct answers: (approximately)

(.29141, -1.08117), (.29141, 1.08117), -5.58283

[Luiz C. Vieira (Brazil)]

Hi.

I did not check it, but wouldn't it be a pole?

[Les Koller]

If you read on the ROOT function of the advanced users manual, it shows you that you enter 1: function 2)variable to solve for 3) guess. The number returned is either the ROOT or the Local EXTREMA. There is a local minimum at x = .189....what your answer was. 0 (your guess) is closer to this value than the actual roots, that's why ROOT gave this answer for this specific guess.

[Eddie W. Shore]

Quote:

---

If you read on the ROOT function of the advanced users manual, it shows you that you enter 1: function 2)variable to solve for 3) guess. The number returned is either the ROOT or the Local EXTREMA. There is a local minimum at x = .189....what your answer was. 0 (your guess) is closer to this value than the actual roots, that's why ROOT gave this answer for this specific guess.

---

Good to know: thank you Les. I'll have to modify my program accordingly.

[Les Koller]

:)

[Gilles Carpentier]

Same result with the 50G, using ROOT

Note that on the 50G, in approx mode :
'X^3+5*X^2-2*X+7' SOLVEVX
gave the 3 roots (in complex mode, 1 in real mode)

{ 'X=(.291415279138,-1.08116667823)' 'X=(.291415279138,1.08116667823)' 'X=(-5.58283055828,0.)' }

but is unable to find the exact roots

I don't remember if SOLVEVX exists in 48 series

[link:http://www.wolframalpha.com/input/?i=%27X^3%2B5*X^2-2*X%2B7%3D0%27]Wolfram math[/link]

Edited: 13 July 2012, 4:22 a.m.

[Les Koller]

Yes, same result on 50G here too, which is what prompted me to go to the Advanced UM. :)

[Les Koller]

Looks like, from the manual, the 48GX does not have solvevx.

[Eddie W. Shore]

Quote:

---

Same result with the 50G, using ROOT

Note that on the 50G, in approx mode :
'X^3+5*X^2-2*X+7' SOLVEVX
gave the 3 roots (in complex mode, 1 in real mode)

{ 'X=(.291415279138,-1.08116667823)' 'X=(.291415279138,1.08116667823)' 'X=(-5.58283055828,0.)' }

but is unable to find the exact roots

I don't remember if SOLVEVX exists in 48 series

[link:http://www.wolframalpha.com/input/?i=%27X^3%2B5*X^2-2*X%2B7%3D0%27]Wolfram math[/link]

---

I think SOLVEVX started with the HP 49G, not in the 48 series.