Re: TI-83 - complex numbers bug?
[Prev][Next][Index][Thread]
Re: TI-83 - complex numbers bug?
In article <342bc9dd.676242@news.btinternet.com>, Reinier Zwitserloot
<R.Zwitserloot@BTInternet.com> writes
>1. My TI-85, ROM 10, gives the radian answer even if in degree mode.
Not mine:
TI-85 (ROM 10) and TI-86 (ROM 1.3):
In mode PolarC
In Radian mode (17.321, 10.000) gives (20/_.524)
In Degree mode (17.321, 10.000) gives (20/_30)
In mode RectC
In Radian mode (20/_30) gives (3.085, -19.761)
In Degree mode (20/_30) gives (17.321, 10.000)
No problem - the way the angle argument is read and returned
depends (consistently) on the angle mode setting.
TI-83 (ROM 1.0200):
In mode re^Oi
In Radian mode 17.321 + 10.000i gives 20e^(.524i)
In Degree mode 17.321 + 10.000i gives 20e^(30i)
In mode a+bi
In Radian mode 20e^(i30) gives 3.085 - 19.761i
OK up to here. The way the angle argument is read and returned
depends on the mode setting. Now the punchline:
In Degree mode 20e^(i30) gives 3.085 - 19.761i
For my money, WRONG. It should give 17.321 + 10.000i.
For another take on it, in Degree mode, test the identity
A-abs(A)e^(angle(A)i) = 0
for any complex number A. It hits the wall at speed.
>2. This isn't a bug. It just so happends to be that e^complex relates
>to angles. You can't give it degrees though. you shouldn't be able to.
>This is a non-trigonometric instruction that happends to operate like
>a trig instruction. the ti-8x is correct in treating it as radian in
>any mode.
I'm not convinced. What we have here is a feature with six legs
and a couple of little antennae. That's not a problem. I can program
round it - but I'd like to know if TI are likely to fix it in future ROM
versions.
--
|>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Stuart Dawson Dawson Engineering
sd@dawson-eng.demon.co.uk
Belfast, Northern Ireland +44 1232 640669
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<|
Follow-Ups:
References: