10bii+ Trig Bug

[Katie Wasserman]

I was just thinking about what the 10bii+ would be like as a re-purposed RPN calculator and playing around with it when I found a trig bug. Try this in degrees mode:

270

SIN

(you'll correctly get -1.00)

270

TAN

(you'll see OFLO flash and get correctly get -1.00E500)

270

COS

(you'll correctly get 0.00)

270

TAN

(you'll see OFLO flash and get 1.00E500 -- the bug)

This is in version 7-13-2010. I hope that this is the only bug, I use the 10bii+ a bunch and find it to be a really nice (if non-RPN) machine.

-Katie

p.s. There's still a yet-to-be-found Easter Egg in there.

Edited: 5 July 2011, 10:51 p.m.

[Tim Wessman]

I have not been able to reproduce the sign change thing. Is that what you are referring to? Or the fact that it does the overflow and then returns a number?

270 is undefined as it is an infinity sign change. Could you please explain more?

TW

Edited: 5 July 2011, 11:52 p.m.

[Katie Wasserman]

Correct, I am referring to the sign change.

In the first case the TAN of 270 returns -1.00E500 in the second it returns +1.00E500. You need to do these steps in the order I wrote starting with a ALL CLEAR. I can reproduce this on both 10bii+ calculators that I have.

I realize that TAN 270 is infinite but in one case it's negative infinity and the other it's positive infinity. Apparently the sign of the last result obtained before the TAN function is the determining factor.

More....

Actually, it turns out that an overflow condition always carries over the sign from the previous calculation. This happens even for something like factorial, for example:

1

-

2

=

300

!

results in -1.00e500

The 10bii (not +) results in +1.00e500

Edited: 6 July 2011, 1:54 a.m.

[Thomas Radtke]

A side note: It would be very esthetic to omit a (likely wrong) result at all on overflow.

Edited: 6 July 2011, 5:21 a.m.

[Katie Wasserman]

I agree and I believe that the 20b/30b takes this approach. However the 10bii+ was intended to replace the 10bii and needs to be upwardly compatible with it in all respects. One of which is just flashing OFLO and returning 1e500 when an overflow occurs.

[Thomas Radtke]

Yes, the 20b at least omits it.

Just noticed the german error message on overflow (300!): "Keine Amortisation". Strange.