HP Forums

I realize this is primarily a financial calculator, but I've just noticed that the logs seem to fail above somewhere near 10^16. I didn't notice a page in the manual listing valid domains.

A few more details: it seems almost to be a hex<->dec conversion mistake. Log(10^16) is 10. And so on, 6 too small, until it then jumps to 12 too small, then 18 etc.

Oh well, I have always found using LOG on a financial calculator to be a little unnatural. :)

Quote:

---

A few more details: it seems almost to be a hex<->dec conversion mistake. Log(10^16) is 10. And so on, 6 too small, until it then jumps to 12 too small, then 18 etc.

---

The firmware in the recently released SDK, v7.21.2008, gives log(1E16)=16. log(5E16)=16.6989

hello,

Darn, you guys are quick at finding bugs, and discovering that they have already been fixed!

I just wonder how you discovered that the LOG of number with exponent >16 (and <-16 BTW) are not working properly? I doubt that you had a real need to calculate the log of such a large number?

regards, cyrille

I admit playing with my calculators as a pastime. I analyzed one resistive divider until I knew all the answers. I would enter the lower resistor into the t register and enter the upper resistor, add and divide and I got the total resistance and then the fraction of the divider. I wondered what would happen if I multiplied and subtracted and the number that resulted was the output resistance of the tap. Totally unexpected. When selecting constants from the list. I wondered what would happen if I keyed a number, it selected the constant numbered from left to right and entered it as well. Worked with the 33S and later the 35S. I find your divider solution on the 50g trivial, I like mine. Sam 80 I started with the first 35. I flat hate your invisible decimal, use a bigger symbol. I use the comma for greater visibility.

Quote:

---

Darn, you guys are quick at finding bugs, and discovering that they have already been fixed!

---