Page 1 of 1
Multiply by 100?
Posted: Thu Apr 06, 2017 3:38 pm
by Jan Lichtenbelt
I try to calculate the percentage of A compared to B. In the simulation I get the right answer of 98%, but with the 16LF1847 completely wrong result.
Who is that possible. And what to do?
- Results
- Test_ULong_V1.jpg (61.32 KiB) Viewed 5517 times
With kind regards
Jan Lichtenbelt
Re: Multiply by 100?
Posted: Thu Apr 06, 2017 4:36 pm
by Benj
Hi Jan,
Is that the right program in the attachment?
Re: Multiply by 100?
Posted: Thu Apr 06, 2017 8:44 pm
by Jan Lichtenbelt
Hi Ben,
Sorry but I used the wrong file. Plaes find the Test_Ulong_V1 above.
Jan
Re: Multiply by 100?
Posted: Fri Apr 07, 2017 10:22 am
by Benj
Hi Jan,
The calculation D = 100 * B is what is causing the problem.
If you instead change to this then it should work as expected.
D = B
D = D * 100
This forces the * 100 to be a 32-bit calculation.
If you use the PrintFormattedNumber function to print out the D variable then you will see the true 32-bit value on the LCD - 590000.
Re: Multiply by 100?
Posted: Fri Apr 07, 2017 11:06 am
by Jan Lichtenbelt
Hi Ben
Thanks a lot. It works.
Multiplying by a number ( eg. 100) is no problem if one uses Uint. But is seems that in case of Ulong the number has to be changed first into a Ulong variable.
Kind regards
Jan Lichtenbelt
Re: Multiply by 100?
Posted: Fri Apr 07, 2017 11:31 am
by medelec35
Hi Jan,
One thing I have noticed.
If you want to display correct 590000 instead of 176 then instead of using LCD
you can use
.
Use 1 in the Bool input box for Unsigned number.
Martin
Re: Multiply by 100?
Posted: Fri Apr 07, 2017 12:12 pm
by LeighM
Multiplying by a number ( eg. 100) is no problem if one uses Uint. But is seems that in case of Ulong the number has to be changed first into a Ulong variable.
Yes, it's a "C" thing. It depends upon the size of an "int" on the platform/processor.
So it isn't a problem on say ARM, which is 32 bit.
For interesting reading search "integer promotion"
