Math with decimals problems.

neoteric

I understand the BS2 does not do decimals.· I also understand there is a limit to the size number I can have.
I have a variable·A that could be any number from 1 to 5000.· I then need to take that variable, multiply it by 4,680 and divide it by 3958.

A=A*4680
A=A/3958
This would be the same as taking the variable and multiplying it by 1.1824, which is fine.

A=Z*1.1824

I have another variable, B, that could be any number from 1 to 5000.· I then need to take that variable, multiply it by 3010 and divide it by 3729.
B=B*3010
B=B/3729···· (this example is slightly different, because I am multiplying a number that is smaller than the number I am dividing by.
This would bae the same as taking the variable B and multiplying it by .0871 which is fine.

I have searched the forums, and have seen references to "this is done with subroutines..."· but I cannot find any description of how it is actually done.··· I have googled, and have tried some of thier "fixed point" logic, but have been wholely unsuccessful.·· I would really appreciate someone at the very least pointed me in the right direction....

Thanks.

T&E Engineer

You may want to look at this link for Math calculations with a BS2:

http://www.emesystems.com/BS2math6.htm

Thanks.

Newzed

Neteoteric, here is the answer to your question.

To mulitply a number by a decimal, use the */ operator.· See Help section).

num = num*/$XXXX

Your decimal 1.824 has to be converted to $XXXX.· The interportion becomes $01.· The decimal portion becomes .1824 * 255 = 46.512
Round the number off to 47 and convert it to·· HEX number with your scientific calculator in Windows.· 47 = 2F, so your */ operator is $012F.
Write num = num*/$012F.

If your number was 500, then 500*/$012F = 591

500 * 4680 / 3958 = ·591.2076.

Close enough?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

Newzed

I meant to say integral instead of interportion.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

Ryan Clarke

Neoteric,

First and foremost I second the suggestion you read Dr. Allen's page that LTE suggested above- however I would suggest you look at this page of his:
http://www.emesystems.com/BS2math1.htm

Now, that being said-

Consider the following:

Result = ( N */ 47 ) + N

or

Result = ( N */ 303 )

ALSO:

Result = ( N ** 11954 )

How did we get these numbers?

Remember that the */ operator is "units" of "1/256" and the ** operator is "units" of "1/65536"- so multiply your 0.1824 * 256 on a calculator. Also try it with 65536. (Why do we add the extra N in the */ example? Figure that out and you will fully understand the ** and */ operators.)

Ryan

(Sid post jumped me [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Ryan Clarke
Parallax Tech Support

RClarke@Parallax.com

neoteric

Well, I think I understand it now.· Thanks to all who guided me.

((Why do we add the extra N in the */ example? Figure that out and you will fully understand the ** and */ operators.)

Because the */ only returns the lower byte.· The N thus has to be added to get the 1.x result.....

Thanks all.

Jon Williams

You've got it mixed: ** -- in practic -- multiplies by units of 1/65536, so it will always be something less than one. The */ operator works in units of 1/256, hence the upper byte is whole units (up to 255 [noparse][[/noparse]$FF]).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Jon Williams
Applications Engineer, Parallax