not quite understanding the MATH function

[Kamikaze47]

yes, variable>>1 is the same as variable/2. >> is a bnary shift. so if a=8 for example, thats 00001000 in binary. when u shift it all 1 place to the right (>>1), you get 00000100 which is 4 (i.e. you have halved the value).

I notice you have some (1/2)'s in your modified program, but im afraid that wont work either. PBP will calculate that 1/2 (or 1>>1) as 0. 1 is represented in binary as 00000001 and if you shift that right 1 place you get 00000000.

Remember that time[lmt] is an allways going to be an integer (whole number). Therefore there is no point in checking if it is equal to 1.5 since it will never be 1.5.

[Rhatidbwoy]

Making sense. Been too long since i last did all of this. ok 1 is set in place for 0001B, 01Q, 1, and 01H then how does the half come. I can follow the counting that takes place when 0-F in Hex or 0-9 in decimal, butthe halves. I think it is safe to ask for reference to fractional literature. I would like to further my knowledge in such.

It is not quite as simple when it come to the basic math as we know of it as (what we learned in grade school) and for the following of fractions using controllers, well... Thanks I would have to say. Just show how far out off the loop I fall.


"yes, variable>>1 is the same as variable/2. >> is a bnary shift. so if a=8 for example, thats 00001000 in binary. when u shift it all 1 place to the right (>>1), you get 00000100 which is 4 (i.e. you have halved the value)."

so if I put >>2 would that be /4 (division by four). The shift right by 2 places. Trying to make sense.



The second question is the label PROGRAM2. I have an equation that has a negative value in the denominator. In paper math it should cancel out to a positive number. You had mention something earlier about this, would I have to change the equation so that there is no negative in the denominator or is there some way to work with the equation as is?



The third question is that I have the 1.5 would this fall to you shift comment. That it is more appropriate to utilize (1 + (>>1)) rather than to what I have.

[Kamikaze47]

yea, >>2 is the same as dividing by 4, >>3 is the same as dividing by 8 etc... you can also multiply by shifting left (e.g. <<1 is the same as multiplying by 2)

the line "time[chl] = (chnl1[chl] - 510) / -255" can of course be re-written as "time[chl] = (510 - chnl1[chl]) / 255". If you dont mind dividing by 256 instead it could also be written as "time[chl] = (510 - chnl1[chl])>>8" and would be much quicker to execute. However you must keep in mind that the result of this line is ALLWAYS going to be an integer even if the real matematical answer is 1.3 for example, the value stored in time[chl] will be 1

the same thing applies to the lines with 1.5 in them. the variables cant not possibly be equal to 1.5. You seem to be trying to check if the variable time[lmt] is bigger than, less than or equal to 1.5. This is not possible since time[lmt] is of course an integer and can only have whole values. You may need to re-think how your program works if you are trying to use a range of numbers between 1 and 2.

Perhaps you can post the equations/algorithms that you are trying to calculate and I may be able to work out what you are trying to do in your program.

Just for your reference, a variable that has been declared as a byte has 256 possible values in the set [0,1,2,...,255]. A variable that has been declared as a word has 65536 possible vlaues in the set [0,1,2,...,65535]

[Rhatidbwoy]

I got the last thing that you have said with the word and byte. Was reviewing that. The other thing is that I will have to change the value of 1-2 to something with higher values.


All to what I am trying to do is create a means to control a servo. The time constant that I was trying to work with first is that of 1-2 sec. That will have to change in means to 1-11 I guess. the only reason to that guess is the tenth place there are so many given posibility that will end up between 1-2 with .1 increment. So I guess then I would have to make the appropriate changes to the equation.

In other words I was using 0 to 255 as my input(y-axis) and 1-2 as my result(x-axis) coming to the equation of what I have there for my normal eq. and my reverse eq. I would just have to figure things out for my x-axis as being 1-11 rather than 1-2 . This whole thing does make sense. This then leave me to further refining the program to get that of the 1-2sec with IF..THEN and PULOUT. Not to bad.


The other thing to is that I did some reviewing and found out to what you are talking about with the shifts. I do have to THANK YOU. on the side note you taught me a few thing along with this forum. Yes there are people that are willing to help, but I believe the power of understanding is learning it. I can say that I did learn something and it is not just this problem but to what this forum has to offer.

[Kamikaze47]

Rhatidbwoy: No problem at all. I'm glad to help.

What you could do is apply a scaler to your result. For example, instead of seconds, work it out in milliseconds (scaler of 1000). Or if you dont need it that accurate, use a smaller scaler like 32 for example. Its good to choose a scaler that is a power of 2 (i.e. 2, 4, 8, 16, etc...) becuase you can multiply by these numbers very quickly and easily using the shifting you have learned.