Mitsubishi Truncate Instruction

baboon · Jan 6, 2021

Hi,

Can anyone tell me if there is an equivalent 'truncate' function in Mitsubishi GX Works 2.

Ive worked with Rockwell and Siemens for years and this instruction, which both have a 'truncate' function but I cant seem to find one in GXWorks

Pleeeeease help

drbitboy · Jan 6, 2021

DIV(dividend - MOD(dividend,divisor),divisor)

If there is no division and you want integer part of A_FLOAT_VALUE, then you want

A_FLOAT_VALUE - MOD(A_FLOAT)VALUE)

?

baboon · Jan 7, 2021

Hi,

Ive tried the MOD instruction, but get an error when using it with a floating point number.

Its fine on a word etc

drbitboy · Jan 7, 2021

It's a little convoluted, but something like this should work.

Assume F32 is 32-bit IEEE-754 floating point, and we want TRN(F32).

COPy F32 bits to 32-bit integer I32
Bit-wise AND I32 with 0x7f800000, divide by 0x800000, subtract 127, call result E (32-bit integer)
If E is less than 0, then TRN(F32) is 0.0
If E is greater than 24, then TRN(F32) is F32 (no fractional part)
calculate integer (2^(23-E))-1, call result MASK (32-bit integer)
Bit-wise AND I32 with MASK, call result I32TRN (32-bit integer)
COPy I32TRN bits to 32-bit IEEE-754 floating point F32TRN

Those numbers near 24 may be off by +/-1.

Cf. here

AlfredoQuintero · Jan 7, 2021

baboon said:
Hi,

Ive tried the MOD instruction, but get an error when using it with a floating point number.

Its fine on a word etc

This is correct. The MOD and MOD_E function only work on integers or double integers. I can find no equivalent function to TRUNC in the manual.

https://dl.mitsubishielectric.com/dl/fa/document/manual/plc/sh080366e/sh080366ek.pdf

Is the information provided by drbitboy useful? What do you need to achieve?

baboon · Jan 7, 2021

Either:

Round down a floating point number.

(REAL_TO_INT or INT dont work as these round up)

OR

Get the decimal part of the floating point.

Basically i want to know if the floating point has ANY decimal part to it.

I.e.
31.123 = False
31.000 = True

baboon · Jan 7, 2021

Hi, There in NO TRN function

baboon · Jan 7, 2021

"Is the information provided by drbitboy useful? "

Unfortunately there isnt a TRN instruction, which is the issue

parky · Jan 7, 2021

Why do you want to truncate a float? for example how many decimal places and why ? The idea of floating point is the accuracy to x number of decimal places

daba · Jan 7, 2021

parky said:
Why do you want to truncate a float? for example how many decimal places and why ? The idea of floating point is the accuracy to x number of decimal places

He explained why in post #6

Truncate a float and see if result is equal to the float gives him what he is after ...

AlfredoQuintero · Jan 7, 2021

baboon said:
Hi,

Basically i want to know if the floating point has ANY decimal part to it.

I.e.
31.123 = False
31.000 = True

Is there any decimal order after which your application does not care?

Based on your example:
31.0001 = is this True or False?

parky · Jan 7, 2021

If you want say 3 decimal places then multiply the float by 1000 then convert it to an integer (or double integer if you think it is going to be above 32767). Convert it back to a real then divide by 1000.
So 3.141593 as a float multiply by 1000
= 3141.593
Convert it to an integer
= 3142
convert it back to a real
= 3142.0
Divide by 1000
= 3.142 (Note: this has increased the 3rd place from 1 to 2 as 593 (the decimal part will truncate to 2.
You could just check the decimal part and sub 0.1 from the original float to stop the rounding up of the float when converting it to an integer

parky · Jan 7, 2021

Did not see that one I was checking my logic at the time lol

baboon · Jan 7, 2021

I did try this, but if the value was 3.0008 for example, I would end up with 3 (Not 3.1), which cause an issue.

daba · Jan 7, 2021

baboon said:
Hi, There in NO TRN function

I do not know how big your float can be, but you might be able to perform a manual MOD function on it and test the final result ....

Keep subtracting 1.0 until the result is less than 1, if result = 0, then the float had no decimal part.

Obviously if the floating point value can be large, this will be detrimental to your scan time ...

You could speed things up (if you know the largest value your float can be) by doing this in "decimal chunks"...

e.g. Largest float = 5000.000

:loop
If {float} > 1000, sub 1000 : jump :loop
If {float} > 100, sub 100 : jump :loop
If {float} > 10, sub 10 : jump :loop
If {float} > 1, sub 1 : jump :loop

If {float} = 0 ......

Sorry I don't know the syntax for your PLC, but the idea should work ...

Mitsubishi Truncate Instruction

baboon

Member

drbitboy

Lifetime Supporting Member

baboon

Member

drbitboy

Lifetime Supporting Member

AlfredoQuintero

Lifetime Supporting Member

baboon

Member

baboon

Member

baboon

Member

parky

Member

daba

Lifetime Supporting Member

AlfredoQuintero

Lifetime Supporting Member

parky

Member

parky

Member

baboon

Member

daba

Lifetime Supporting Member

Similar Topics