Nachrichten

1

oscat.lib fuer Step 7 / Re: Arithmetik doppelter Genauigkeit

« am: 16. Februar 2016, 11:25:54 »

I'm sorry for resurrecting an old topic, but I figured it's worth posting since I couldn't find any similar solution on the forum. I hope someone finds this useful while S7-1500 with native double float is gaining traction.


If I understood e3's post correctly, he saw similar problem in R2_ADD2 function as we did. My colleague "mme" came up with this solution based on R2_ADD:

Code: [Auswählen]

tR2_ADD2.RX = Y.R1 + Y.RX + X.RX + X.R1;
tR2_ADD2.R1 = X.RX + Y.RX - tR2_ADD2.RX + Y.R1 + X.R1
I did a quick and dirty test in C# using floats in place of S7 Real and with double as a reference. The code above appears to be working as expected.
Test code:

Code: [Auswählen]

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

namespace Real2Test
{
    class Program
    {
        static void Main(string[] args)
        {
            float a = 10000000F;
            float b = 0.00123F;

            double c = a;
            Real2 d = new Real2();
            d.RX = a;
            d.R1 = 0F;

            Real2 e = new Real2();
            e.RX = b;
            e.R1 = 0F;


            for (int i = 0; i < 3333333; i++)
            {
                c -= (double)b;
                d = R2AddNew(d, R2Mul(e, -1F));
            }
            Console.WriteLine("Double: {0:F8}\tR2: {1:F8}\t{2:F8}", c, d.RX, d.R1);

            for (int i = 0; i < 3333333; i++)
            {
                c += (double)b;
                d = R2AddNew(d, R2Mul(e, 1F));
            }
            Console.WriteLine("Double: {0:F8}\tR2: {1:F8}\t{2:F8}", c, d.RX, d.R1);

            for (int i = 0; i < 3333333; i++)
            {
                c += (double)b;
                d = R2AddNew(d, R2Mul(e, 1F));
            }
            Console.WriteLine("Double: {0:F8}\tR2: {1:F8}\t{2:F8}", c, d.RX, d.R1);
            Console.ReadKey();
        }

        static Real2 R2Mul(Real2 x, float y)
        {
            Real2 tmp = new Real2();
            tmp.RX = x.RX * y;
            tmp.R1 = x.R1 * y;

            return tmp;
        }

        static Real2 R2AddNew(Real2 x, Real2 y)
        {
            Real2 z = new Real2();
            z.RX = y.R1 + y.RX + x.RX + x.R1;
            z.R1 = x.RX + y.RX - z.RX + y.R1 + x.R1;
            return z;
        }
       
    }

    class Real2
    {
        public float R1 { get; set; }
        public float RX { get; set; }
    }
}

Thank you for this great library!
W