[turbothis]

i got a nice powerfull spindle motor for my mill and would like to use the resolver it has on it now for ease of installation and also to learn about how they work.
i see the c code for the excitation of the resolver in the dynomomtion site. i assume one could not just run 5v to the resolver ref-/ref+ and then wire the sin/cosin to the dac inputs?
resolver pin out =

Pin A - S3 (black), Sin Hi
Pin B - S1 (red), Sin Lo
Pin C - S4 (blue), Cos Lo
Pin D - S2 (yellow), Cos Hi
Pin E - R1 (red/white), Ref Lo
Pin F - R2 (yellow/white), Ref Hi

thanks matt

[TomKerekes]

Hi Matt,

You have a tendency to take on difficult tasks

Resolvers are fairly complex, it depends what kind you have but normally you can not just run DC+5V to the excitation coil. Did you watch our video? A resolver acts more like an AC Transformer with two output windings. The AC output voltages vary based on the rotor position. Sampled at the right time can get a measurement of the amplitude. You might be able to use a relay driver output on Kanalog to excite the resolver. It helps to have a scope to see the outputs.

Regards

[turbothis]

ya, a lot of things in my life are this way.
i did watch the video but have zero experience with them so that was step 1.
i think i might just pull it out and install another encoder then if this is going to be a whole new can of worms.

i would like to get a scope to see all this stuff first hand. is there a cheap radio shack style scope for beginners?
thanks matt

[Al_The_Man]

QUOTE=turbothis;1397014]
i think i might just pull it out and install another encoder then if this is going to be a whole new can of worms.

i would like to get a scope to see all this stuff first hand. is there a cheap radio shack style scope for beginners?
thanks matt[/QUOTE]

That may be your best option, you need to know the excitation frequency/amplitude for the primary, there are USB double trace storage scope's such as Syscomp make fairly cheap that you can use with a PC to observe the output.
The natural result of resolvers are fairly low resolution for the outputs of the sin/cos, usually the resultant electronics use the arctangent function to calculate a high resolution result for position purposes.
The main advantage is they are usually a very rugged alternative to a optical encoder.
Not used that often now.
Al.

[RUSJIaN]

Help me. I have resolver end kflop+kanalog.KMotion with Resolver

[RUSJIaN]

Where is the switch?

[RUSJIaN]

Hi Tom, sorry for the delay.
For example, the x-axis (ADC(0)). I doing it right?

// Two Axis Resolver Program  1/6/08
//
// outputs square wave to both resolvers using Aux Switch Output
//
// then samples output coils near positive and negative peaks
// takes the difference to compute magnitudes
//

// these ratios are used to match the amplitudes of sine:cosine


#define RATIO0 (978.0f/768.0f)  // size j/size k
#define RATIO1 (978.0f/768.0f)  // size n/size m
#define RATIO2 (978.0f/768.0f)  // size p/size q

 

main()
{
    int i=0;
    int k0,j0,k1,j1;
    int m0,n0,m1,n1;
    int q0,p0,q1,p1;
    

    SetBit(167);  //+/-15V on
    SetBitDirection(152,1);      // configure AUX switch as output
    DefineCoordSystem(0,1,2,-1);    // Define 2 axis system
    

    Delay_sec(0.1);            // wait for +/- 15V to be stable

    for (;;)            // repeat forever
    {
        WaitNextTimeSlice();    // wait a few servo cycles
        WaitNextTimeSlice();
        Delay_sec(10e-6);     // wait for ADC conversion to complete
        k0=ADC(0);        // Sample all the ADCs
        j0=ADC(1);
        m0=ADC(2);
        n0=ADC(3);
        q0=ADC(4);
        p0=ADC(5);
        SetBit(152);        // Switch the resolver excitation
        

        //compute & track position based on measured magnitudes
        DoResolverInput2(ch0,(k1-k0)*RATIO0,j1-j0);  
        DoResolverInput2(ch1,(m1-m0)*RATIO1,n1-n0);
        DoResolverInput2(ch2,(q1-q0)*RATIO2,p1-p0); 
        

        WaitNextTimeSlice();     // wait a few servo cycles
        WaitNextTimeSlice();
        Delay_sec(10e-6);     // wait for ADC conversion to complete
        k1=ADC(0);        // Sample all the ADCs
        j1=ADC(1);
        m1=ADC(2);
        n1=ADC(3);
        q1=ADC(4);
        p1=ADC(5);
        ClearBit(152);        // Switch the resolver excitation
        

        //compute & track position based on measured magnitudes
        DoResolverInput2(ch0,(k1-k0)*RATIO0,j1-j0);
        DoResolverInput2(ch1,(m1-m0)*RATIO1,n1-n0);
        DoResolverInput2(ch2,(q1-q0)*RATIO2,p1-p0);

#if 0    // enable this to print the magnitudes occasionally
        if (++i==1000)
        {
            i=0;
            printf("%5.0f %5d\n",k1-k0,j1-j0);
        }
#endif
    }
}

[TomKerekes]

Hi RUSJIaN,

Resolvers are fairly difficult to setup and tune and often have nonlinear distortion. I'd recommend switching to digital encoders if possible.

Do you have any specs on the resolver? You will most likely need a scope to check the resolver signals. Do you have one and know how to use it?

See:
KMotion with Resolver

Kanalog has 8 relay driver outputs (ie SW0) that might work for the stimulus.

Regards

[RUSJIaN]

HI Tom,
I was able to connect resolver, but there are problems with the settings in the step response.
When P<1, there is no output to the motors. When P>1 there is hesitation.
Resolver connected to the shaft through a multiplier of 1:2.5

[RUSJIaN]

Hi Tom. Obtained the following results. Perhaps the resolver on the y-axis has a problem.

[TomKerekes]

Hi RUSJIaN,

You would need to answer my previous questions and provide a lot more information in order for us to help you.

Regards

- - - Updated - - -

Hi RUSJIaN,

You would need to answer my previous questions and provide a lot more information in order for us to help you.

Regards

[RUSJIaN]

It seems to be turned

[RUSJIaN]

Hi RUSJIaN,

You would need to answer my previous questions and provide a lot more information in order for us to help you.

Regards

- - - Updated - - -

Hi RUSJIaN,

You would need to answer my previous questions and provide a lot more information in order for us to help you.

Regards

Specs on the resolver:
1. Supply voltage:12 V
2. AC input frequency: 400 Hz
3.Output voltage: 6 V
4. Number of poles: 2

I don't have a scope, but if it will help I can get.

I was able to get feedback from two axes, the signal fluctuations of +/- 1 pulse. But for the third axis you need to edit the C code, because the fluctuations for the third axis +/- 5 and 15 pulses. How should look the C code for the three axes?

[bhurts]

Lower your velocity to 300000 and increase your jerk to 2-5000000. Then try another plot. Right now your position is lagging behind the command which means your motors can't spin as fast as your trying to command them to. You have a large following error. The 2 plot lines red and blue should be almost exactly the same.

Ben

[RUSJIaN]

bhurt, Ok.

Configuration for three axes(С code)?
#define RATIO0 (978.0f/768.0f) // size j/size k
#define RATIO1 (950.0f/709.0f) // size n/size m
#define RATIO2 ????????????????????????

[RUSJIaN]

Any ideas?

[TomKerekes]

Hi RUSJIaN,

I don't understand what you mena by "fluctuations +/-5 15"?

Please post the code you are using.

The Ratios are used to match the resolver sine and cos magnitudes so that they form a circle and not an ellipse. Such as:

Attachment 298808

You should print the ADC readings as the axis is moving to plot and determine the ratio to match the magnitudes.

It doesn't look like you did this for the other axes properly because of the oscillations per cycle in the position when moving. The position has distortion if the sine and cosine values form an ellipse rather than a circle.

Regards

[RUSJIaN]

Hi RUSJIaN,

You should print the ADC readings as the axis is moving to plot and determine the ratio to match the magnitudes.

Regards

How is that easier to do?

[TomKerekes]

Hi RUSJIaN,

You didn't post your c program as we requested.

Add a printf statement like in the example.

Please put more effort into your questions.

Regards

[TomKerekes]

Hi RUSJIaN,

That looks reasonable. But you did not explain what you are doing or showing. The time scale seems wrong (square wave high and low times should be 2x180us = 360us)

You can turn on printing the sine and cosine ADC readings by changing the

#if 0

to

#if 1

There is a bug in the print statement (it should print two integers). For axis 0 change to:
printf("%5d %5d\n",k1-k0,j1-j0);

For axis 1 change to
printf("%5d %5d\n",m1-m0,n1-n0);

For axis 2 change to
printf("%5d %5d\n",q1-q0,p1-p0);



Move the axis and determine the peak-peak amplitudes for each ADC to determine the proper Ratios. The ADC ranges can be different because of when the ADC happens to sample the signals. The 8 ADCs Sample in sequence every 90us.

Regards