Re: freqmot.o settup problems

[Ray Henry <rehenry-at-up.net>]

Hi Jason

Well this post proves that I'm perfectly willing to answer on the basis 
of old knowledge and some ignorance of the deep inner workings of current 
code.  I guess I will have to compile a copy of it and try some of the 
new stepper stuff this weekend. (comments mixed into a snipped ini)

On Friday 27 September 2002 07:04 am, Jason wrote:
> hi all,
> 	finaly got my stepper driver back on line today. recompiled the latest
> cvs and all goes well. However I calculated the steps per inch at 5080
> ( 200step 1mm pitch ) , when i put it into the ini and start emc in mdi
> mode it dosnt drive correctly. I set it to g21 y1 and it only pulses
> aprox. 5 times ( nowhere near the required 5080 times ) .
> Any help is welcomed
>
> Thanks
> Jason Cox
>
>
> stepper.ini
>
> ; Stepping type, 0 = step/direction, 1 = phasing
> STEPPING_TYPE = 	4

I know that the code uses the following switch

 switch(STEPPING_TYPE){
    case 0: pstype = "Step/Dir"; break;
    case 1: pstype = "Phase"; break;
    case 2: pstype = "Quadrature/other"; break;
    case 3: pstype = "Four-Phase"; break;
    case 4: pstype = "Binning Step/Dir"; break;
    default: pstype = "Invalid";
  }

but I don't know that they have all been tested out with real machines 
doing real motion.  If four doesn't work as promised you might try 1 as 
it is the old freqmod.  If you do you will need to change a bunch of 
values. 

> LINEAR_UNITS =		0.03937007874016

You could set this to one and us mm all the way through your setup.  
Unless the number of steps per mm are very small this will work as 
promised.  Since you have 200 steps per mm you should be fine with mm but 
you will need to change several of the other variables to match mm values 
rather than inch.

> P = 1.000

This is much to low for freqmod but shouldn't make a difference at least 
with the old binning routine -- if it works as it did with steppermod.  

> FF0 = 1.000
> FF1 = 1.000
> FF2 = 1.000

These will give you really unexpected results with the old freqmod.  The 
value in FF0 will offset final position which I don't think you want.  
The value in FF1 should be okay but not used with binning.  Never tried 
FF2.

I'd set everything but (P) to zero for a starting setup and then work 
from there.  I'd set P around 500 to start if you decide to use type 1.

> INPUT_SCALE =			5080	0
> OUTPUT_SCALE = 5080.000 0.000

If these are not the same, they really trip up freqmod.  I am not certain 
what affect they have on the new binning.  

> MIN_OUTPUT =			-1000000
> MAX_OUTPUT =			1000000

These should have no effect on any of the stepper modules.  They are way 
big for any of the servo stuff as they represent the analog(ue) range in 
volts. 

> COMP_FILE =			x.cmp

I'd comment this out until you get the rest of it working to your 
satisfaction.  Then bring it back in and see what breaks.

> ; address for parallel port used for auxiliary IO
> IO_BASE_ADDRESS =	0x378

I see that you have set the same address for stepper and io.  I do this 
in order to use the parport's second output byte for io but you need to 
know what is being stepped on by what in order to get away with it.

Hope this helps

Ray