Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
Posted by
David G. LeVine
on 2007-11-13 11:06:31 UTC
At 12:16 PM 11/12/2007, you wrote:
discussing) or as an offset voltage. You get the same resolution for
the same number of bits since sign is a bit.
If we let 0000 (binary) be zero volts, one count will be 0001, minus
one count will be 1111, just the same as integer arithmetic is done
in computers. With 4 bits, there are 16 states.
For sign and magnitude, we see sign => +, 000 => zero AND sign =>
minus, 000 => zero for zero volts out, note that TWO states are zero
valued. Full scale is sign => plus, 111 => max and sign => minus 111
=> negative full scale. For 4 bits, there are 16 states but only 15
*different* states.
At 8 bits it is a crapshoot, 1 in 256 is not that important.
Let's assume the PWM is set for 50% duty cycle = zero and 100% duty
cycle is plus full scale, 0% duty cycle becomes minus full
scale. The problem here is that you need a PID controller to change
the instaneous voltage to something the servo will do well with -- it
can be in software or hardware. If the PWM is at a fixed rate, a
fast filter is pretty easy, if not it needs a slow filter.
David G. LeVine
Nashua, NH 03060
>This will almost work, but it will be unidirectional. You'd needActually not. You can either do it as sign and magnitude (as you are
>additional circuitry (pretty simple, though) to flip the aritmetic
>sign of the analog output for the reverse direction.
discussing) or as an offset voltage. You get the same resolution for
the same number of bits since sign is a bit.
If we let 0000 (binary) be zero volts, one count will be 0001, minus
one count will be 1111, just the same as integer arithmetic is done
in computers. With 4 bits, there are 16 states.
For sign and magnitude, we see sign => +, 000 => zero AND sign =>
minus, 000 => zero for zero volts out, note that TWO states are zero
valued. Full scale is sign => plus, 111 => max and sign => minus 111
=> negative full scale. For 4 bits, there are 16 states but only 15
*different* states.
At 8 bits it is a crapshoot, 1 in 256 is not that important.
Let's assume the PWM is set for 50% duty cycle = zero and 100% duty
cycle is plus full scale, 0% duty cycle becomes minus full
scale. The problem here is that you need a PID controller to change
the instaneous voltage to something the servo will do well with -- it
can be in software or hardware. If the PWM is at a fixed rate, a
fast filter is pretty easy, if not it needs a slow filter.
David G. LeVine
Nashua, NH 03060
Discussion Thread
turbulatordude
2007-11-11 18:22:55 UTC
voltage based servo controller ?
Jon Elson
2007-11-11 20:24:30 UTC
Re: [CAD_CAM_EDM_DRO] voltage based servo controller ?
turbulatordude
2007-11-11 22:04:52 UTC
Re: voltage based servo controller ?
Stephen Wille Padnos
2007-11-11 22:13:59 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
Jon Elson
2007-11-12 09:10:56 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
Jon Elson
2007-11-12 09:12:47 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
Jon Elson
2007-11-12 09:15:50 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
Stephen Wille Padnos
2007-11-12 09:21:47 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?
David G. LeVine
2007-11-13 11:06:31 UTC
Re: [CAD_CAM_EDM_DRO] Re: voltage based servo controller ?