Filters, and a ? for Mariss was Re: Rotor inertia.

Jon,

The early drives didn't have encoder filtering and the ERR/RES was
sensitive. The drives have evolved as I have learned how they are
applied. I had erroneously assumed you were refering to drives built
within the last 12 to 18 months, so I apologise for that mistake and
my testy reply.

The cause of my bad attitude is learning HDTL and ISE6 for my new
Xilinx XC2S30 FPGA development system. Learning hurts and I'm in a
hurry to get this thing up and running; I have a lot of circuitry to
shove into it. Right now I've got a few gates blinking a few LEDs on
the development board, but those gates are in the FPGA. Not too bad
for being on the steep part of the learning curve.

Mariss

--- In CAD_CAM_EDM_DRO@yahoogroups.com, Jon Elson <elson@p...> wrote:
>
>
> Mariss Freimanis wrote:
>
> >Jon,
> >
> >Not true.
> >
> >The encoder input chain / channel is:
> >
> >1) 1K pullup resistor to +5VDC, input to schmitt NAND gate
(CD4093).
> >2) Output of NAND to 47K / 100pF low-pass (5uS RC delay).
> >
> OK, well a 5 uS filter is all you can have on a drive spec'd for
200 K
> steps/counts per second.
>
> >3) Output of low-pass filter into 2nd schmitt NAND gate.
> >4) Output of 2nd NAND to quadrature decoder.
> >
> >This filter will not pass any signal artifact narrower than 5uS
and
> >squares-up slow rising/falling encoder edges. The low-pass
bandwidth
> >matches the expected encoder roll-off frequency of 100kHz.
> >
> >This filter has been tested with a 20' foot motor cable bundled
with
> >unsheilded encoder cable in a common harness. The behavior and
> >accuracy of the servo was unaffected.
> >
> Well, that is interesting. Some users reported they needed shielded
> encoder cables to make their system work. But, I must admit, I was
not
> there, and didn't see shielded cable fix the problem.
>
> Jon