Re: 3 axis allegro 3977 stepper motor board

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "turbulatordude" <davemucha@j...>
wrote:

> Nice pictures and a nice board.
>
> Can you elaborate on the noise to which you speak ?
>
> Of course another question is about the schematic, is that available ?

I based my board off the Allegro demo board for this chip. the schematic is here:

http://www.allegromicro.com/demo/a3977sed/schematic.pdf

However, after building a single axis board with the step and direction pins right next
to the chip and flying leads soldered to those pads, I noticed that the stepper motor
noise was coupling into the control lines and making the stepper "walk" under certain
conditions (more on this later). I've dealt with lab grade stepper motor controllers
that use PWM and it's been hard to keep the noise out of sensitive instruments
nearby. My solution on this board was to add a pair of inverter chips over near the
dB25 connector which will buffer the input pulse and direction signals, and also
provide a low impedance output to minimize noise pickup as those traces head into
that noisy chip. With a shielded parallel cable the motors no longer "walk" and I can't
induce the walking behavior even if I set a stepper on the parallel cable.

My board design is essentially the above schematic repeated three times with the
above described inverters and a 5 volt linear regulator for the logic side of the chip
which steals power from the motor supply. So this board only requires 24 volt input
and the 5 volt logic line is quite clean with the 70 dB of rejection of the garden variety
linear regulator.

The noise I speak of is a nasty side effect of PWM (pulse width modulation). This and
all PWM controllers are never quiet, they are always spitting out a stream of pulses
(square waves) to the stepper motor coils. The duty cycle (amount of "on" time vs.
"off time") is varied to get the integrated current desired. The effect is that there is a
never ending stream of 24 volt (or whatever your power supply voltage is) square
waves running out to the motors. the leads to the motors and the motors themselves
make great antenna's (really great depending on your lead length) radiating all that
energy out everywhere. You see the effects of this with an AM radio (there was a
thread about this on the xylotex forum), and with the problems that many people
have getting the xylotex driver to behave. It's a great chip, and PWM can be very
useful (it's great for microstepping), but it adds a level of complexity and problems to
controller and system design.

I'm going to be doing a half-stepping design as well which won't be PWM and will be
quiet (at least when the motors aren't running). The beauty of a half-stepping (or
whole stepping) design is that there isn't a need for feedback around the output
devices (it's pretty brain dead), so the outputs can be scaled up easily. You want 2
amp outputs? You want 20 amp outputs? you want 130 amp outputs? no problem,
the same driver can drive all of them. the allegro chip is limited to 2.5 amps output
and there's not a lot more you can do about it.