Re: [CAD_CAM_EDM_DRO] Re: Big servo motors drive recommendations

Mark Vaughan wrote:

>
>
> Rutex do have what they call a fold back system, which does seem to start
> clipping the pwm short when it starts to get into problems. Makes a horrible
> grunchy sound.

My servo amps whistle quite audibly when in current limit. It
is a cycle-by-cycle limit, a flip flop gets reset at the
beginning of every PWM cycle, and gets set (shutting off the
transistors) when current exceeds the threshold.
It just doesn't do anything when real problems occur bar blow

> the fets, and when into foldback the noise this creates causes massive
> encoder / position errors.
>

Assuming the system isn't being asked to exceed its
capabilities, the current limiting doesn't cause position
errors. The "grunchy sound" and position errors sound like the
current limiting there is done at a slow speed, by a CPU. What
you don't want to do is make any discontinuities in the transfer
function of the servo amp that aren't necessary to save
hardware. If there is a large discontinuity, it upsets the
control loop.

> The Tek's do not have any circuitry to achieve this, they seem to have
> conduction commutation problems when they are starting to get into trouble
> just before they trip, they just trip way too quick.
>
> Shuting off within the pwm cycle but allowing it to continue is a good
> method as you say, it doesn't compensate for dynamic range problems, but it
> stops them becoming a problem, and provides a lot of protection just as it
> is required with good recovery. With yout FF1 and FF2 it should make quite
> an aggressive acceleration viable and also allows much wilder inductance
> figures, though your filters will mask this issue as you say.
>

My filters are pretty small in comparison, just 47 uH on each
motor wire.

Jon