Re: [CAD_CAM_EDM_DRO] Re: Lost steps

Ozzie@... wrote:

> Next I wrapped the motor wires from that axis in foil and grounded
> it,
> and wrapped the fan motor wires and grounded it. I reasoned that
> these
> might be generating electrical noise, and though I didn't know if the
> wrap would help, it couldn't hurt. (I'd already wrapped the signal
> wires, you might remember.)I tested all the wraps for ground.
>
> I put it all together and the motor ran much better but not without a
> few stutters, and the bare shaft does not get back to the same radial
> position after an increment that should have brought it to the same
> position, e.g. an inch.

Ah, HA! Now, you are getting somewhere. Clearly, it is some sort of
interference
problem, radiation from the motor wires and other things are
contaminating
the step and/or direction inputs to the driver.

> It may be my imagination but it seemed the longer I tested the worse
> the condition became, but maybe I just became more observant. In any
> case it never became as bad as yesterday.

If the system is on the very edge of noise sensitivity, then small
changes in
temperature may cause the input threshold to make it dramatically more
(or
less) sensitive.

> I also tested for continuity between computer case and driver case,
> and that was good.

Well, DC continuity is necessary, but you need good AC continuity, too.
And, you also need good SIGNAL grounds, not just chassis grounds!
How many ground wires are attached at the parallel port end of the
cable, and how many are attached to the stepper drivers. More wires
means lower impedance, hence better noise immunity.
You may need to connect a heavy (say, #10 stranded) wire between
the two equipment chassis, or even a length of ground braid.

> What next, any ideas.
> I suppose there are other wires I could wrap, but if that were needed
> wouldn't Dan at Camtronics have mentioned it. I've seen pictures of
> his drivers and the wires are not shielded??
> The insulators between the mosfets and the heatsinks are the ones
> supplied by Dan, as is every part in the case, and the case.

How heavy are the ground wires running from the driver to the
(presumably)
minus side of the power supply? These should be quite heavy, otherwise
large current pulses from the switching action will cause substantial
voltage
difference between the 3 axis drivers, and these currents will be forced
to flow
through signal grounds back to the PC motherboard, unless you have
opto-isolated inputs on the drivers (which, by the symptoms, I suspect
you don't.) These current pulses in the signal grounds will cause
everything
to go haywire, I'm sure.

Is there a separate logic voltage supply (like +5V or +12V) to the
drivers,
or do they derive logic power from the main motor supply? If there is a

separate PS for the driver logic, then you may need to put capacitors
from the + to - of these power input terminals. 22 to 470 uF at 16 V
would probably work fine, here.

If there is no filtering of the motor supply input to the drivers
already built
into the drivers (again, a capacitor connected right to the input
terminals)
an external capacitor of at least 100 uF, preferably 470 uF or more,
should be placed right on the motor power input terminals. This
capacitor
will supply the circulating current pulses from the switching circuits,
preventing most of that current from flowing through power supply ground

wires. I know that several other people on this group had this exact
problem, and this was a major part of the cure.

> Should I try replacing all the mosfets in one board, I did that once
> before I got a board to test right.

No, there is almost no chance any MOSFET is bad. If even one is bad,
the
motor will not step through 4 full steps, period. There is NOTHING
burned out. This is a CLASSIC EMI (Electro-Magnetic Interference)
problem.

Jon