Re: [CAD_CAM_EDM_DRO] Electronics gurus please help

Vajk Fekete wrote:

> hi,
>
> i am building a stepper driver for my homemade mill, that i am going to use with EMC.
>
> i have salvaged some steppers (60ozin, 6 wire), and also the driver chip from old printers.
> the driver is a sla7026, 3A max chopper, unipolar.
>
> question1: is a chopper unipolar driver much worse than a chopper bipolar drive?

It depends entirely on how you do it. The general advantage of the bipolar drive
is it applies the full supply voltage across the winding backwards to reverse the current
in the winding quickly. This is the generally the fastest way to reverse the current in the
winding. With a unipolar setup, you have current in one half of the winding, and to reverse
that phase, you have to stop the current in that half, and start the current flowing in the
other half. If you had high-voltage transistors and zener diodes to absorb the energy
in the shutting-off winding, you might be able to do it about as fast, but you are taking
energy from the windings and turning it into heat at every step. With the bipolar drive,
you recycle much of the energy in the windings back through diodes, the power
supply, and finally back into those (or other) windings. So, energy is recycled, rather
than heating up a bunch of expensive parts that then need big heatsinks, etc.

> question2: a bit longer, and i badly need advice here:
>
> i have built an axis drive on veroboard, with rather long wires, etc, and it worked ok.
> then i made a rather simple pcb, built a sigle axis on that, and after correcting some errors, it works ok.
> then i made a final design for the pcb, built one axis, and it does not work. built another, and the with same symptoms it is not working.
>
> so experts: at a chopping frequency a couple 10Khz, could the pcb layout have this much influence? i checked all the components, all the connections, spent about 2 days on it, so i am 98% that the
> working and not working circuit only differs in the shape of the traces.

Well, just because your chopping frequency is tens of KHz, doesn't mean that the energy
content of the signals is in that region only. If the current and/or voltage changes are
abrupt, the frequency domain can range to the hundreds of MHz! Power FETs turn
on and off in 40 -150 nS, and can generate strong RF pulses containing strong energy
components in the 10 - 300 MHz range, which can couple VERY well from small lengths
of traces to other traces, driving the logic haywire. If you are using power FETs, the first
thing to try is to slow them down with series resistors in the gate leads. If you are
using bipolar transistors, small ferrite beads on the base leads may help. Small resistors
in the base lead may also be helpful.

>
> i am no EE, but have some electronics building experience, more from the digital world. and i do not have a scope, so cannot check if the timer is working ok or not.

Without a scope of some sort, you are just blind, unfortunately. An experienced hand
may be able to try a few good tricks without being able to see the problems in detail
with a scope, but the amount of time saved in diagnosing these sorts of gremlins rapidly
makes the cost of a second-hand scope worthwhile.

Jon