Yes, one can calculate the peak current. I think the overlaying issue is that lead
length between the inductive reactance of the motor coil (or a relay) and the diode
should be kept as short as possible to prevent unwanted RF emission, not just
device protection. This is why you may see diodes connected right across the motor
(or coil), and not on the circuit board. There are some dip relays that contain a
diode encapsulated within the package itself. It is far cheaper to place an
external diode in the circuit near the load than the whole ULN device should it
fail.
~Rick

Spehro Pefhany wrote:

> At 01:29 PM 2/28/2004 -0500, you wrote:
> >Never rely on internal diodes to do the job of external snubbing. There's
> >no way
> >the substrate of the ULN device can handle the spike supression of any motor.
> >The internal diodes are there as a last resort, not the first line of defense.
> >~Rick
>
> Can you predict how much the peak current through the catch diodes (which
> are not substrate diodes, BTW) will conduct, given the motor coil current
> of, say, 170mA per phase?
>
> Best regards,
>
> Spehro Pefhany --

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.