On 26.02.2012 19:08, Mouse wrote:
>>> What i see are very very short (<1us) drops of the 12 down that are
>>> larger than 1V. [...]
>
> It occurs to me that perhaps the pulses are real and the underlying
> problem is that there's a hardware fault causing the decoupling to not
> work as well as it should. Anything from a broken etch run to a
> pick-and-place failure causing that particular board to not have all
> the decoupling it's supposed to could cause this - I'm speculating, of
> course, but such failures _do_ happen.
OTOH even if the 1 V voltage drops were real, they most probably
would not cause problems. While being worse than usual +- 5%
spec of a PSU, any sane circuit should handle -10%.
Soren Kristensen wrote:
> If you supply 11-13V to the unit then a mosfet switch direct
> the power input directly to the 12V rail, you can then draw
> max 2A.
>
> Otherwise a dc-dc converter make 12V from 5V
An interesting fact is that Attila is seeing pulses with a 12 V
supply and not seeing them with 16 V one.
If this circuit is misbehaving due to some noise, out of spec part,
some funny coupling with the PSU circuitry or other problem and
"cannot decide" how to make the 12V, such pulses are definitely
possible. Is there any hysteresis in this logic and if yes,
how large?
> But I also know that from my own experience that if you just
> take a standard scope probe with the typical 10 cm ground
> lead, that 10 cm is very good at picking up unwanted pulses....
Agreed, this is also my experience.
And at this load a capacitance required to safely filter out
< 1 us pulses is on the order of microfarads - so unless
there is something producing huge loads momentarily I don't
really think there is a way to produce > 1 V drops with
multiple caps along the path.
Regards
--
Stano
_______________________________________________
Soekris-tech mailing list
[email protected]
http://lists.soekris.com/mailman/listinfo/soekris-tech
