Re: Building computer - power supplies
On Sat, Sep 28, 2013 at 2:29 AM, Doug <dmcgarrett@optonline.net> wrote:
> On 09/27/2013 09:21 AM, Ralf Mardorf wrote:
> /snip/
>
>> However, undersized voltage for capacitors for switching power supply is
>> an often used fraud against consumers and switching power supplies
>> pollute the mains. Exotic resistors sometimes make them a PITA when you
>> want to repair one and you don't have those resistors on stock. I
>> dislike all switching power supplies.
>>
>> Regards,
>> Ralf
>>
>>
> Not sure what you mean by "undersized voltage for capacitors."
Ralf seems to speak German more natively than English, and sometimes
he reverses grammatical order. I think, from the general tenor of his
posts, he intended to say something like, "capacitors for undersized
voltages."
Hmm. Anyway, what you would call capacitors rated at voltages too low
for the circuit. Underrated in the bad sense.
My impression is the specification rating in the power supply design
is less the culprit than the capacitor manufacturer pushing the edges
on their QC. But the upshot is that capacitors are exposed to higher
voltages and/or effective power than they can handle, and get burned,
and it is a manufacturing problem, and sometimes an engineering
problem.
> Good
> enigineering practice says that there should be plenty of headroom on
> capacitor voltage rating. IOW, if you're filtering a 15 volt supply,
> you shouldn't use a 16 volt capacitor, but at least a 20 Volt one.
> Or more. MIL-specs require double the working voltage, in most cases.
Well, you definitely don't generally see double in consumer designs.
Sometimes I've seen 5V rated caps filtering 5V power output in bad
switching designs.
> While it's true that switchers produce RF interference, that can be
> handled by proper shielding of the supply and filtering of the AC line,
> usually using ferrite components.
And consumer grade power supplies often cut corners on their design,
sometimes even failing to use ferrite ferrite. (Ever seen a fake
ferrite? (Fool's gold? heh.)
> Modern power supplies for sensitive
> radio receivers are designed using switching techniques and proper
> filtering, and the receivers work without interference.
>
> A further point: Common computer supplies are rated around 450 Watts.
> You wouldn't want to lift the transformer that would supply 450 Watts,
> much less take up the space for it. So switching supplies are the only
> reasonable alternative.
Yeah, linear is no panacea. (From my limited hobby experience, I can
tell stories of linear power supplies that were way noisier than your
common marginal design switching PS. )
The problem (for anyone reading who is not aware of this) is not
switching vs. linear, it's the sales and manufacturing practices
encouraged by stingy boards of directors.
> --Doug, retired electronic engineer and radio amateur, WA2SAY
>
>
> --
> Blessed are the peacemakers..for they shall be shot at from both sides.
> --A.M.Greeley
--
Joel Rees
Be careful where you see conspiracy.
Look first in your own heart.
Reply to: