Power Supply recomendation?

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

Wes Newell <w.newell@TAKEOUTverizon.net> wrote in message news:<pan.2004.06.10.19.28.06.691636@TAKEOUTverizon.net>...
> On Thu, 10 Jun 2004 05:36:17 -0700, larrymoencurly wrote:

> > Considering that many cheap 500W+ PSUs look less substantial
> > than many of the better 300W PSUs, I find it hard to believe
> > that a 550W selling for $15 can be trusted when lots of power
> > is needed, unless it was a surplus special.
>
> So now you want me to to provide data for all cheap power
> supplies. I wouldn't hold my breath.
>
> Now, since you say many cheap 500W PSU's are less substantial
> than many of the better 300W's, why don't you prvide the data
> to prove this. Rediculous as that sounds, it's exactly what
> was asked of me. So for get it.

Compare this: www.bit-tech.net/images/review/123/7.jpg

to this: http://terasan.okiraku-pc.net/dengen/no20/open.jpg
http://terasan.okiraku-pc.net/dengen/no20/up02.jpg

(lots of PSU internals pictured at
http://terasan.okiraku-pc.net/dengen/ )

If you didn't know anything else about either PSU, which one would you
think had the higher power rating? The first PSU is a 550W Q-tec
while the second is a 300W Powerman (Fortron-Source, also makes
Sparkle), and unless the Q-tec's transformer has higher capacity for
the same dimensions or is run at a higher frequency, how can it put
out more power than the Powerman's? Similarly, unless the Q-tec is
more efficient (fat chance), won't its heatsinks likely run hotter at
any given power level than the Powerman's? But most of all, why do
companies like PC Power & Cooling, Antec, and Fortron-Source use
bigger capacitors, heatsinks, and transformers if they're unnecessary,
especially when most customers don't care about the insides?

Except for an Enermax, all the PSU failures I've experienced have been
with cheapos. One didn't have good overload protection for the +3.3V
(almost identical to the one shown by PC Power & Cooling as an example
of a bad PSU), a low voltage capacitor in a Powmax failed (could have
been one of those made with faulty electrolyte), and a Deer shorted
several 16V capacitors, probably because it sent 50V spikes into them.
And that failed Enermax was badly made 250W without an anti-surge
thermistor.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On 11 Jun 2004 07:25:12 -0700, larrymoencurly@my-deja.com (larrymoencurly)
wrote:

>Except for an Enermax, all the PSU failures I've experienced have been
>with cheapos. One didn't have good overload protection for the +3.3V

I once plugged the cables in backwards to an internal USB header. The PSU
breaker kept tripping, and the board kept re-booting. It's AMAZING that
something wasn't smoked. That sold me on Enlight PSU's and Gigabyte boards!

So... what did YOU plug in backwards?

Bob

Remove "kins" to reply by e-mail.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On Fri, 11 Jun 2004 11:32:46 -0400, w_tom wrote:

> Wes again posts as the classic naive 'expert'. Just because
> his computer turned on means everything is 100% correct?

As with any recommendation, you can take it or leave it. It works for me,
100% of the time. Every voltage is within tolerances. To me, that means it
works. What happens when it fails is yet to be seen, but I've never had a
failed PSU take out hardware. You have, too bad.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.htm

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On Fri, 11 Jun 2004 19:08:23 GMT, Wes Newell <w.newell@TAKEOUTverizon.net>
wrote:

>On Fri, 11 Jun 2004 11:32:46 -0400, w_tom wrote:
>
>> Wes again posts as the classic naive 'expert'. Just because
>> his computer turned on means everything is 100% correct?
>
>As with any recommendation, you can take it or leave it. It works for me,
>100% of the time. Every voltage is within tolerances. To me, that means it
>works. What happens when it fails is yet to be seen, but I've never had a
>failed PSU take out hardware. You have, too bad.

PSU's normally fail with low or no voltage and current, not a spike in
voltage or current. So it stands to reason that failures are annoying but
not dangerous.

Hey, instead of spending $60, maybe it makes sense to buy 4 of the $15 PSU's
(as long as they make adequate clean power)

Bob

Remove "kins" to reply by e-mail.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On Fri, 11 Jun 2004 11:32:46 -0400, w_tom wrote:

> Wes again posts as the classic naive 'expert'. Just because
> his computer turned on means everything is 100% correct?

And second, I never claimed to be a power supply expert. Although I have
lots of experiences with them among many other things.

--
Abit KT7-Raid (KT133) Tbred B core CPU @2400MHz (24x100FSB)
http://mysite.verizon.net/res0exft/cpu.htm

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On 11 Jun 2004 07:24:52 -0700, larrymoencurly@my-deja.com
(larrymoencurly) put finger to keyboard and composed:

>But most of all, why do
>companies like PC Power & Cooling, Antec, and Fortron-Source use
>bigger capacitors, heatsinks, and transformers if they're unnecessary,
>especially when most customers don't care about the insides?

Here is a PC Power & Cooling 450W ATX PSU:
http://terasan.okiraku-pc.net/dengen/no41/index.html

These photos show two 200V 1000uF 85 degC caps parked next to
heatsinks:

http://terasan.okiraku-pc.net/dengen/no41/open.jpg
http://terasan.okiraku-pc.net/dengen/no41/side01.jpg

A quality PSU would use 105 degC capacitors and would locate them away
from sources of heat, if at all possible.

Here's a 250W Nipron PCSA-300P-X2S PSU that uses a 105 degC cap:
http://terasan.okiraku-pc.net/dengen/no09/index.html
http://terasan.okiraku-pc.net/dengen/no09/main2.jpg

>Except for an Enermax, all the PSU failures I've experienced have been
>with cheapos.

Perhaps it's because "cheapos" outnumber the more expensive brands.
Ten times as many cheapos means ten times as many failures, all things
being equal. Or perhaps it's because the people who buy expensive PSUs
opt for the higher rated ones.

> One didn't have good overload protection for the +3.3V
>(almost identical to the one shown by PC Power & Cooling as an example
>of a bad PSU),

I'm intrigued as to what PC Power & Cooling considers a bad PSU. Can
you elaborate on this? Any URLs, photos, diagrams?

> a low voltage capacitor in a Powmax failed (could have
>been one of those made with faulty electrolyte), and a Deer shorted
>several 16V capacitors, probably because it sent 50V spikes into them.
> And that failed Enermax was badly made 250W without an anti-surge
>thermistor.

.... which proves that a higher price is no guarantee of quality.


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

Franc Zabkar <fzabkar@optussnet.com.au> wrote in message news:<uljkc0h6a9hul9m56uce8j0419j09190cb@4ax.com>...

> Here is a PC Power & Cooling 450W ATX PSU:
> http://terasan.okiraku-pc.net/dengen/no41/index.html
>
> These photos show two 200V 1000uF 85 degC caps parked next to
> heatsinks:
>
> http://terasan.okiraku-pc.net/dengen/no41/open.jpg
> http://terasan.okiraku-pc.net/dengen/no41/side01.jpg
>
> A quality PSU would use 105 degC capacitors and would locate them away
> from sources of heat, if at all possible.

> Here's a 250W Nipron PCSA-300P-X2S PSU that uses a 105 degC cap:
> http://terasan.okiraku-pc.net/dengen/no09/index.html
> http://terasan.okiraku-pc.net/dengen/no09/main2.jpg

I don't have any PSUs that automatically adjust to the AC voltage the
way that one does, but I noticed that all of mine, good and bad, use
85C high voltage capacitors. Is this because those caps work only at
low frequency and don't have square waves passing through them?

> > Except for an Enermax, all the PSU failures I've experienced
> > have been with cheapos.
>
> Perhaps it's because "cheapos" outnumber the more expensive brands.
> Ten times as many cheapos means ten times as many failures, all things
> being equal. Or perhaps it's because the people who buy expensive PSUs
> opt for the higher rated ones.

In my case they failed at much higher rates than the better ones did.

> > One didn't have good overload protection for the +3.3V
> >(almost identical to the one shown by PC Power & Cooling as
> > an example of a bad PSU),
>
> I'm intrigued as to what PC Power & Cooling considers a bad PSU. Can
> you elaborate on this? Any URLs, photos, diagrams?

www.pcpowercooling.com/products/power_supplies/insidestory/

I think that the comparison is a bit unfair because it's between a
425W PCP&C and a 300W King Case/King Star, although this is disclosed.
On my 250W King Star the high voltage filters were just 220-330uF,
the +12V wasn't regulated, and there was no RFI filter at all (several
empty spots on circuit board, tons of AM radio interference). Also
that small circuit board on top of the heatsinks for the +5V standby
regulator was held in place with just a single screw, and when that
screw was loose it could almost short to the top of the case. I don't
remember if mine had an anti-surge thermistor.

PCP&C also has this comparison, between their 510W and an Enermax
550W:

www.pcpowercooling.com/pdf/Turbo-Cool_510_vs.pdf

> > a low voltage capacitor in a Powmax failed (could have
> >been one of those made with faulty electrolyte), and a Deer shorted
> >several 16V capacitors, probably because it sent 50V spikes into them.
> >And that failed Enermax was badly made 250W without an anti-surge
> >thermistor.
>
> ... which proves that a higher price is no guarantee of quality.

I shouldn't have included my Powmax because any PSU can have a bad
capacitor. But I thought that Powmax was considered bad, even if it
was manufacturered several years ago, before the brand was cheapened a
lot.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On 13 Jun 2004 05:30:13 -0700, larrymoencurly@my-deja.com (larrymoencurly)
wrote:


>
>I don't have any PSUs that automatically adjust to the AC voltage the
>way that one does, but I noticed that all of mine, good and bad, use
>85C high voltage capacitors. Is this because those caps work only at
>low frequency and don't have square waves passing through them?

It is less important for those to be of high temp grade, because there is
less change in current at that point, voltage. Their ESR is not so
significant a factor in "self" heating.


>
>> > Except for an Enermax, all the PSU failures I've experienced
>> > have been with cheapos.
>>
>> Perhaps it's because "cheapos" outnumber the more expensive brands.
>> Ten times as many cheapos means ten times as many failures, all things
>> being equal. Or perhaps it's because the people who buy expensive PSUs
>> opt for the higher rated ones.
>
>In my case they failed at much higher rates than the better ones did.

The oddity of an Enermax, is that it's build quality is optimized towards
"looking" good rather than performing well. Most name-brands would spend
the $ on electrical design before adding gold-colored anodized heatsinks
or cable sheaths, etc.

>
>> > One didn't have good overload protection for the +3.3V
>> >(almost identical to the one shown by PC Power & Cooling as
>> > an example of a bad PSU),
>>
>> I'm intrigued as to what PC Power & Cooling considers a bad PSU. Can
>> you elaborate on this? Any URLs, photos, diagrams?
>
>www.pcpowercooling.com/products/power_supplies/insidestory/
>
>I think that the comparison is a bit unfair because it's between a
>425W PCP&C and a 300W King Case/King Star, although this is disclosed.
> On my 250W King Star the high voltage filters were just 220-330uF,
>the +12V wasn't regulated, and there was no RFI filter at all (several
>empty spots on circuit board, tons of AM radio interference). Also
>that small circuit board on top of the heatsinks for the +5V standby
>regulator was held in place with just a single screw, and when that
>screw was loose it could almost short to the top of the case. I don't
>remember if mine had an anti-surge thermistor.

Sadly, they get even worst than the one pictured, it has a full frame
rigid fan... probably sleeve bearing but likely better balanced.

>
>PCP&C also has this comparison, between their 510W and an Enermax
>550W:
>
> www.pcpowercooling.com/pdf/Turbo-Cool_510_vs.pdf
>
>> > a low voltage capacitor in a Powmax failed (could have
>> >been one of those made with faulty electrolyte), and a Deer shorted
>> >several 16V capacitors, probably because it sent 50V spikes into them.
>> >And that failed Enermax was badly made 250W without an anti-surge
>> >thermistor.
>>
>> ... which proves that a higher price is no guarantee of quality.
>
>I shouldn't have included my Powmax because any PSU can have a bad
>capacitor. But I thought that Powmax was considered bad, even if it
>was manufacturered several years ago, before the brand was cheapened a
>lot.

Yes, todays power supplies follow same trend as yesteryears',
manufacturers seem to target certain market segments, seldom jump from
producing good units to junk or vice-versa.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On 13 Jun 2004 05:30:13 -0700, larrymoencurly@my-deja.com
(larrymoencurly) put finger to keyboard and composed:

>Franc Zabkar <fzabkar@optussnet.com.au> wrote in message news:<uljkc0h6a9hul9m56uce8j0419j09190cb@4ax.com>...
>
>> Here is a PC Power & Cooling 450W ATX PSU:
>> http://terasan.okiraku-pc.net/dengen/no41/index.html
>>
>> These photos show two 200V 1000uF 85 degC caps parked next to
>> heatsinks:
>>
>> http://terasan.okiraku-pc.net/dengen/no41/open.jpg
>> http://terasan.okiraku-pc.net/dengen/no41/side01.jpg
>>
>> A quality PSU would use 105 degC capacitors and would locate them away
>> from sources of heat, if at all possible.
>
>> Here's a 250W Nipron PCSA-300P-X2S PSU that uses a 105 degC cap:
>> http://terasan.okiraku-pc.net/dengen/no09/index.html
>> http://terasan.okiraku-pc.net/dengen/no09/main2.jpg
>
>I don't have any PSUs that automatically adjust to the AC voltage the
>way that one does, but I noticed that all of mine, good and bad, use
>85C high voltage capacitors. Is this because those caps work only at
>low frequency and don't have square waves passing through them?

During their charging cycle these caps draw very large current pulses
at twice the mains frequency (100Hz or 120Hz). During the discharge
cycle the caps are supplying smaller current pulses to the chopper
transistor(s) at the switchmode frequency (40kHz - 100kHz ?). As Kony
states, these caps would have a relatively low ESR, so they would
probably not be subject to the same self heating effects as smaller
electrolytics. And space *is* a serious constraint in PC PSUs, so
proximity to heatsinks is probably unavoidable. IME the most common
problem with switchmode PSUs in general is the smaller electros on the
primary side of the supply. When these leak and/or develop high ESR,
the PSU will fail to regulate properly.

Having said that, the additional cost to the manufacturer of a 105
degC cap over an 85 degC one would have been pennies.

>> > Except for an Enermax, all the PSU failures I've experienced
>> > have been with cheapos.
>>
>> Perhaps it's because "cheapos" outnumber the more expensive brands.
>> Ten times as many cheapos means ten times as many failures, all things
>> being equal. Or perhaps it's because the people who buy expensive PSUs
>> opt for the higher rated ones.
>
>In my case they failed at much higher rates than the better ones did.
>
>> > One didn't have good overload protection for the +3.3V
>> >(almost identical to the one shown by PC Power & Cooling as
>> > an example of a bad PSU),
>>
>> I'm intrigued as to what PC Power & Cooling considers a bad PSU. Can
>> you elaborate on this? Any URLs, photos, diagrams?
>
>www.pcpowercooling.com/products/power_supplies/insidestory/

Hmmm, it's technical, but it still sounds like little more than a
sales pitch. And I notice that not all PC Power & Cooling PSUs will
have all these touted features.

>I think that the comparison is a bit unfair because it's between a
>425W PCP&C and a 300W King Case/King Star, although this is disclosed.

Well, that's marketing for you. Notice that PCP&C also have 300W
"cheapo" PSUs which they euphemistically refer to as their "economy"
line. ;-)

> On my 250W King Star the high voltage filters were just 220-330uF,

I bet they are no smaller than what you'll find in PCP&C's "standard
250 ATX":
http://www.pcpowercooling.com/pdf/econ-atx.pdf

Notice that some specs are no better than what you would expect from a
"cheapo" PSU:

Operating Range: 95-132 VAC
Regulation: 5% (+3, +5, +12)
10% (-5, -12)

Then there's this 300W PSU:
http://www.pcpowercooling.com/pdf/T301U.pdf

The regulation specs and safety features (OC & OV protection) are
fairly basic.

The same can be said for the 300 ATX-PFC:
http://www.pcpowercooling.com/pdf/hp-atx-pfc.pdf

This PSU has OV protection on the +3.3V and +5V rails, but not on the
+12V rail.

>the +12V wasn't regulated, ...

I find this hard to understand. All the supplies I have seen regulate
by sensing a weighted average of the +5V and +12V rails. That's why
the +5V rail is affected by a change in the +12V load (eg extra HDs).
FWIW, I sometimes adapt PC PSUs for 6V or 13.8V operation by removing
and recalculating the sense resistors.

>and there was no RFI filter at all (several
>empty spots on circuit board, tons of AM radio interference).

Yep, that's bad. I presume that the empty spots would normally have
been populated by inductors and capacitors in a pi configuration?

> Also
>that small circuit board on top of the heatsinks for the +5V standby
>regulator was held in place with just a single screw, and when that
>screw was loose it could almost short to the top of the case. I don't
>remember if mine had an anti-surge thermistor.

I've had a PSU with a rivet rattling around inside. I don't recall the
brand, though.

>PCP&C also has this comparison, between their 510W and an Enermax
>550W:
>
> www.pcpowercooling.com/pdf/Turbo-Cool_510_vs.pdf
>
>> > a low voltage capacitor in a Powmax failed (could have
>> >been one of those made with faulty electrolyte), and a Deer shorted
>> >several 16V capacitors, probably because it sent 50V spikes into them.
>> >And that failed Enermax was badly made 250W without an anti-surge
>> >thermistor.
>>
>> ... which proves that a higher price is no guarantee of quality.
>
>I shouldn't have included my Powmax because any PSU can have a bad
>capacitor. But I thought that Powmax was considered bad, even if it
>was manufacturered several years ago, before the brand was cheapened a
>lot.

You should never buy on brand alone. Some respected manufacturers will
produce lesser quality products and sell them for more than they are
worth simply by cashing in on their reputation. For example, I've seen
the same mechanisms inside Sony and Palsonic VCRs, and in early Sonys
and Sanyos. US Robotics, a respected modem manufacturer, is now
rebranding Conexant softmodem chipsets. Badge engineering is one of my
pet hates.

Incidentally, I've been through PCP&C's online PSU selection process.
The recommended PSU for my purposes was the Turbo-Cool 300 ATX-PFC.
Their worst case home scenario, with max HDs and max RAM, requires
only a Turbo-Cool® 425 ATX-Deluxe. Curiously, the on-line selector
does not ask for the CPU type, or the type of graphics card, but it
does ask how much memory will be used. I would have thought that power
consumption for memory was comparatively small. <shrug>


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

Franc Zabkar <fzabkar@optussnet.com.au> wrote in message news:<r6ctc01iurk2apgicsp0oga3hp40686kkh@4ax.com>...
> On 13 Jun 2004 05:30:13 -0700, larrymoencurly@my-deja.com
> (larrymoencurly) put finger to keyboard and composed:

> >www.pcpowercooling.com/products/power_supplies/insidestory/

> >the +12V wasn't regulated, ...
>
> I find this hard to understand. All the supplies I have seen regulate
> by sensing a weighted average of the +5V and +12V rails. That's why
> the +5V rail is affected by a change in the +12V load (eg extra HDs).
> FWIW, I sometimes adapt PC PSUs for 6V or 13.8V operation by removing
> and recalculating the sense resistors.

I traced it out and found that the +12V output consisted of just a
transformer winding, 1-2 diodes, and some filter capacitors and coils.
The output wasn't connected to any feedback circuit, so if the +12V
was regulated it must have been done on the primary side.

> >and there was no RFI filter at all (several
> >empty spots on circuit board, tons of AM radio interference).
>
> Yep, that's bad. I presume that the empty spots would normally have
> been populated by inductors and capacitors in a pi configuration?

Yes, and transplanting some from an old AT PSU eliminated all the
noticeable interference.

 

[Guest]

Archived from groups: alt.comp.hardware.homebuilt (More info?)

On 15 Jun 2004 10:28:28 -0700, larrymoencurly@my-deja.com
(larrymoencurly) put finger to keyboard and composed:

>Franc Zabkar <fzabkar@optussnet.com.au> wrote in message news:<r6ctc01iurk2apgicsp0oga3hp40686kkh@4ax.com>...
>> On 13 Jun 2004 05:30:13 -0700, larrymoencurly@my-deja.com
>> (larrymoencurly) put finger to keyboard and composed:
>
>> >www.pcpowercooling.com/products/power_supplies/insidestory/
>
>> >the +12V wasn't regulated, ...
>>
>> I find this hard to understand. All the supplies I have seen regulate
>> by sensing a weighted average of the +5V and +12V rails. That's why
>> the +5V rail is affected by a change in the +12V load (eg extra HDs).
>> FWIW, I sometimes adapt PC PSUs for 6V or 13.8V operation by removing
>> and recalculating the sense resistors.
>
>I traced it out and found that the +12V output consisted of just a
>transformer winding, 1-2 diodes, and some filter capacitors and coils.
> The output wasn't connected to any feedback circuit, so if the +12V
>was regulated it must have been done on the primary side.

I agree that it was probably unregulated. If so, then IMHO your PSU's
design makes more sense than "typical" designs. By that I mean that
the +5V rail should be accurately regulated because the +5V logic
depends on it. OTOH, the +12V rail may be permitted to vary a great
deal before it will cause problems for those devices that use it, eg
HDs and COM ports.

See this typical (?) "generic" PSU design which senses both the +5V
(R26) and +12V (R25) rails:

http://www.pavouk.comp.cz/hw/en_atxps.html

>> >and there was no RFI filter at all (several
>> >empty spots on circuit board, tons of AM radio interference).
>>
>> Yep, that's bad. I presume that the empty spots would normally have
>> been populated by inductors and capacitors in a pi configuration?
>
>Yes, and transplanting some from an old AT PSU eliminated all the
>noticeable interference.

You must have loved that cheapo PSU to go to all that trouble. ;-)


- Franc Zabkar
--
Please remove one 's' from my address when replying by email.