Power Supply keep blowing

[manny]

I have never seen a switcher with the PWM on the low side.

Yes more inductors can be added to the galvanic isolation.
But that is on the order of something under ten dollars per inductor.

Something like at least $9 under. If the transformers were
anywhere close to $10 they would not be used by so many of
the worst, most cost-conscious manufacturers, such as Deer,
in their products retailing for $10-20, yet they are.

 

[w_tom]

Excellent post. It's blunt. It's technically accurate. It includes
reams of supporting facts. The SG6105 includes features beyond what is
required including AC voltage monitor and negative voltage OVP. Also
includes the necessary delay timing for Power Good, for other various
inputs and outputs, apparently those timings are generated by a well
defined internal oscillator, and some TL431 voltage references. I
don't see where it describes how/if error states are latching events or
how those error states are cleared. Where does it describe how errors
are latched or cleared?

Transformers costs were one reason why few functions were exchanged
across the galvanic isolation barrier. Even telephone type
transformers costs $4. But with so many functions (including delays)
integrated inside this single IC, then higher costs of additional
transfomers would be negated by less components - both labor and
material costs reduced.

This your longer post truly is a benchmark answer. One of the best
replies I have read for sometime. BTW, which hyperlink was I suppose
to look at from terasan.info?

 

[larry moe 'n curly]

larry moe 'n curly wrote:

I have never seen a switcher with the PWM on the low side.

The SG6105 includes features beyond what is required including
AC voltage monitor and negative voltage OVP. Also includes the
necessary delay timing for Power Good, for other various inputs
and outputs, apparently those timings are generated by a well
defined internal oscillator, and some TL431 voltage references.
I don't see where it describes how/if error states are latching
events or how those error states are cleared. Where does it
describe how errors are latched or cleared?

The www.sg.com.tw website also has an application note for this chip
that may answer your questions, but I didn't look at it. However I
have a PSU built around the chip, and the overcurrent protection did
latch when I intentionally created a short, and I couldn't restart the
PSU until I first pulled the AC cord.

Transformers costs were one reason why few functions were exchanged
across the galvanic isolation barrier. Even telephone type
transformers costs $4. But with so many functions (including delays)
integrated inside this single IC, then higher costs of additional
transfomers would be negated by less components - both labor and
material costs reduced.

I looked in a Mouser Electronics catalog and found lots of small
transformers listing for about $2 in single quantities, so I would
expect a big PSU maker to be able to get them for much less. Also
almost every cheapo ATX PSU I've owned contained three transformers
(optical feedback, no transformer for this), meaning they were
regulated from the low side. And I think that every AT PSU I've had
contained two transformers, excluding the one for feedback (lots used a
transformer for this instead of an optical isolator), and because AT
PSUs don't have a standby voltage, doesn't this imply control from the
low side? Almost of these PSUs, AT and ATX, were controlled by either
a KA7500 or TL494 chip located very close to the low voltage
components..

BTW, which hyperlink was I suppose to look at from terasan.info?

The one in Japanese. Try Babelfish translation -- the laughs
alone are worth the effort. I included that link for people who
wanted to verify that it was common for ATX PSUs to have the regulator
chip on the low voltage side.

 

[w_tom]

Interesting that your experience contradicts what I thought
its schematic said. Doing only a quick logical assessment, an
error condition would be cleared by cycling the PS-ON switch.
However 'reset by power cord removal' might be a function of
whatever controls that PS-ON line. Or does an internal
lockout function exist that is not indicated on that
schmatic? Or I misread the logic.

Am not concerned enough to search application notes for
details. But your post that says, "the overcurrent protection
did latch when I intentionally created a short" suggests I
should watch for this (that might be a) new function in power
supplies that use this and equivalent chips.

If I don't hear otherwise, then I guess any one of those 30+
hypelinks in terasan.info were examples. As long as a
schematic is provided, then text is not important - unless I
misread schematic logic again.

Again, these wonderful little gems - your post with numbers
from a datasheet - make newsgroups informative and useful.