SMPS wall wart failure.

[Phil Allison]

Arfa Daily wrote:

In short, the output electros in a SMPS are critical to it's operation and
in many cases there is nothing to prevent the output voltage going high
when they wear out.

I have seen serious damage done to 5V logic when this happens.



David - see Phil's answer regarding the choke being first. And +1 on his
observations about serious damage occurring when a cap goes high ESR and the
control loop 'lets go' in its efforts to correct for what it sees as the low
output voltage. Notable among these cases are cheap Chinese DVD players and
set-top boxes, but I have also seen it happen on equipment from what you
would normally consider to be 'reputable' manufacturers ...

** The examples I had in mind were Yamaha Multi FX processors like the SPX90 and a few relatives.

The SMPS was on its own PCB and there was enough heat to dry out the electros on the DC rails - particularly the 5V one. That rail could rise to nearly double voltage and take out a whole bunch of 74LS ICs and others that were strictly 5V.

The same PCB had the dreaded "yellow glue" dobbed all over the place eating the leads off small electros, zeners and resistors in the control loop.

After a full clean up and fitting all new electros, I added a SCR crow bar to the 5V output to protect the main board in future.



..... Phil

 

[Arfa Daily]

"Phil Allison" <pallison49@gmail.com> wrote in message
news:f5a1ba52-ed7d-4ef8-8aab-578ef04e9373@googlegroups.com...

David Farber wrote:


I was wondering what would happen if the choke were placed before the
capacitor and zener diode.

** A choke would present a high impedance in series with each current
pulse from the switching tranny - causing a big drop in the voltage
appearing on the electro cap.


Wouldn't the choke filter out the spikes and then
make the zener diode unnecessary or at least less likely to short?

** The zener (probably 6.2 V) conducted heavily and failed short when the
electro went high ESR and caused the peak output voltage to go high.

An electro that has developed high ESR cannot smooth the current pulses
being delivered by the switching tranny and diode, so the output wave has
continuous high peaks with a low *average* value. The control loop
responds to the low average and tries to correct it by making each current
pulse stronger, which only makes things worse.

In short, the output electros in a SMPS are critical to it operation and
in many cases there in nothing to prevent the output voltage going high
when they wear out.

I have seen serious damage done to 5V logic when this happens.



... Phil

David - see Phil's answer regarding the choke being first. And +1 on his
observations about serious damage occurring when a cap goes high ESR and the
control loop 'lets go' in its efforts to correct for what it sees as the low
output voltage. Notable among these cases are cheap Chinese DVD players and
set-top boxes, but I have also seen it happen on equipment from what you
would normally consider to be 'reputable' manufacturers ...

Arfa

 

[Arfa Daily]

"Phil Allison" <pallison49@gmail.com> wrote in message
news:5942d2d7-2abb-4cd1-8e4d-306032a9222a@googlegroups.com...

Arfa Daily wrote:



In short, the output electros in a SMPS are critical to it's operation
and
in many cases there is nothing to prevent the output voltage going high
when they wear out.

I have seen serious damage done to 5V logic when this happens.



David - see Phil's answer regarding the choke being first. And +1 on his
observations about serious damage occurring when a cap goes high ESR and
the
control loop 'lets go' in its efforts to correct for what it sees as the
low
output voltage. Notable among these cases are cheap Chinese DVD players
and
set-top boxes, but I have also seen it happen on equipment from what you
would normally consider to be 'reputable' manufacturers ...


** The examples I had in mind were Yamaha Multi FX processors like the
SPX90 and a few relatives.

The SMPS was on its own PCB and there was enough heat to dry out the
electros on the DC rails - particularly the 5V one. That rail could rise
to nearly double voltage and take out a whole bunch of 74LS ICs and others
that were strictly 5V.

The same PCB had the dreaded "yellow glue" dobbed all over the place
eating the leads off small electros, zeners and resistors in the control
loop.

After a full clean up and fitting all new electros, I added a SCR crow bar
to the 5V output to protect the main board in future.



.... Phil

Seems like a good move ...

Arfa