For my ten pence worth in the Linear / Switcher discussion.
I've worked on switchers for nearly 10 years and I've seen just about every
possible mode of failure and believe me if the opto coupler between the
primary and seconday sides goes open circuit you can wave goodbye to
anything connected to +5 on the output 'cos that suckers thinks the +5 is
about 0.6V and goes into overdrive to try to get the feedback (and by
default the +5) back to what it thinks is normal. The first you know about
it is a couple of loud bangs as the output caps go into orbit. We used to
call the Switcher covers 'Blast sheilds' and for good reason.
Whatever you fit you could be in for a melt down. If you want to fit a
linear fit a zener crowbar circuit across the output of it to protect your
+5 by blowing the fuse before it does to much damage.
Chris
> ----------
> From: Jess Askey[SMTP:jess@magenta.com]
> Reply To: vectorlist@mcfeeley.cc.utexas.edu
> Sent: Tuesday, September 29, 1998 4:15 PM
> To: vectorlist@mcfeeley.cc.utexas.edu
> Subject: Re: G08 power supply
>
> In my experience a linear supply is much more reliable than a switcher.
> It may weight 25 lbs. but
> it seems like if you fix all the cold solder joints and the burned molex
> connectors that are
> classic problems with the Williams power supplies (after 10 years mind
> you) you will have a nice
> rock solid beast of a power supply. I would take that over a switcher
> any day. I tried putting a
> switcher into my Major Havoc once and while it did work nicely, there
> was some noticable noise in
> the vectors.
> Also, a switcher can indeed fail and send overvoltage on the +5V, I
> have seen it happen. :-(
>
> jess
>
> Christopher V. Moore wrote:
> > Hi all, just butting in here :-)
> >
> > Correct me if I'm wrong but I thought the major point to using a
> switching
> > supply over a linear supply wasn't the increased MTBF but the fact that
> if
> > the switcher fails it stops sending any kind of voltage. This will
> protect
> > the components it feeds instead of feeding who knows what on to the
> supply
> > rails when a linear supply goes. I'd rather replace a cheap switcher
> as
> > a unit rather than have to debug/repair the game board when the linear
> supply
> > goes.
> > --
> > Christopher V. Moore -- Principal Engineeer
> > Heartlab, Inc. - 101 Airport Rd - Westerly, RI 02891 -- www.heartlab.com
> > Phone: (401) 596-0592 x113 - Fax: (401) 596-8562 - Email:
> cmoore@heartlab.com
> > "Let's get this mother out of here!" -- Last words spoken on the moon by
> > Gene Cernin, Apollo 17.
> >
>
>
Received on Tue Sep 29 10:48:16 1998
This archive was generated by hypermail 2.1.8 : Fri Aug 01 2003 - 00:32:58 EDT