Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!olivea!samsung!sdd.hp.com!hp-col!col!kenw
From: kenw@col.hp.com (Ken Wyatt)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: XT Power supply high-frequency noise?
Message-ID: <25580010@col.hp.com>
Date: 20 Mar 91 16:22:39 GMT
References: <1463@culhua.prg.ox.ac.uk>
Organization: HP Colorado Springs Division
Lines: 47

In have designed switching power supplies in one of my past lives.  I am not
sure what the switching frequency of XT power supplies is, but let's assume
that it is around 20 kHz (a common freq).  What you are likely hearing is
a subharmonic (10 kHz) or the fundamental (20 kHz - if your hearing is good!)
causing the power transformer to vibrate.  Transformer manufacturers reduce
this effect by dipping them (under pressure) into a varnish-like goo which
acts to hold all the transformer "hardware" together tightly.  Some trans-
former manufacturers skip this step to "save cost" and the consequence is
sometimes a high-pitched squeal.  By the way, this squeal could also be
coming from any of the magnetic components, such as output filter chokes;
but most likely, the power xfmr.

If you are daring (!) you may try to locate the offending "squealer" by
probing around with a PLASTIC rod (such as a BIC pen, or equivalent).
A slight pressure may stop the squeal and perhaps jamming a NON-CONDUCTIVE
object under/over/between the transformer may cause the squeal to cease.

NOTE:  To perform the above experiment requires the removal of the top lid
of the power supply (about a dozen screws) - do not remove the mounting
screws for the fan.  There may be as much as 500 VOLTS connected to the
power transformer and main switching transistor/FET (largest device near
the power transformer).  BE CAREFUL - KEEP ONE HAND IN POCKET - REMOVE ANY
RINGS - HAVE A SECOND PERSON THERE READY TO UNPLUG THE LINE CORD - ETC.

Interesting Note:  Switching power supplies may be modeled as a negative
resistance.  Thus, they often draw more power when lightly loaded than
when fully loaded.  This might explain the squealing symptom during "idle".

By the way --- my XT power supply started to squeal at start-up and then
after a couple minutes fades away until it is not objectionable.  It has
been doing this for a year now with no bad effects.  From my experience,
squealing components do not usually mean "doomsday"; they are just a bother.
One design I did for the MX missile required a switching frequency of 5 kHz.
Now THAT was a bother!  There was nothing we could do to reduce the "audio"
from this supply.

Bottom Line:  In my opinion, the squealing does not mean impending failure.
If it bothers you and you don't wish to fool around inside have it replaced.

 ----------------------------------------------------------------------- 
|  Ken Wyatt                    |         HPDesk:  Ken Wyatt / HP0800   |
|  Product Regulations Mngr     |           UNIX:  kenw@hpcol.hp.com    |
|  HP/Colorado Springs Div.     |            FAX:  (719) 590-7679       |
|  PO Box 2197                  |     Compuserve:  75665,1727           |
|  Colorado Springs, CO 80901   |  Amateur Radio:  WA6TTY               |
|  Phone:  719-590-2852         |                                       |
 -----------------------------------------------------------------------