Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!tut.cis.ohio-state.edu!att!cbnewsm!wg
From: wg@cbnewsm.ATT.COM (Bill Gieske)
Newsgroups: comp.sys.ibm.pc
Subject: Re: bad power supplies (was: Problems with my Dell)
Message-ID: <8821@cbnewsm.ATT.COM>
Date: 24 Jan 90 16:45:37 GMT
References: <75971@tut.cis.ohio-state.edu> <4417@pegasus.ATT.COM>
Organization: AT&T Bell Laboratories
Lines: 29

In article <4417@pegasus.ATT.COM>, psrc@pegasus.ATT.COM (Paul S. R. Chisholm) writes:
> Does anyone know if there's any way to test a power supply (other than
> hooking it up to an expensive mother board and seeing if it blows?-)

My power supply went last month, and how I was told to verify it was bad before
shipping it back was to measure the voltages - +5, -5, +12, -12.  If I recall
correctly the +/- 5 volts must be within 10%, or no less than 4.6V, and the
=/- 12 volts must be within 10%, or no less than 11V.  The power supply may
have a note on the case indicating which wires contain which volts.  If not,
you'll have to punt.  There is also a wire coming from the power supply that
indicates there is sufficient power coming in so that the system should con-
sider itself as capable of running - called power good (AT only?).  If this
is too low, the system will not attempt to come up.  This is supposed to pre-
vent the system from thrashing if the power level fluctuates.

Try testing the power supply with no load, with only the motherboard hooked 
up, and then with the hard disk and possibly the monitor card.  Voltage levels
should stay within tolerance.

The good news here is that although the system potentially thrashed about as 
the power supply died, nothing on the hard disk was corrupted, nor were any
cards damaged.  I sent both the power supply and motherboard back and got
replacements for both, but was not clear as to whether the motherboard was
really damaged.

Good luck.

Bill Gieske
cbnewsm!wg