Mark said:
No need to kow-tow to Tom's superiority, man. My recent problems led me
down the same path, and I picked up a meter to do some testing. Well,
the instructions for it were horribly written and assumed a significant
knowledge of electric circuitry. Figuring out how to even set the thing
up to test anything was almost beyond me. The information I could find
on the web was very limited, but what I did find there were plenty of
warnings about the dangers of tyros digging around in hot circuitry with
these things. And Tom provides only enough information to make you
dangerous. I'm sure one can determine if a power supply is good with one
of these, and more about the mobo as well, and I do want to explore its
use further, when I have more time. But right now Tom's
apparently-habitual approach to the whole thing with the totally
uninitiated strikes me as quite irresponsible.
Just my opinion...
To give a little more help, than Tom might be comfortable giving, here
is a meter primer. HTH.
******* Using a multimeter *******
The multimeter has lots of ranges and inputs, but you can use just the
ones that are easy to understand to start with. For example, my multimeter
has a "transistor beta" feature, where you are supposed to stuff the leads
of a transistor, into a bunch of tiny holes on the face of the meter, and
I've never used that feature. Maybe it is in the manual, but I've never had
need of it. Since I don't need it, I don't touch it. No big deal.
This meter has four holes on the front, for meter leads. Everything is
referenced to the "Common" or COM terminal. In this picture, COM is in the
lower right corner, and usually is black in color. You can start by plugging
the black test lead into the COM hole. That is your reference point that the
meter will measure the voltage with respect to.
http://mechatronics.mech.northwestern.edu/design_ref/tools/multi-jacks.gif
The two on the left are for measuring current, and can be ignored.
To start with, you want to try simple things, like measuring voltage.
The red lead plugs into the "Volts, Ohms, Diode Test" hole. The
black goes to the COM hole. You can probe with the meter, while it
is in volts or "voltage measuring mode", without much danger. If you
plugged the red lead into one of the current holes, and dropped the
meter leads on a live circuit, that would be a close to a dead short,
so you have to be a bit more careful with the current functions. And
the dead short in the current measuring ranges, is present even when
the meter is turned off. So just keep your red lead out of the "Amps"
or "Milliamps" holes for now. In fact, I seldom use the "Amps" or
"Milliamps" holes, and the test leads stay in the same two holes (on
the right) almost all the time.
OK, with the leads plugged in, you select a safe operating range.
Meters can be "autoranging" or "manual". Mine is manual, because
I like them that way (they respond faster to changes). I've also
used plenty of different autoranging meters, and they're fine too.
If I was working in a PC, and the scale on the meter offered 1000V,
200V, 20V, 2V, 200m as options, I would select 20V range. That is
because the most positive and negative voltages expected on a PC are
+12V and -12V. Both should be readable within that range. Since I've
also recently been measuring the A.C. voltage in my house, I use 200V
for that, since my house voltage is supposed to be 113VAC or so,
according to Hydro.
So, we've plugged in the probes, set the meter to 20V. Well, what
function do we want ? It would be Volts DC, because direct current
or DC is what comes out of the PC power supply. VDC means we
are expecting a "flat", "non-varying" voltage. (AC, like my house
power, is time varying, and I'd select the volts A.C. function for
that.)
Once you selected a range which is slightly larger than the thing to
be measured, you can switch on the meter. For volts ranges, zero is
the normal, unused reading. On ohms, the meter basically shoots
offscale, and would give an overrange indication, and the display
might flash. But for volts, we get a calming zero value, that won't
jangle the nerves.
When working in the PC, you really don't want your hands full. You can
buy alligator clips at Radio Shack, and I have small ones and slightly
larger ones. The small ones snap onto the tips of my meter probes
(after some bending with pliers
). You can also get meter leads with
an alligator clip already on the ends. Using an alligator clip, I clip
the black meter lead (COM) to a screw on one of the I/O connectors
on the back of the computer. By doing that, I only need one hand to
make measurements. The screw on the I/O connector is at ground
potential.
Using the red lead, you probe where you want to measure a voltage.
It may take the meter several readings to settle down. If I was
reading the 12V rail, I might see "+12.23", which means I made
good use of my meter's output digits. If the display read " 12",
which is what I'd see on the 1000V DC scale, I'd know to increase
the sensitivity of the meter, by selecting the next range down,
which is 200V. Then I'd see "+ 12.2". And finally, when I select
the 20V scale, I'd see "+12.23", which is all the digits my meter
has got.
The power supply has a label on the side, which will mention a number
of useful things. It might tell you the expected uncertainty of its
output voltages, like +/- 5%. On the 12V rail, the expected voltage
ranges would be +11.4 VDC to +12.6 VDC. The +12.23 value I just
measured, is within the allowed max and min value, so the supply is
working within spec.
So, that is a basic measurement, and how to interpret it. On my meter,
I have one set of knob settings, that select 1000V, 200V, 20V, 2V etc,
and a separate push button selects between VAC and VDC. Some meters have
separate knob areas for the volts DC and volts AC. So selecting VDC, which
is what you want for checking the outputs of the PSU, may vary from
meter to meter. And with a cheap, "super autoranging" meter, you
may have very few knob options indeed on the meter. It might just
say VDC, and do all the work of measuring the voltage for you. The
meter in that case, keeps adjusting the display, and the decimal point,
until the meter makes good use of all available digits on the display.
Note that the meter has limited accuracy. For example, my meter has
a "3 1/2 digit" display, which reads from 0 to 2000. But the spec for
the meter, notes that it is 1.5% accurate. That means, when I measured
the "12.23" value above, the _true_ value could be anywhere from 12.05
to 12.41 volts. Which means the least significant digit is pretty useless
on my cheap meter. So when making the measurements, you still have to
keep the measurement qualities of the instrument, in perspective. The
instrument is probably not going to be quite as bad as the 1.5% number,
but it could be.
To access motherboard voltages, the 20 or 24 pin shell on the
connector, is open at the back. You should be able to touch the
metal contained within each nylon shell hole, with your red probe tip.
That is how you can measure all the rails, while the fans are spinning
on your rig. Naturally, it is easier to work on a PC, if you
disassemble it and put the guts on your work table. That makes it
easier to see what you're doing, and makes it easier to probe in
just the right place with your new multimeter.
*******
The main danger inside the PC, would be shorting something. If I made
an accidental connection between +5V and GND for example, the wire
I used to do it, would got hot, so hot that it might burn me. The
circuitry could get damaged. But that is about the extent of the danger
to a human - short of taking a bath, and dropping the entire PC
in the bath with you, the risk of electrocution is minimal, since
you aren't working with the AC voltages from the wall plug. (You'd
have to do something extremely stupid, like jam a screwdriver tip through
the metal housing of the PSU casing, to get a shock from the AC. So
don't do that
)
Anyway, have fun,
Paul
