Archived from groups: alt.comp.periphs.mainboard.asus (
More info?)
In article <7cfbf61b.0409051144.3bb3d427@posting.google.com>,
rdoc2@comcast.net (RDOC) wrote:
> nospam@needed.com (Paul) wrote in message
news:<nospam-0509040025450001@192.168.1.177>...
> > In article <7cfbf61b.0409041944.29b2db32@posting.google.com>,
> > rdoc2@comcast.net (RDOC) wrote:
> >
> > > My OS is Windows 2000 Pro and my mother board is an Asus P4B266. I
> > > purchased this motherboard in March 2004. I installed it and
> > > everything was working fine. Out of nowhere my LCD lights on the front
> > > of the case stopped lighting up all together. The lights are for the
> > > power and the hard drive (a red and a green light). I checked the
> > > connection of the lights to the motherboard and pulled the connectors
> > > off and put back them right back on. I did this one at a time so as to
> > > be sure that I didn't mix them up. That didn't help the problem. I
> > > contacted Asus by filling out their support form and they won't even
> > > answer me at all.
> > > Can someone please help me and let me know if there is anything I can
> > > do to fix this problem or is there anything that I can check to
> > > determine what is wrong. Thanks!
> > > RDOC is online now Add to RDOC's Reputation Edit/Delete Message
> >
> > The lights are probably light emitting diodes (LED lights).
> >
> > Here is a typical driver circuit -
> >
> > +5V
> > |
> > |
> > Resistor
> > |
> > ------ (+) <--- +
> > LED
> > ------ (-) <--- -
> > |
> > |/
> > ----| Transistor switch
> > |\
> > |
> > GND
> >
> > There are two ways the LED can be connected. If the LED leads are
> > rotated 180 degrees, the LED will not light. If the LED is rotated
> > 180 degrees again and connected, this time it will light. Of course
> > the transistor switch has to pull the (-) on the LED to ground, to
> > allow current to flow. In the case of the power LED, the circuit
> > is energized virtually all the time, so reversing the LED on the power
> > terminals is a quick test of this theory. The hard drive, on the
> > other hand, only blinks the transistor switch, when the hard drive
> > is being used.
> >
> > How can the circuit fail ? Well, the LED itself could be dead.
> > Applying a large reverse bias to it, could cause it to fail.
> > Diodes sometimes fail shorted, so using an ohmmeter, you can
> > measure the LED both possible ways, and if the readings are the
> > same with the meter leads reversed, then the LED itself could be
> > dead. Normally, the forward and reverse characteristics of the
> > LED will be quite different on an ohmmeter. The LED can also be
> > blown, if it is connected straight to a power supply (like
> > straight to +5V and GND) without a current limiting resistor.
> > With enough current, the bond wires burn, and the LED can fail
> > open circuit. The resistor needed is included on the motherboard,
> > in the circuit diagram above.
> >
> > You can also damage the transistor circuit shown above, but it
> > would take "malice of forethought" to find a way to do that with
> > just the case wiring. If you aren't fooling around with it, the
> > motherboard side of things is pretty bullet proof.
> >
> > If one of the power supply outputs has died, it will be pretty
> > obvious, because the computer won't be booting, and those two
> > LEDs might not light because there is nothing to drive the
> > transistor switches. If the computer is behaving normally except
> > for the LEDs, rotating the LED leads is the first thing I'd
> > check.
> >
> > HTH,
> > Paul
>
> The green light on the motherboard is on when the computer is on. I
> haven't looked at it when the computer is off. You say check the power
> supply connection to the motherboard. Ok that means? Where does the
> power supply connect to the motherboard? I may have knocked something
> lose when working inside the case? Paul how can I check that there is
> power going to the leds. I have a multimeter and a bunch of normal
> test lights and one that has two double a batteries in it that is used
> to check if a wire is good by putting one connection on each end of
> the wire and the light will light if the wire is good?
First of all, you have to get your priorities straight. Is the
rest of the computer working normally ? If the answer is no,
then there is no point in fixing the LEDs. Generally, whatever
makes the whole computer work, will give working LEDs. It is
quite unlikely that just a LED driver or two is failing here.
If the green LED on the motherboard is lit, that means +5VSB (standby
supply) is coming through the 20 pin ATX cable, from the PSU.
That means the 20 pin ATX power cable is connected. Push the
connector down on the motherboard, to make sure the connector
is fully engaged. Do the same for the 2x2 ATX 12V connector,
which is near the three audio jacks (and/or the drive power
cable that may be connected right next to the ATX 20 pin power
connector).
As for what can you do with a multimeter, to explain that, would be
a small course on electricity plus a small course on using a
multimeter
This would take me a great deal of typing, and in
the end, I'm not sure it will help you that much with your problem.
Starting with the multimeter, I virtually never use the current
measuring scale. The reason being, it is easy to short the circuit
under test when using the meter to measure current, and the meter can be
damaged. In years and years of using meters, I have never damaged
a multimeter, and that is one of the success ingredients - the current
measurement function is only to be used under controlled circumstances,
like if you are using a power supply with a known or adjustable
current limit. There is nothing inside a PC that meets these
requirements, and that leaves us with the voltage scales and the
ohmmeter.
When it comes to the resistance scales, resistance measurement is fine
if there are parts that can be accesses without interference from
other components. For example, if two resistors are in parallel, you
can only measure the combined resistance of the two, so you don't know
the value for each resistor individually. Again, this implies that
resistance measurement has limited value, when working with a whole
motherboard (the ohmmeter works fine for sorting a bin full of unknown
resistor values).
Now, we are down to volts. Let us look at our circuit again. The
voltmeter has a really high input impedance (should be 10megohms
or greater). In the case of our LED circuit, this is high enough
that we can neglect the loading effect of the meter. Not all circuits
meet that requirement. We can use the voltmeter to make two measurements,
a no load measurement and a loaded measurement. For the loaded
measurement, we keep a box of assorted leaded resistors handy, the
sort you find at Radio Shack for the outrageous price of 2 for $0.99
when they are actually worth a couple cents each.
If we use the voltmeter, to measure from the (+) pin on the PANEL
header, to the chassis of the computer, it will read 5.0 volts.
That reading doesn't tell us a lot, other than that there could be
a 5 volt or higher voltage somewhere in the circuit. (This is with
the motherboard powered.)
+5V
|
|
Resistor
|
------ (+) <----------------------
| (red)
------ (-) |
| voltmeter (reads 5.0)
|/ |
----| Transistor switch | (black)
|\ |
| chassis = GND
GND or actual GND
Next, we stick a 1K ohm resistor as a load, and measure
the voltage again. (This is with the motherboard powered.)
+5V
|
|
Resistor
|
------ (+) <------------+---------
| | (red)
------ (-) | |
| 1Kohm voltmeter
|/ | |
----| Transistor switch ---------+ (black)
|\ |
| chassis = GND
GND or actual GND
Now, the voltage is split between the "Resistor" and the 1Kohm
test resistor. If the voltage measured read 2.5V, then we would
know the "Resistor" had a value of 1Kohm as well. If the voltage
reads higher than 2.5V, the "Resistor" is less than 1Kohm.
You can certainly use the current measuring function, and strap
it right from (+) to chassis, but if the "Resistor" just happens
to be zero ohms (i.e. there isn't one), the fuse will pop on the
meter, and the motherboard could be damaged. That is why, working
with the voltmeter scale and doing load testing, helps tell what
is on the other side of the circuit.
As I said earlier, knowing this probably isn't going to get this
problem debugged any faster. The +5V in our circuit above, is
the same +5V used to power the disk drive, and the rest of the
motherboard +5V circuits, so if the motherboard is otherwise working,
it is hard to imagine a fault that will just kill the +5V
feeding the LED circuit. On the low side of the circuit, we can
check the transistor, by doing a load test using the 1K ohm
resistor, but this time feeding the resistor from +5V.
+5V
|
|
1Kohm
|
|
------ (-) <----+---- <------
| |
|/ |
----| Transistor switch Voltmeter
|\ |
| |
GND GND
If the voltmeter reads zero, then the transistor is doing a
good job of switching on the LED. If the voltmeter reads
one or two volts, that isn't quite as good. While you could
use the ohmmeter to do this test, even with the power on,
the ohmmeter has limits as to how much current it can source,
and it might only be using 1ma of current. Using resistors,
you have the freedom to put more current into the circuit.
For example, the switch should be able to handle 10ma, which
is 5V divided by 500 ohms. So, a 510 ohm resistor could be
used, along with +5V working on the motherboard, to test whether
the transistor saturates well (measuring Vce_sat). If the voltmeter
reads 5V doing this test, then the transistor isn't working at all.
The power LED needs to have a 1x3 connector on the end, with the
middle pin missing. It connects here:
20 pin PANEL
X X +5V <----- (+)
+5V X X Power LED
MLED X X PLED <----- (-)
ExtSMI# X X Keylock
Ground X X Ground
Power_sw X X
Ground X X +5V
X X Ground
Reset_sw X X Ground
Ground X X Speaker
The Drive Activity LED on the computer case, needs a 1x2 connector
on the end, and it mates with the HDDLED pins on a separate header.
HDDLED X (-) Hard Drive
X (+) Activity LED
HTH,
Paul
Facebook
Twitter
Instagram