Failures can be due to a "short circuit" or due to an "open circuit".
A short circuit overloads the switching power converter. An open
circuit prevents current flowing to the load.
Your LED symptoms are telling you something. The LED inside the adapter
end, could be monitoring the single voltage which feeds the laptop. The
LED could be connected to the 18V rail, through a current limiting
resistor in series with it. So it may be giving you direct feedback
about the output.
The adapter could fail to produce any output, due to a primary
side or secondary side circuit problem inside the "brick". That
would require a new adapter, as getting inside it, may not be
that easy.
If the cable is frayed, near the laptop end, the wires could short
temporarily. The adapter may have overcurrent protection, which
causes the LED on the adapter to go out. Overcurrent protection
means the adapter does not get hot when it gets shorted. It might
get hot if operated at its rated load. So there should be
some protection from short circuits in the thing. That helps
to reduce the possibility of a fire.
Inside the laptop, there is probably some means to prevent an external
electrical failure, from draining the battery in a spectacular way.
Perhaps a diode or equivalent, only allows current to flow from the
charger barrel connector input, rather than back out of it. But if the
center connector pin on the laptop end, broke, and touched the other
conductor, that would cause the LED to go out. That could be your
short circuit. So the short could be at the laptop connector, or
anywhere along the adapter barrel plug and cable.
If the connector inside the laptop end broke, it could also fail open
circuit. The laptop wouldn't get any current in that case, but
the adapter LED would likely stay operating. Your symptoms don't match
that case.
So that leaves a short in the cable somewhere, or some kind
of failure in the AC to DC bits inside the brick.
Typical tests a person would do on an adapter, would consist
of two tests. These tests are intended to prove the adapter
operates over its ratings range.
1) With adapter unplugged from load, verify the open circuit
voltage. The adapter is likely regulated DC, and it might
deliver a steady 18V whether the load is zero amps or 3.5 amps.
An open circuit test verifies the switching converter inside
the brick, is doing something. The LED should come on.
If the switching regulator is "weak", it might still pass
this test, and the LED might be on.
2) load test. If the adapter is rated 18V at up to 3.5A, then
a person could devise a load test. This tests that the
adapter meets its stated rating. Doing the math, 18V / 3.5A = 5.14
ohms.
Nine of the following resistors in parallel, would give a 5.55 ohm
load, with a 90 watt dissipation rating (point a fan at the
resistors, to cool them off). There should be some clearance
around each resistor for air flow. The slightly higher resistance
value I get by using just nine resistors, means the load test
will be testing the adapter at slightly less than 100% of its
rating. I don't want the final load resistance value to be
lower than 5.14 ohms.
http://www.radioshack.com/product/index.jsp?productId=2062292
The load test verifies the brick still delivers the rated power.
If the LED goes out during the load test, then the adapter
should be replaced. It could very well be a problem with the
switching regulator inside the brick. During the load test,
you can also verify the regulated output voltage is still
at 18V or whatever.
Different brands of adapters, will operate at different
voltage and current values. There are 65W adapters and 90W
adapters. There can be subtle differences between them. You
can adjust the arithmetic above to suit.
The only thing I don't like about that load test above,
is the resistors are wire wound. That means there could be
a slight inductive kick when the load is applied or
disconnected. I would leave the load connected before
plugging and unplugging the adapter, during the load
test, to reduce the insult to the adapter. The adapter
may have "slow start" output, which would reduce
inductive kickback. Once the adapter is disconnected
from the wall, then I'd disconnect my home made load.
I use resistors like that, for a home made ATX power supply
tester. I point an 80mm fan at my resistors, during the
test. I load test new ATX power supplies, before I use
them, but not at their rated power. (I run for a couple
hours, to verify there aren't any infant mortality problems
with my new product. I check the voltages during the test.)
A laptop is more likely to use all the watts that the
adapter can provide, so getting closer to the rating makes
more sense for those. (You don't need to run the load test
for two hours on the laptop adapter - even 30 seconds is
enough to prove it still runs at close to full load.) Since
an ATX supply has multiple rails, it is more difficult
(and expensive) to devise meaningful load tests. I selected
a total loading of about 100W or so, for my tester, suitable
for most power supplies I might buy. It isn't intended to
be stressful to the supply and wasn't part of my plan.
It allows me to check the voltages before actual usage,
without danger to a new motherboard.
(An example of a home ATX power supply tester.)
http://groups.google.ca/group/alt.comp.periphs.mainboard.asus/msg/19647caf2c65504b?dmode=source
To make a reliable connection to a barrel power plug, you
could use something like this. Otherwise, it could be
difficult to hold a load in place without the contact
being intermittent.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CP-048H-ND
HTH,
Paul
