A power supply is one of the
most important components in
a computer, yet it is often
the least appreciated due to
it’s "low-tech nature". When
a power supply is dead, your
entire system is dead. A
faulty power supply could
also cause other parts of
your system to fail. As
personal computers become
ever more powerful, the
importance of a reliable
power supply is greater than
ever before.
It’s easy to tell if a power
supply is working or not.
However, until now it has
been hard for a PC end-user
to measure the quality and
reliability of a power
supply. Herein we attempt to
document a very simple way
to estimate the quality of a
power supply by the weight.
We measured the weights of
38 standard ATX power
supplies of different
wattages, models and
manufacturers. The weight
data were then correlated
with the current price and
wattage on the label. The
findings are in agreement
with the return rates and
general experience we
accumulated from many years
of selling and servicing
power supplies.
Equipment
The weight is measured by a
digital scale with the
accuracy of +/- 0.01 lb.
Results
Table 1 shows the
models, wattage, net
weights, and current prices
(on 12-20-00) of 38 power
supplies from 23
manufacturers. The weights
were measured with a digital
scale. The error margin was
within +/- 0.1lb. The market
prices of the products (as
they were, not changed or
adjusted for this study)
were plotted as a function
of the weight in Figure 1.
The wattage of the power
supplies was plotted as a
function of the weight in
Figure 2. The wattage data
were what were shown on the
power supply stickers.
Observations
Several observations can be
made from the data:
- There is a large
variation in the weight
among the same wattage.
For instance, among all
the 300W power supplies,
the weight ranged from
2.15 to 3.8lb, a 1.65lb
difference. The lightest
one is only 57% of the
heaviest in weight.
Among the 250W, the
weight ranged from 1.7
to 3.1lb, a 1.4lb gap.
The heaviest one is 82%
heavier than the
lightest one. Judging by
these data, some of the
good-quality 250W power
supplies are clearly
better than the
poor-quality 300W’s.
Therefore, don’t be
fooled by the wattage on
the sticker.
- Generally, the
Athlon-approved power
supplies weigh more than
the non-Athlon approved
of the same wattage,
confirming that Athlon-approved
ones are of better
quality. It's possible
that manufacturers only
submit better models for
the AMD approval
process.
- The name-brand power
supplies typically weigh
more than the
less-known, generic
brands of the same
wattage.
- FCC approved power
supplies often weigh
more than those without
FCC labeled.
- Figure 1 clearly
shows that the market
price of the power
supply is, to a degree,
proportional to the
weight, especially for
items heavier than
2.25lb. The higher the
weight, the higher the
price tends to be. You
pay for the pound! For
power supplies weighing
lower than 2.25lb (the
blue dots), the
correlation does not
work well, suggesting
either the existence of
a minimum market value
or some of these power
supplies should be
priced even lower (due
to low quality).
- Figure 2 shows that
the wattage of a
good-quality power
supply is, to a degree,
proportional to the
weight for the
good-quality power
supplies such as Athlon
approved 250W's and
300W's. This diagram
also shows that some of
the low-weight power
supplies (blue dots)
most likely do not have
the true wattage as
labeled.
Why Weight Matters?
The more appropriate
question is why size
matters.
The weight of a power supply
is directly related to the
quantity, quality, and size
of the material (thus cost)
used to build the power
supply. We opened a light
and a heavy 300W-labeled
power supplies. The heavy
one has larger capacitors,
thicker wires, larger
transformer, larger heat
sinks, more connectors, and
more capacitors than the
light one, all of which are
important factors for the
overall cost and quality of
a power supply. The
difference is illustrated
below:

Conclusions
This brief study clearly
confirms an empirical
knowledge: the quality of a
power supply can be estimatd
by its weight. While this is
not a true scientific or
thorough measurement of the
power supply reliability, it
is nevertheless a very
simple and easy way for
ordinary PC users to
estimate and compare the
quality of a power supply.
Directron.com has pioneered
this measurement and has now
published the weights of all
power supplies sold on
Directron.com’s web site.
It’s important for the
end-users to know the
quality of a power supply.
The specs of power supplies
are simply printed on
stickers. Don’t be fooled by
them! The stickers can be
easily removed and replaced.
Even the manufacturers might
place a 300W sticker on an
otherwise 250W power supply.
Once products are shipped it
is quiet rare for any
private or public entity to
go back and verify the specs
on the sticker. However,
it’s difficult, if not
impossible, for a dishonest
vendor to alter the weight
of a power supply.
By no means is the weight
the best way to measure the
quality of a power supply.
There are other factors one
should consider in
purchasing a power supply
such as wattage, output,
brand, price, noise level,
and long-term stability.
However, to most end-users,
measuring the weight may be
the only viable method
that's widely available.
The large variation in the
quality of power supplies
with identical wattage
rating suggests that an
industry agency is needed to
rate, confirm, and police
the labeling of power
supplies to protect the
consumers.
Table 2. A guideline to
classify the quality of a
power supply by its weight
(lb).
| |
Low Quality |
Average Quality |
High Quality |
|
250W |
1.7-2.4 |
2.4-2.8 |
2.8-3.1 |
|
300W |
2.1-2.7 |
2.7-3.2 |
3.3-3.8 |
Based on the above criteria
and our general experience
in return rates, these are
consistently high quality
brands on the market:
EnerMax, Delta, PC Power &
Cooling, Enhance, Antec,
CWT, and SPI. These brands
are consistently associated
with low-eight models: L&C,
Deer, Skyhawk, Star, KME,
KingStar, and LCT. Our
experience shows that the
return rates of a low-weight
power supply could be as
high as twice of those of
high-quality ones. How does
your power supply measure
up?
Physical Dimensions
Besides the specs and form
factors, the physical
dimensions are also
important factors in
selecting a compatible power
supply. Here is an outline
of the physical dimensions
of most standard power
supplies:
ATX: 6x3.5x5.5", HxWxD.
Most common. Uses 4 mounting
screws.
Mini-ATX: 5x3.5x5",
HxWxD. Rare size. Uses 4
mounting screws. Can be used
in a regular ATX case, but
often not the other way
around.
MicroATX: 5x3x4", HxWxD.
Use 3 mounting screws. Not
interchangeable with ATX or
miniATX.
Flex ATX: Even smaller
than Micro ATX. Various
sizes according to case
specs; often not
interchangeable.
Use the data above to
determine if a particular
power supply would fit your
case.
Notes: (1) Most case
manufacturers do not make
their own power supplies.
They often use power
supplies that are re-labeled
as their brands. For
instance, PowerMan brand
from In-Win is made by SPI
(same as FPS). One good way
to identify the true
manufacturer of a power
supply is to look at the
circuit board on which the
manufacture name is often
printed. (2) Power supplies
are also heavily
cross-branded. For instance,
Enhance power supplies are
branded as FKI and HP. (3)
The largest power supply
manufacturer is Delta since
Delta makes power supplies
for those large OEMs;
however, Delta does not make
power supplies for the
distribution channels. (4)
It’s possible that some
power supply manufacturers
produce two grades of the
power supply for different
market segments. An example
is the CWT-300ATX (FCC
approved) and ATX-300
(non-FCC), which are 0.6lb
apart. (5) Other factors may
also contribute to the
weight differences, for
instance, the number of
power connectors and fans
may vary from one to the
other. However, such factors
typically are in agreement
with the trend. The heavier
the better - assuming other
conditions are the same. (6)
It'd be nice to do a further
study to correlate the
weight of power supplies
with their output
performances. (7) The
purpose of this article and
study is not to promote or
demote any particular brand
or model, but to provide a
subjective and factual
results. (8) One may ask why
some manufacturers choose to
cut the corners to produce
below-average power
supplies. The main driving
force is the demand for
cost-reduction of a complete
system in the reselling
market. Nowadays, one can
purchase an ATX case with
power supply for less than
$25, which can significantly
reduce the price of a system
for those consumers who may
have never pronounced the
phrase "power supply." (9)
Most resellers who sell
systems would not quote the
brand of the power supply
included in the system. If
you buy computers instead of
build your own, ask for the
brand of the power supply.
How Much Power Do You
Need?
|
Component |
Requirement
|
Line(s) Used |
|
AGP Video Card |
30 – 50W |
+3.3V |
|
Average PCI Card |
5 – 10W |
+5V |
|
10/100 NIC |
4W |
+3.3V |
|
SCSI Controller
PCI Card |
20W |
+3.3V and +5V |
|
Floppy Drive |
5W |
+5V |
|
CD-ROM |
10 – 25W |
+5V and +12V |
|
DVD-ROM |
10 – 25W |
+5V and +12V |
|
CD-RW |
10 – 25W
|
+5V and +12V |
|
7200rpm IDE Hard
Drive |
5 – 20W |
+5V and +12V |
|
10,000rpm SCSI
Drive |
10 – 40W |
+5V and +12V |
|
Case/CPU Fans |
3W
(ea.) |
+12V |
|
Motherboard
(w/o CPU or RAM) |
25 – 40W |
+3.3V and +5V |
|
RAM |
8W per 128MB |
+3.3V |
|
Pentium III
Processor |
38W |
+5V |
|
Pentium 4
Processor |
70W |
+12V |
|
AMD Athlon
Processor |
70W |
+12V |
For overall power supply
wattage, add the requirement
for each device in your
system, then multiply by
1.8. (The multiplier takes
into account that today’s
systems draw
disproportionally on the
+12V output. Furthermore,
power supplies are more
efficient and reliable when
loaded to 30% - 70% of
maximum capacity.)
|