Linear or SMPS?

[Jeff Layman]

I was thinking of changing a wallwart 12v 1000mA PS from linear to SM, as
the latter is more efficient, but now I'm not so sure.

What is the no-load power consumption of each type? I understand from
Wikipedia that a SWPS must operate with a minimum current, so needs a dummy
load when it is not being used.
(http://en.wikipedia.org/wiki/Power_supply#Switched-mode_power_supply). In
that case, depending on the full and standby consumption of the device being
powered, the SMPS may actually cost more to run if the SMPS "standby"
consumption is higher than the equivalent linear supply. It will, of
course, depend on the relative full/standby operation times, but I assume
there is a crossover point where one type of power supply will cost less to
run in the long term than the other.

Also, I have generally assumed that SMPS efficiency is in the region of
70%+, whereas linear efficiency is only 30 - 40%. However, this 1250mA
multi-voltage SMPS has an efficiency of only 51%
(http://www.maplin.co.uk/Module.aspx?ModuleNo=96864). To be fair, the
higher-power version (2500mA) has an efficiency of 75%.

Does that mean there is little point in going to an SMPS if the power
consumption if on the low side? Or will it always win out in terms of cost
of operation?

 

[Charles]

I was thinking of changing a wallwart 12v 1000mA PS from linear to SM, as
the latter is more efficient, but now I'm not so sure.

What is the no-load power consumption of each type? I understand from
Wikipedia that a SWPS must operate with a minimum current, so needs a
dummy load when it is not being used.
(http://en.wikipedia.org/wiki/Power_supply#Switched-mode_power_supply).
In that case, depending on the full and standby consumption of the device
being powered, the SMPS may actually cost more to run if the SMPS
"standby" consumption is higher than the equivalent linear supply. It
will, of course, depend on the relative full/standby operation times, but
I assume there is a crossover point where one type of power supply will
cost less to run in the long term than the other.

Also, I have generally assumed that SMPS efficiency is in the region of
70%+, whereas linear efficiency is only 30 - 40%. However, this 1250mA
multi-voltage SMPS has an efficiency of only 51%
(http://www.maplin.co.uk/Module.aspx?ModuleNo=96864). To be fair, the
higher-power version (2500mA) has an efficiency of 75%.

Does that mean there is little point in going to an SMPS if the power
consumption if on the low side? Or will it always win out in terms of
cost of operation?

30 days x 24 hours x $.15 per kWh x 12 W = $1.30 per month (at full load,
assuming 50% efficiency, that's $2.60 per month). Is that important? With
20 of these, I suppose it could be worth thinking about.

I don't think wallwart SMPS do anything crazy if not loaded ... would not
worry about it. I guess that the no-load energy use is similar for both
types.

It appears SMPS are becoming more and more popular, in any case. Silicon is
lighter and cheaper than iron and copper.

 

[Jeff Layman]

30 days x 24 hours x $.15 per kWh x 12 W = $1.30 per month (at full load,
assuming 50% efficiency, that's $2.60 per month). Is that important? With
20 of these, I suppose it could be worth thinking about.

Here in the UK electricity is about $.24 per kWh. As to whether or not it is
important, even at the US cost, 3 or 4 would save over $100 a year. Would
you turn down a $100 bill if one was offered to you? That's the trouble
with continuous small costs, we never bother to work out what the total is
over a long period of time!

I don't think wallwart SMPS do anything crazy if not loaded ... would not
worry about it. I guess that the no-load energy use is similar for both
types.

It appears SMPS are becoming more and more popular, in any case. Silicon
is lighter and cheaper than iron and copper.

Now that is an interesting observation, considering the cost of metals these
days.