# How do laptops get their power?

#### Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

•  Subject
• Author
• Posted on
Seems like this ought to be a faq, I'd like to know where to look.

The question is, when you plug a laptop into the wall, does the
battery serve as a power conditioner for the laptop?  I may have to
use an old generator to power my laptop, and want to know if I will be
better off never running it while it is plugged in to the generator,
and rather only use the generator to charge the battery and then run
the laptop on battery only.

If it matters, this is a 5 yr. old thinkpad.

## Re: How do laptops get their power?

The battery normally does not serve as a power conditioner.

Laptops have an internal switching power supply.  Both the battery and
the external AC adapter are just inputs to this switching power supply,
which also charges the battery when on AC.

Don't worry about the generator.  Both the AC adapter and the power
supply in the laptop are switching power supplies, and switching power
supplies are extremely tolerant of abberations in their own input power.

grizdog@gmail.com wrote:

## Re: How do laptops get their power?

Check to find out what kind of power the generator will be producing.  Some
generate a sine wave like the power company provides while others produce a
modified power output using some sort of square wave or other simulated sine
wave output.

If the output is a true sine wave then you will have no problem but if it is
one of the square or modified wave shapes be very careful when attaching
computers to it.  Many power supplies, computer or other electronic
equipment, do a melt down when they are plugged into a square or modified
sine wave output.

## Re: How do laptops get their power?

Why?  The first thing encountered by the mains current is a filter circuit
(which filters nasty spikes and things from either direction).  The second
thing encountered is a full wave rectifier to convert the AC to DC.
Rectifiers actually work better from square waves than sine waves (The
modified sine wave is somewhere between the two).  The next thing is a
reservior capacitor which will absorb pretty well much of any spikes that
are left.  Next is a choke followed by a filter capacitor which together
filter out any ripple left on the mains supply.  Any noise or spikes on the
incoming supply left by the reservoir capacitor (and their won't be much)
will get stopped here.  It is followed by a flyback switch mode converter
whose very principle of operation relies on the generation of voltage spikes
and so if there are any left, it just won't care.

The reality is that switch mode power converters found as laptop power
supplies are pretty unfussy devices digesting pretty well anything that you
throw at them, including much that is outside of the ratings shown on the
tally plate.  They will even work perfectly well off a DC mains supply of
150 to 300 volts.

## Re: How do laptops get their power?

M.I.5¾ wrote:

The current in the filter cap
Ipk = c x dv/dt.  The dv/dt of a sine wave is pretty low near the peak
where the diodes turn on.  The square wave/modified sine wave has a very
high dv/dt where the diodes turn on.  Higher peak diode current, higher
peak cap current, higher fuse current...all increase the failure rate of
the input components.  This is mitigated somewhat by the resistance in
the input filter.

So, noise gets filtered, yes.  Failure rate goes up, yes.
Maybe not enough to matter, or maybe enough to matter.  Depends on the
overdesign of the input circuit.  Most of the time it will work just
fine...except when it doesn't.

(The

--

## Re: How do laptops get their power?

Can you smell something? Gas? Drains?  No, It's Bullshit!

You are talking out of your arse hole.

First the impressive looking differential formula that you give is
completely invalid (and it certainly didn't even look right).  If you
perform a dimensional analysis on the c x dv/dt part, it gives a result in
coulomb seconds (dimensions [IT^2] - which does not correspond to any
recognised unit) and *not* amps (dimensions [I]).  In any event peak current
in diodes is also largely completely irrelevant.  It is the average diode
current that matters, failure rate being determined largely by power
dissipation (i.e. heat).  The only other thing you have to worry about
beyond that is that the I^2t rating for any current pulses does not exceed
the I^2t rating of the diode (essentially its ability to emulate a fuse).

For an ideal square wave the output voltage of an unloaded full wave
rectifier is contstant, so you arguement fails completely as there are no
perceptible current pulses anyway.   The only current flowing through the
diodes is the load current (on each half cycle).  The current flowing
through the reservoir capacitor should be zero.  In any event the dv/dt in
your flawed formuala would be infinity for an ideal square wave and *very*
large for a practical one.

## Re: How do laptops get their power?

M.I.5¾ wrote:

In any event the dv/dt in

THAT IS EXACTLY MY POINT!!!
It's true that for a perfect square wave, the infinite current only
happens at turn on.  But for a perfect (typical shape) modified
sine wave, it happens twice every cycle.

LOL
You never cease to amaze me.  You should write comedy.

So, you're saying that you slept thru math AND circuit analysis class...
time-variant voltage and current flow in a capacitor.

The depth of your "understanding" is...well...bottomless.
The concern I have is that someone may take your input as relevant
and get hurt.
These news articles hang around forever.

So, keep up the good work entertaining us.  I'm sure there's no stopping
you anyway.
strange or misleading,...well...I just don't have enough time for that.

I can't stop laughing....
mike

--

## Re: How do laptops get their power?

Thanks, folks.  I'm very glad I asked - this has been very
educational.

Now I'm going to go read up on switched power supplies.

Thanks again.

## Re: How do laptops get their power?

You formula is invalid so your point doesn't exist.

Where is this infinite current going to go?  Certainly not through the
reservoir capacitor because the difference between the diode output and the
capacitor voltage is zero, so no current can flow.  There is no current
pulse.

My dear fellow, I taught circuit analysis for many years.

What time variant voltage?  Full wave rectification of a square wave input
gives a constant steady voltage output.  Therefore there is no current flow.
Plot it out on a piece of graph paper and see for yourself.

Well they would be better off than following your complete and utter
bollocks based on made up fomulae where the left and right hand side don't
even match.

and I will post where your response is flawed which I have no doubt it wll
be.

## Re: How do laptops get their power?

I wondered if the above formula had been misquoted from something else, but
my searching through reference books has turned up nothing that looks like
it.  It did have a look through a work on inverter design which noted that a
square wave output is the ideal wave shape for subsequent full wave
rectification as it required the minimum of filtering.  However, it didn't
mention anything about large current pulses through the rectifier diodes or
indeed any current pulses at all (not unsurprisingly because you made them
and the formual up).  It just noted that IrmsAC = Idc (for a square wave -
this is not true for any other wave shape). A statement that would not be
true if their were current pulses.  None of this came as any surprise to me.

I showed you post to some colleagues (were are all engineers here).  All of
them spotted that your formula and post couldn't possibly represent reality.

Sorry sunshine, but you're completely wrong.

## Re: How do laptops get their power?

M.I.5¾ wrote:

Here's my take on this - and I'm a chartered engineer, with a first
degree in electrical and electronic engineering and a PhD in the
engineering area.

The current in a capacitor *is* C dv/dt. Where dv/dt is the "rate of
chance of voltage with time". If your engineer friends are not aware of
that, they are certainly not competent electrical engineers. You will
find that formula on Wikipedia and various other places. It is very
basic electronics.

You state "It just noted that IrmsAC = Idc (for a square wave - this is
not true for any other wave shape)" I'm not quite sure what you mean by
this. Since a square wave is not DC, it is hard to see exactly how you
can define Idc.

Also, anything with the statement "this is not true for any other wave
shape" is almost certainly wrong. Unless you have a proof that there is
only one single waveform that can have a given property, I would tend to
not say it. It's quite likely there will be another waveform with the
properties you refer too, even if that waveform is of no practical
significance.

Electronic items tend to fail more on switch-on / switch-off due to
transients. I can see the transients are higher for a square wave. BUT
once running, I would think the SMPS has an easier time with a square
wave drive than it does with a sine wave. The capacitor current would be
fairly low in this case, as the capacitor will be being charged almost
all the time.

On balance I would not like to say what is best. I would certainly
myself not worry about it. I'd just use it.

Well, I beg to differ. He is not completely wrong at all.

## Re: How do laptops get their power?

His reasoning is vague or ambiguous and his conclusion is clearly
wrong.  I'm ignoring that analysis until he can explain things such as
which dv/dt is low at what point, how that is relevant, and other
conclusions that can only be made using other facts not in evidence.

Is strain on a power supply greater with square wave inputs?  Yes.
But if increased strain is so destructive, then also do not drive a
car up hills.  That also increases strain and causes premature car
failure.  IOW his conclusions are completely irrelevant once we put it
into perspective.  The user will see no difference between typically
cleaner electricity from the utility or from serious UPSes compared to
significantly 'dirtier' electricity output by the typical plug-in UPS.

This post to put into perspective for grizdog the latest posts.
Grizdog, if you think the latest posts are confusing, well, so do I.
And I am suppose to know this stuff.

Bottom line - the switch mode power supply (not a battery) makes
those electrical anomalies irrelevant.  M.I.5=BE demonstrated why in a
short summary of how the supply works.

## Re: How do laptops get their power?

That's correct - but note that the voltage being differentiated is the
voltage *on the capacitor*, not the input voltage to the supply.  The
person who first used this formula was talking about the dV/dt of the
voltage *out of the SPS* and thus at the input to the supply.  But when
any current is flowing in to the capacitor, the capacitor dV/dt is going
to be lower than the supply input dV/dt due to any resistance or
inductance in the input filter and the diode bridge itself.

A fast-rising input waveform allows the input voltage to temporarily
diverge further from the capacitor voltage than it would with a pure
sine wave, so the peak current will be higher.  But as long as it's
within the rectifier's peak current rating, it should be OK.

Dave

## Re: How do laptops get their power?

Hmm, Just run the dimensional analysis again.  I have this wonderful all
singing all dancing calculator that does such things.  I have just
discovered that the firmware has a bug or a misfeature or something.  If you
don't put the 't' in the above equation in brackets it multiplies instead of
divides.  In any event the point is entirely accademic becauae the capacitor
does not see the square wave input being, as it is, on the other side of the
rectifier.  Hence the equation does not turn up in power supply reference
material in this regard(which probably also why no one recognised it -
probably like me, maybe encountered it once at uni but never used it since
and forgotten about it.).  The voltage on the capacitor at the end of one
half cycle is exactly the same as the voltage at the beginning of the next,
so if you do apply the equation you still get a result of zero and not
infinity as the poster claimed.

The IrmsAC *before* the recitifier is equal to the Idc *after* the
rectifier.

Indeed for sine wave inputs to the recitifier the ratio of IrmsAC to Idc can
vary quite widely depending on the exact configuration of filtering
following the rectifier.  For a capacitive and resistive filter, IrmsAC =
1.56xIdc.  Change the resistor for a choke and the ratio is lower.  But they
are never equal.

He was completely wrong in that the capacitor does not take infinite current
pulses as he claimed.

## Re: How do laptops get their power?

snip

It's the FIRST formula you learn in AC circuit analysis class.
THE VERY FIRST!!!!  You can't do AC analysis without it!  It's the
AC DEFINITION of a capacitor.

It did have a look through a work on inverter design which noted that a

I've said it three times now, you're right in the degenerate case of a
pure square wave.  Problem is that most modern DC/AC converters use
(what they call) a modified sine wave.  Why?  Because you need the peak
voltage to
be the same as a sine wave for electronic stuff to work.  And you need
the RMS value of the waveform to be the same as a sine wave so light
bulbs work.  The result of that approximation is alternating positive
and negative pulses whose width and amplitude are adjusted to meet the
above criteria.  There's dead time between the two.  That results in
"ripple" on the output and the result is a current pulse into the filter
cap twice/cycle.  Fast edges cause more peak diode current than sine waves.
Want a picture?
Sure.
http://mike.liveline.de/pictures.html
I measured the AC input current because I was too lazy to take the
It's a Toshiba PA2450U Laptop adapter.
Tektronix TDS540 Scope
Tektronix P6042 Current Probe
Sinewave came from the mains.
Modified sine wave from a noname auto DC/AC converter.
The sharp peaked trace is from the DC/AC converter.
The smoother one from the mains sine wave.
Some of the peak is from the supply input filter.  Extrapolating
the exponential back to the transient should give a good
approximation to what's happening to the diode current.

Quality diodes are incredibly tolerant of peak currents.  Typically,
this won't be a problem.  But there are marginal designs out there.
Judging from the response of your colleagues, a LOT of marginal designs.

It would be interesting to have some inside information on how thoroughly
laptop power supplies are tested with non-sinewave inputs.

The early Tektronix 200 series mini-scopes used a charge dump
battery charger that was basically a capacitor and some diodes
connected from the input mains to the battery.  Worked great on sine
wave power.
I plugged one into a modified sinewave inverter.  The charge circuit
went up in smoke in a flash...literally.  Wish I'd looked at the schematic
before I did it.

However, it didn't

I'm not at all surprised.  There are a lot of "engineers" out there who
lack the basic understanding of electronics fundamentals.  Hiring an
engineeer was always a very depressing experience for me.

Thanks for the compliment.  Sunshine is the exact opposite of having
one's head in a dark place.

Are we having fun yet?
mike

## Re: How do laptops get their power?

See my response to Dave.

Virtually everybody else has stated that these supplies run from square
wave, modified square wave or anything else without any problem, so there is
little point in dicussing your fundamental error any further.

## Re: How do laptops get their power?

Apparently you have not looked at outputs from most computer grade
(plug-in) UPSes when in battery backup mode.  For example this one
outputs two 200 volt square waves with a spike of up to 270 volts
between those square waves.  This is perfectly good for electronics
but may harm some small electric motors.  Output from UPSes may be so
'dirty' as to harm power strip surge protectors but is perfectly
sufficient for electronics.

Computers are some of the most robust appliances in the house which
is why a computer grade UPS typically delivers creates some of the
'dirtiest' electricity in the house.

Barry Watzman has accurately described how a laptop works.  Battery
requires computer control circuitry to charge and operate.  Therefore
battery needs much cleaner power from a power supply.  Clean power is
created by power supplies in all electronics - to make dirtiest
electricity irrelevant - so that battery can even be recharged.  Some
'assume' a battery operates as if a filter.  It cannot because the
typical NMHd or lithium battery needs cleaner power for its battery
controller.  Electronics is powered either directly from AC mains or
from battery.  AC power does not arrive via (get filtered by) battery
circuits.  Power supply must 'clean' electricity so well as to make a
computer one of the most robust household appliances.

## Re: How do laptops get their power?

As far as I know, all "generators" generate pure sine wave.  INVERTERS
often generate "modified sine wave" (and having looked at some APC UPS'
on a scope, I think that such a description is a joke).  But he said
"generator", which implies a mechanical rotating device.

Regardless, however, all switching power supplies simply take the
incoming power (whatever it is), rectify it, brute force capacity filter
it, and use it as DC input to the electronic switch.  The nature of how
they work makes them incredibly tolerant of almost unbelievable
abberations in their own power input.

I find your assertion that "Many ... computer or other electronic
equipment, do a melt down when they are plugged into a square or
modified sine wave output" to be out of touch with reality.  That simply
is not the case.  If it was the case, then everything plugged into any
of the UPS' sold at retail (including tens or hundreds of millions of
models by APC) would have, in your terms, "melted down", and that simply
has not happened.

GlowingBlueMist wrote:

## Re: How do laptops get their power?

The output of many inverters is better described as "modified square
wave" instead of "modified sine wave".

Once upon a time, there were true square wave inverters.  The output
voltage transitioned directly from +120 V to -120 V (for example),
spending essentially no time at zero voltage.  This waveform is easy to
generate with almost no electronics, but it has a bunch of nasty
properties.  If you make the peak voltage 120 V, you get 120 V RMS out,
which is good for light bulbs and other heating devices.  But normal 120 V
sine-wave AC has a peak voltage of 170 V, and many electronic devices
draw power from the line only near peak voltage.  Such devices, designed
for a peak of 170 V, may not work at all on a 120 V square wave.  And if
you increase the peak voltage of the square wave to make electronics
happier, RMS-dependent loads like lamps see too much voltage.

Also, a pure square wave has a high amplitude of third and fifth
harmonic content (180 and 300 Hz).  Some devices with iron cores like
transformers and motors may overheat when subjected to these harmonics.

That's where the "modified square wave" comes in.  The inverter logic is
made slightly more complicated to give a waveform that spends some time
at zero volts in between the high and low voltage peaks.  It's still
square-looking, and still efficient to generate.  But by keeping the
voltage non-zero for only a portion of the time, the peak voltage can be
raised while keeping the RMS voltage constant.  For example, if you make
the peak voltage 170 V but have the waveform at zero voltage 50% of the
time, you get 170 V peak and 120 V RMS.  Both electronic and resistive
loads are pretty happy with this.

In practice, it seems that many inverters use a slightly lower peak
25-35%).  I've never seen a good explanation for this, but I have a
guess: this reduces the amplitude of the third harmonic in the output
compared to the 170 V/50% case.  There's a particular combination of
voltage and time somewhere in this range that cancels the third
harmonic entirely, and many of the inverters I've looked at on a scope
operate somewhere near this.  So this is a pretty good compromise
waveform: 120 VRMS for the benefit of resistive loads, 140-160 V peak to
keep electronics happy, and low third harmonic for the benefit of
transformers and motors.

But it's still awfully square-looking, and still has higher harmonics
even if the third harmonic is missing.

To be correct, you should note that some standby power supplies
(including some APC models) have "true sine wave" output.  This is
actually a PWM approximation to a sine wave, but it's reasonably
sine-shaped.  These don't seem to be nearly as common as the "modified
sine wave" ones, though.

Dave

## Re: How do laptops get their power?

On Sep 4, 9:36 pm, da...@cs.ubc.ca (Dave Martindale) wrote:

Bottom line to what both Barry and Dave have posted - even
'dirtiest' output from a computer grade UPS (electricity so dirty as
to even damage some small electric motors) is irrelevant to computers.
Even generators would not create electricity that dirty.  Computer's
power supply must make those harmonic and noise problems completely
irrelevant.

One generator problem can cause computer damage.  Some generators
are so inferior as to not adapt well to changing load.  If generator
is also powering a circular saw or a freezer, and if that load
suddenly disconnects, then a generator high voltage may then cause
electronics damage.  Some generators are not very good at voltage
regulation.

Well, laptops are supposed to work on any voltage from 90 to 265
volts.  If operating at 120 volts when a generator suddenly 'over
voltages', well, that voltage should never exceed 265 volts.  Laptop
should not suffer damage.  Still, this is why many spend a little more
money for a generator that has superior regulation and also tends to
be extremely fuel thrifty when power demand is low.  Generators such
as Hondas should be considered so that when a large load disconnects,
then generator does not spike out a large voltage.

Multiple posters are disputing claims from GlowingBlueMist who has
inaccurately assumed a UPS outputs 'clean' power - a popular myth.
Also provided are multiple reasons why 'dirty' electricity is not
harmful to computers (including laptops).