Actual experiences mod'ing laptop bricks?

[Don Y]

Hi,

I've got a ~200W laptop brick on which I'd like to tweek the
output voltage (~5% -- 19.5V down to 18.5V). Anyone know what
sort of margins they design into these? I'm not keen on risking
the laptop for the sake of a $10 power supply... (i.e., "it
*should* work" is not an acceptable reply :> )

Thx,
--don

 

[mike]

Hi,

I've got a ~200W laptop brick on which I'd like to tweek the
output voltage (~5% -- 19.5V down to 18.5V). Anyone know what
sort of margins they design into these? I'm not keen on risking
the laptop for the sake of a $10 power supply... (i.e., "it
*should* work" is not an acceptable reply :> )

Thx,
--don

I use laptop power supplies for all kinds of stuff.
Just change the sense resistor value.
Changing 5% shouldn't be a problem.
But the answer to your concern is in the risk/reward ratio.
If you can get the right one for $10 and your laptop
is worth much more than $10, it's a no-brainer.

You might get better help if you listed more detail
on what/why you're doing it?
At that power level, there might be negotiation between
the laptop and the brick.
I've seen some power supply interface designs that I'd
classify as stupid.

My experience with sourcing custom OEM power supplies is
that they're like everybody else...half of them
are below average. And average ain't nearly good enough.
So asking them ain't gonna be helpful either.

 

[Don Y]

Hi Mike,

I use laptop power supplies for all kinds of stuff.
Just change the sense resistor value.
Changing 5% shouldn't be a problem.

That was my plan -- but, I'm concerned with the downside
risk (derating plus any "smarts" in the power supply that
might be "confused" because they are unaware of this change)

But the answer to your concern is in the risk/reward ratio.
If you can get the right one for $10 and your laptop
is worth much more than $10, it's a no-brainer.

Not my laptop. These are laptops that were donated (without
power supplies... sheesh!), will be "refurbished" and "gifted"
to disadvantaged kids. (I.e., the sorts of folks who can least
afford to bear the consequences for a bad decision on my part!)

I can get my hands on a supply of these larger 200W supplies
(virtually) "free" but am hesitant to *use* them "as is"
(not knowing how tolerant the laptop is of variations in input
voltage *above* spec). So, it's basically just my (donated)
time to make the changes.

OTOH, if I have to purchase power supplies (even "used"), they're
in the $20+ range, each. (assuming I can find a single source
at that price!)

You might get better help if you listed more detail
on what/why you're doing it?

See above.

At that power level, there might be negotiation between
the laptop and the brick.

Exactly. And, I'm sure any "smarts" therein won't take into
account an *observation* of any component changes that have been
made "post production" :<

I know some of these power supplies refuse to work on "competitor's"
products (despite having the same V+I characteristics)

I've seen some power supply interface designs that I'd
classify as stupid.

My experience with sourcing custom OEM power supplies is
that they're like everybody else...half of them
are below average. And average ain't nearly good enough.
So asking them ain't gonna be helpful either.

This is one reason why I want to avoid the "off brand"
replacement products. So, I'm stuck with either "name brand"
*new*... or *used*.

 

[dp]

Hi,

I've got a ~200W laptop brick on which I'd like to tweek the
output voltage (~5% -- 19.5V down to 18.5V). Anyone know what
sort of margins they design into these? I'm not keen on risking
the laptop for the sake of a $10 power supply... (i.e., "it
*should* work" is not an acceptable reply :> )

Thx,
--don

Hi Don,
some years ago I tried this on a small 15V (IIRC) adaptor. It was
somewhat smaller, not an ordinary laptop one but I got it because
it stated its output was settable. It was not so I opened it hoping
to find the feedback divider and change it (it was a cheap DX box).
Turned out the designer had been pretty good and had done things
not that obvious without having the circuit. The feedback was somewhat
hard to modify, a "zener" (TL431) serially feeding current into
the feedback optocoupler - and there was more to it than that
which would have taken hours to understand so I just abandoned
the idea (got another brick).
I suppose there are many different designs out there and most of
them likely (the one I abandoned unchanged included) allow a small change
like the one you need but I think your only reasonable option is
to open the thing and see for yourself. May be you'll know what to
do within a few minutes, may be you'll give it up as impractical.
Opening may not necessarily be easy; recently I had to cut the
plastic of my laptops brick in order to open it to fix the outgoing
cable....

Dimiter

 

[Don Y]

Hi Dimiter,

Happy New Year!

some years ago I tried this on a small 15V (IIRC) adaptor. It was
somewhat smaller, not an ordinary laptop one but I got it because
it stated its output was settable. It was not so I opened it hoping
to find the feedback divider and change it (it was a cheap DX box).
Turned out the designer had been pretty good and had done things
not that obvious without having the circuit. The feedback was somewhat
hard to modify, a "zener" (TL431) serially feeding current into
the feedback optocoupler - and there was more to it than that
which would have taken hours to understand so I just abandoned
the idea (got another brick).

The adapter is rather large -- probably 20+ cu in. Hopefully, not
as cramped as some of the tinier beasts out there.

And, as it's so large (power), I suspect (?) it may be a bit
better designed than the "500mA disposables".

I suppose there are many different designs out there and most of
them likely (the one I abandoned unchanged included) allow a small change
like the one you need but I think your only reasonable option is
to open the thing and see for yourself. May be you'll know what to
do within a few minutes, may be you'll give it up as impractical.
Opening may not necessarily be easy; recently I had to cut the
plastic of my laptops brick in order to open it to fix the outgoing
cable....

My main concern is no knowledge of the information that is exchanged
(one way? two way? *interactive*??) between the adapter and the
laptop.

I.e., at the very least, the laptop will know that this is a "native"
adapter (same manufacturer) vs. a "foreign" one. Will it know which
*model* adapter? Will it only be willing to operate with *known*
models (BIOS update)?

Will it *command* the adapter to "switch to normal output voltage"
(perhaps it is initially in a low voltage mode that enables the
two devices to converse without putting either in jeopardy)?
Will it query the adapter dynamically to determine current operating
conditions?

If there are smarts in the adapter, will it notice that Vout is not
what it *should* be? Will it try to compensate? etc.

(once you open the door for smarts *in* the power supply, it no
longer is a "trivial component")

I'm not keen on discovering these issues via "smoke signals"!
(they aren't my laptops to "destroy")

--don

 

[dp]

Hi Dimiter,

Happy New Year!

.....
....

My main concern is no knowledge of the information that is exchanged
(one way? two way? *interactive*??) between the adapter and the
laptop.

I have not seen that many laptops but those I have seen don't have
such a thing, just a plain two wire connection. These I have opened
have been far from suggesting the cable could be carrying anything
else but power. I'd be very surprised to see such a thing.

I.e., at the very least, the laptop will know that this is a "native"
adapter (same manufacturer) vs. a "foreign" one. Will it know which
*model* adapter? Will it only be willing to operate with *known*
models (BIOS update)?

Again, I have not seen this done but then I am far from having seen
them all... If you dedicate a day's work on opening the laptop
and investigating (oops, just opening might take you a day...
I remember opening one for cleaning the air path 1-2 years ago,
took me hours).
I have bought 3-4 type of different "laptop chargers" from DX
(to use on anything but laptops, that is) and none of them had
any indications of such a thing (and I have opened... let me see,
I think 4 if not 5 types, one really old, all models would be older
than say 3 years, may be more).

Dimiter

 

[Don Y]

Hi Dimiter,

I have not seen that many laptops but those I have seen don't have
such a thing, just a plain two wire connection. These I have opened
have been far from suggesting the cable could be carrying anything
else but power. I'd be very surprised to see such a thing.

These have a third connector that carries *some* sort of a signal.
E.g., if I take a power supply from HP and try to plug it into a
Dell laptop (same Vout, Imax, connector dimensions) the Dell laptop
will complain. Most typically, it will refuse to charge the
battery (or, will not charge it completely, etc.)

Again, I have not seen this done but then I am far from having seen
them all... If you dedicate a day's work on opening the laptop
and investigating (oops, just opening might take you a day...
I remember opening one for cleaning the air path 1-2 years ago,
took me hours).
I have bought 3-4 type of different "laptop chargers" from DX
(to use on anything but laptops, that is) and none of them had
any indications of such a thing (and I have opened... let me see,
I think 4 if not 5 types, one really old, all models would be older
than say 3 years, may be more).

The two conductor power supplies are admittedly dumb. Easy to
repurpose a power adapter from vendor A for vendor B's equipment
(as long as V & I are compatible and you can get an appropriate
connector on the end).

I've done this many times in the past. E.g., one of our TV's is
powered by a laptop power supply (because the original power supply
died).

 

[Don Y]

I have bought 3-4 type of different "laptop chargers" from DX
(to use on anything but laptops, that is) and none of them had
any indications of such a thing (and I have opened... let me see,
I think 4 if not 5 types, one really old, all models would be older
than say 3 years, may be more).

Have a look at:

<http://www.laptop-junction.com/toast/content/inside-dell-ac-power-adapter-mystery-revealed>

 

[dp]

Have a look at:

<http://www.laptop-junction.com/toast/content/inside-dell-ac-power-adapter-mystery-revealed>

Oh so some have started to put some I2C eeprom or whatever to make
people buy exactly their charger. The laptops I have been buying
are Acer, no such thing. Why should I be surprised really. I think
I have seen a Sony which did not have that sort of thing either
but I am not so sure about that.
If you have one of these laptops with "encrypted" power I don't think
you have many options but to pay them up (likely around $100 rather
than $10...).

One more thing to watch for when buying, one could easily overlook it
I guess.

Dimiter

 

[Jasen Betts]

Hi,

I've got a ~200W laptop brick on which I'd like to tweek the
output voltage (~5% -- 19.5V down to 18.5V). Anyone know what
sort of margins they design into these?

If ~188W output is acceptable at 18.5V tweaking the voltage should be
possible with no ill effects.
(in other words same maximum DC output current)

Typically switchers use a slosed loop matching a (resistive) voltage
divider off the output against a fixed reference, you just need to
find the divider and tweak the ratio.

I'm not keen on risking
the laptop for the sake of a $10 power supply...
(i.e., "it *should* work" is not an acceptable reply :> )

measure it and test it on a dummy load before connecting the laptop.

 

[Tim Williams]

I would be surprised if the laptop cared; most are probably good from
16-25V. Inspecting a few laptop motherboards, it looks like there's a "HV
/ bulk" supply from which all the others are derived; some battery
management circuitry selects the battery or charger (the adapter of course
being higher in voltage so a sync buck and battery controller handle the
actual charging). 14.4V batteries are common (four cells in series, plus
however many in parallel), and so are 18-19V chargers.

Of course, if "probably" isn't good enough, trimming resistors will be in
your future.

Assuming it's a flyback supply, the lower limit will be due to UVLO on the
probably UC3842 controller, supplied by the aux (line side) winding. Once
you crack it open, you could probe the controller or the little
electrolytic in there to see what it runs at, and from there, figure it
can go as low as the turnoff limit. Any lower and it'll always bob on and
off, which is obviously bad. But that's probably 30% lower than you need.

Tim

 

[dp]

I would be surprised if the laptop cared; most are probably good from
16-25V. Inspecting a few laptop motherboards, it looks like there's a "HV
/ bulk" supply from which all the others are derived; some battery
management circuitry selects the battery or charger (the adapter of course
being higher in voltage so a sync buck and battery controller handle the
actual charging).

I would expect to see something like that, too. At least this is how I
did it on the nukeman (design 1993, http://tgi-sci.com/tgi/fr64.gif ,
then around 2000 http://tgi-sci.com/tgi/pntb.htm ).
Basically all convertors were powered off the battery voltage which
was 9 NiMH in series. I had put an overvoltage limit to the charger
stepdown which limited that voltage to about 15 (or was it 15.5)V;
in case of no battery etc., I had opamps powered powered off that
voltage which would not survive over 15.5 (lmc6482).
But the input voltage range was not that generous, I had used a
PMOSFET as the charger stepdown switch and its gate was driven by
an MC33151; which was powered from the adaptor directly, I think
it would not survive anything above 20V. Not so sure how things
were really. I remember I did burn a few 33151s until I got things
to work OK though .

Anyway, Don's issue is with the "smart" part of the adaptor, which
has little to do with voltages.

Dimiter

 

[Don Y]

Hi Dimiter,

Oh so some have started to put some I2C eeprom or whatever to make
people buy exactly their charger.

That's the case for the example I *showed* you. But, I2C can easily
allow more than "identification" to reside in the power supply!
You could just as easily put an MCU in there that monitors voltages,
gates the output on/off, etc.

I.e., the power supply can now become an *integrated* part of the
laptop instead of just some "external voltage source".

The laptops I have been buying
are Acer, no such thing. Why should I be surprised really. I think
I have seen a Sony which did not have that sort of thing either
but I am not so sure about that.
If you have one of these laptops with "encrypted" power I don't think
you have many options but to pay them up (likely around $100 rather
than $10...).

In my case, the concern is whether a *different* power supply
(by the same manufacturer as the "ideal" power supply) would
"behave" when connected to the laptop.

The laptop could decide not to recognize the power supply
("wrong ID"). Or, some feature may become unavailable
(e.g., battery charging seems to be the most common one).
Or, something even more nefarious (e.g., if the power supply
is "actively smart")

One more thing to watch for when buying, one could easily overlook it
I guess.

Yup. I've seen "universal power supplies" with assortments of
interchangeable "connectors" (and adjustable Vout). Even those
including these "three pin" connectors. But, if they don't
know how to identify themselves as the appropriate *actual*
power supply, then what good are they?

 

[Don Y]

Hi Jasen,

If ~188W output is acceptable at 18.5V tweaking the voltage should be
possible with no ill effects.
(in other words same maximum DC output current)

That would depend on the actual topology used. For a trivial
example, imagine a *linear* where the change would result in
an extra ~10W going up as heat at that same output current.

Typically switchers use a slosed loop matching a (resistive) voltage
divider off the output against a fixed reference, you just need to
find the divider and tweak the ratio.


measure it and test it on a dummy load before connecting the laptop.

Measure *what*? Vout is a no brainer. But, how do I know
something inside the supply isn't operating closer to its
limits?

Remember, this is consumer kit. Everything seems to pride itself
on how close to the edge it can be designed *and* how few days
beyond warranty it can survive!

The same mentality that saves fractional cents on caps is at play, here.

 

[Don Y]

Hi Dimiter,

I would expect to see something like that, too. At least this is how I
did it on the nukeman (design 1993, http://tgi-sci.com/tgi/fr64.gif ,
then around 2000 http://tgi-sci.com/tgi/pntb.htm ).
Basically all convertors were powered off the battery voltage which
was 9 NiMH in series. I had put an overvoltage limit to the charger
stepdown which limited that voltage to about 15 (or was it 15.5)V;
in case of no battery etc., I had opamps powered powered off that
voltage which would not survive over 15.5 (lmc6482).
But the input voltage range was not that generous, I had used a
PMOSFET as the charger stepdown switch and its gate was driven by
an MC33151; which was powered from the adaptor directly, I think
it would not survive anything above 20V. Not so sure how things
were really. I remember I did burn a few 33151s until I got things
to work OK though .

Anyway, Don's issue is with the "smart" part of the adaptor, which
has little to do with voltages.

I'm concerned about the entire "subsystem" -- as the power supplies
are obviously moving beyond "dumb voltage sources" to more of an
"integrated system". When you start integrating things, you tend
to use them (and design them) in ways that leverage other aspects
of the system to improve their performance, decrease their cost
or provide added functionality.

It seems like there is little value to encoding *just* an "ID"
in a power supply. AFAICT, you can't make your initial purchase
(of the laptop) without getting a power supply in the deal. So,
the ID only protects sales of *replacement* power supplies.

[It also could give some liability protection if third party
power supplies are resulting in battery fires, etc. "Due diligence"]

Is there a big problem with folks *losing* their power supplies?
Or, are there plans to artificially grow the power supply market
with a move to even crappier designs/reliability??

It looks like the safer approach, for me, is to find a deal on
the "Right" power supplies (buying them by the dozen can help).
Or, convince a donor to foot the bill for them...

(though it would be an interesting question to have answered!)

 

[Lasse Langwadt Christensen]

Den onsdag den 8. januar 2014 21.30.54 UTC+1 skrev Don Y:

Hi Dimiter,

I would expect to see something like that, too. At least this is how I

did it on the nukeman (design 1993, http://tgi-sci.com/tgi/fr64.gif ,

then around 2000 http://tgi-sci.com/tgi/pntb.htm ).

Basically all convertors were powered off the battery voltage which

was 9 NiMH in series. I had put an overvoltage limit to the charger

stepdown which limited that voltage to about 15 (or was it 15.5)V;

in case of no battery etc., I had opamps powered powered off that

voltage which would not survive over 15.5 (lmc6482).

But the input voltage range was not that generous, I had used a

PMOSFET as the charger stepdown switch and its gate was driven by

an MC33151; which was powered from the adaptor directly, I think

it would not survive anything above 20V. Not so sure how things

were really. I remember I did burn a few 33151s until I got things

to work OK though .

Anyway, Don's issue is with the "smart" part of the adaptor, which

has little to do with voltages.

I'm concerned about the entire "subsystem" -- as the power supplies

are obviously moving beyond "dumb voltage sources" to more of an

"integrated system". When you start integrating things, you tend

to use them (and design them) in ways that leverage other aspects

of the system to improve their performance, decrease their cost

or provide added functionality.



It seems like there is little value to encoding *just* an "ID"

in a power supply. AFAICT, you can't make your initial purchase

(of the laptop) without getting a power supply in the deal. So,

the ID only protects sales of *replacement* power supplies.



[It also could give some liability protection if third party

power supplies are resulting in battery fires, etc. "Due diligence"]



Is there a big problem with folks *losing* their power supplies?

Or, are there plans to artificially grow the power supply market

with a move to even crappier designs/reliability??



It looks like the safer approach, for me, is to find a deal on

the "Right" power supplies (buying them by the dozen can help).

Or, convince a donor to foot the bill for them...



(though it would be an interesting question to have answered!)

most laptop chargers have a similar barrel connector, only the voltage and power varies all the smarts are in the laptop

and you can get universal car adaptors that just a buck/boost to set the
right voltage and a set of connectors

There's an effort to standardize laptop chargers, similar I guess to how EU mandated that cell phones must be able to charge from USB


-Lasse

 

[Greegor]

If the idiotic little ID chips are doing nothing
but ID, then wouldn't it be easier to modify the
BIOS code on the laptop so that it doesn't care
about the ID code? The BIOS is flashable, right?

 

[Don Y]

Hi Lasse,

Den onsdag den 8. januar 2014 21.30.54 UTC+1 skrev Don Y:

Anyway, Don's issue is with the "smart" part of the adaptor, which
has little to do with voltages.

I'm concerned about the entire "subsystem" -- as the power supplies
are obviously moving beyond "dumb voltage sources" to more of an
"integrated system". When you start integrating things, you tend
to use them (and design them) in ways that leverage other aspects
of the system to improve their performance, decrease their cost
or provide added functionality.

It seems like there is little value to encoding *just* an "ID"
in a power supply. AFAICT, you can't make your initial purchase
(of the laptop) without getting a power supply in the deal. So,
the ID only protects sales of *replacement* power supplies.

[It also could give some liability protection if third party
power supplies are resulting in battery fires, etc. "Due diligence"]

Is there a big problem with folks *losing* their power supplies?
Or, are there plans to artificially grow the power supply market
with a move to even crappier designs/reliability??

It looks like the safer approach, for me, is to find a deal on
the "Right" power supplies (buying them by the dozen can help).
Or, convince a donor to foot the bill for them...

(though it would be an interesting question to have answered!)

most laptop chargers have a similar barrel connector, only the voltage
and power varies all the smarts are in the laptop

That's not the case with these. *Three* conductors. The third of which
carries "data" from (to?) the charger. At the very least, the laptop
can say, "This is not an approved (i.e., sold by me) power adaptor.
I refuse to ___________"

As seen in the URL I posted (elsewhere) for Dimiter, at least one
such adapter has a ROM inside that the laptop queries. Presumably,
if the ROM is absent or contains "funky" data, the laptop could opt
not to perform (or, perform incorrectly).

It's unclear if other power supplies (including the ones in
question) do anything *active* with the comm link (e.g., transfer
information about the temperature of the power supply, load current,
etc.). Or, if the laptop *commands* the power supply ("enable
charging mode") during operation.

and you can get universal car adaptors that just a buck/boost to set the
right voltage and a set of connectors

Different beast.

There's an effort to standardize laptop chargers, similar I guess to
how EU mandated that cell phones must be able to charge from USB

The fact that vendors (at least IBM, Dell and HP, from my experience)
are adding this "feature" to their power adaptors suggests they
want to hold onto the power adapter business (?). Note you don't
see much effort to standardize ink jet cartridges :> Nor do
I imagine any printer vendors will be quick to embrace such an idea!

 

[Ralph Barone]

Hi Lasse,

Den onsdag den 8. januar 2014 21.30.54 UTC+1 skrev Don Y:

On 1/8/2014 12:51 AM, dp wrote:

Anyway, Don's issue is with the "smart" part of the adaptor, which
has little to do with voltages.



I'm concerned about the entire "subsystem" -- as the power supplies
are obviously moving beyond "dumb voltage sources" to more of an
"integrated system". When you start integrating things, you tend
to use them (and design them) in ways that leverage other aspects
of the system to improve their performance, decrease their cost
or provide added functionality.

It seems like there is little value to encoding *just* an "ID"
in a power supply. AFAICT, you can't make your initial purchase
(of the laptop) without getting a power supply in the deal. So,
the ID only protects sales of *replacement* power supplies.

[It also could give some liability protection if third party
power supplies are resulting in battery fires, etc. "Due diligence"]

Is there a big problem with folks *losing* their power supplies?
Or, are there plans to artificially grow the power supply market
with a move to even crappier designs/reliability??

It looks like the safer approach, for me, is to find a deal on
the "Right" power supplies (buying them by the dozen can help).
Or, convince a donor to foot the bill for them...

(though it would be an interesting question to have answered!)

most laptop chargers have a similar barrel connector, only the voltage
and power varies all the smarts are in the laptop

That's not the case with these. *Three* conductors. The third of which
carries "data" from (to?) the charger. At the very least, the laptop
can say, "This is not an approved (i.e., sold by me) power adaptor.
I refuse to ___________"

As seen in the URL I posted (elsewhere) for Dimiter, at least one
such adapter has a ROM inside that the laptop queries. Presumably,
if the ROM is absent or contains "funky" data, the laptop could opt
not to perform (or, perform incorrectly).

It's unclear if other power supplies (including the ones in
question) do anything *active* with the comm link (e.g., transfer
information about the temperature of the power supply, load current,
etc.). Or, if the laptop *commands* the power supply ("enable
charging mode") during operation.

and you can get universal car adaptors that just a buck/boost to set the
right voltage and a set of connectors

Different beast.

There's an effort to standardize laptop chargers, similar I guess to
how EU mandated that cell phones must be able to charge from USB

The fact that vendors (at least IBM, Dell and HP, from my experience)
are adding this "feature" to their power adaptors suggests they
want to hold onto the power adapter business (?). Note you don't
see much effort to standardize ink jet cartridges :> Nor do
I imagine any printer vendors will be quick to embrace such an idea!

About the only good reason I can think of to have the power supply "talk"
to the laptop is to allow different battery charge rates for different
sized chargers, therefore allowing the manufacturer to sell either a big
"home" charger or a small "road" charger and have the laptop charge as fast
as possible with either one.

 

[Don Y]

Hi Ralph,

About the only good reason I can think of to have the power supply "talk"
to the laptop is to allow different battery charge rates for different
sized chargers, therefore allowing the manufacturer to sell either a big
"home" charger or a small "road" charger and have the laptop charge as fast
as possible with either one.

I can imagine a power supply "telling" the laptop that it (the power
supply) is running hot and, thus, should reduce its load (esp given
that the laptop has control over a *big* piece of its load -- the
charger!).

I can also imagine a power supply telling the laptop how much power
it is *actually* using (adding those monitoring components *outside*
the physical constraints of the laptop itself) so the laptop can
better calibrate its battery monitoring algorithms (instead of
just watching Vbat without any real idea as to how quickly it
*should* be falling -- when operating on battery).

And, the mere presence of a third conductor could be exploited to
*move* the charger into the power adapter using the third conductor
to deliver "charge current" separate from "operating power" for the
laptop. Even potentially multiplexing (TDM) data onto that conductor!

As I said, once you open the door to smarts in the power supply, you
can come up with lots of more interesting designs with increased
levels of integration across the "power cord" boundary!

[But, other than the URL I posted suggesting "ID" information,
I have no idea how power supplies currently use the conductor]