Mount external modem inside PC? Power from PSU?

[Noozer]

If I have an external phone modem that uses a 1amp 9vDC tranformer, could I
power this from the 12v line of my PC PSU? (assuming that the modem will run
on 12v)

 

[CBFalconer]

If I have an external phone modem that uses a 1amp 9vDC tranformer,
could I power this from the 12v line of my PC PSU? (assuming that
the modem will run on 12v)

As usual, that depends. If you publish the schematic of the modem,
we can probably better evaluate it. The most likely thing is that
the 9 VAC gets full wave rectified resulting in about 12 VDC inside
the unit, in which case it should work (but the turn on load will
be high, to charge the input filter capacitor).

Oops, you said it was 9 VDC. Now we need the rest of the
schematic, including chip data sheets. However a set of 4 silicon
power diodes (say 1N4001) in series will drop the 12V DC to about 9
VDC, which should also work. They will be dissipating about 3
watts.

--
<http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>

"A man who is right every time is not likely to do very much."
-- Francis Crick, co-discover of DNA
"There is nothing more amazing than stupidity in action."
-- Thomas Matthews

 

[paulmd]

If I have an external phone modem that uses a 1amp 9vDC tranformer, could I
power this from the 12v line of my PC PSU? (assuming that the modem will run
on 12v)

Mismatching power supplies is a craps shoot.

 

[Rod Speed]

If I have an external phone modem that uses a 1amp 9vDC tranformer,

That would be unusual.

could I power this from the 12v line of my PC PSU? (assuming that the modem will run on 12v)

It should be ok voltage wise, it should be regulated in the modem.

BUT there may be only enough heatsink on the regulator
to be able to handle 9V and not 12V without overheating
on the hottest days of summer if you arent airconditioned.

And it isnt a great idea to be taking the 12V rail outside
the case either, its a very high current rail in modern ATX
supplys, so the effect of a short can be rather spectacular
and electronically fatal if its a cheap power supply.

 

[Skeleton Man]

If I have an external phone modem that uses a 1amp 9vDC tranformer, could I

power this from the 12v line of my PC PSU? (assuming that the modem will run
on 12v)

Might I ask what the purpose of this is ? (vs a PCI modem instead)

Chris

 

[kony]

That would be unusual.

True, but if it is really 9VDC, next is to determine how
much current it uses (if there's a multimeter handy) or the
more conservative (higher) equipment current rating, or the
overly conservative and probably less useful original PSU
(wall wart?) current rating.

I seriously doubt it comes remotely close to 1A, none of the
external modems I've had got that hot for one thing, maybe
closer to 250mA would be reasonable.

Anyway, the easiest thing to to here is probably put an
LM7809 linear regulator between the 12V PSU rail and the
modem. It'll produce closer to a watt of heat, more than
you really want on a TO220 sized part so put a heatsink on
it. It need not be a large 'sink, particularly if there's
some ventilation.

It should be ok voltage wise, it should be regulated in the modem.

BUT there may be only enough heatsink on the regulator
to be able to handle 9V and not 12V without overheating
on the hottest days of summer if you arent airconditioned.

It would be easier if we knew what modem this was, perhaps a
link to a reasonably high resolution picture of the inside,
top-down.

And it isnt a great idea to be taking the 12V rail outside
the case either, its a very high current rail in modern ATX
supplys, so the effect of a short can be rather spectacular
and electronically fatal if its a cheap power supply.

So?
Use
A
Fuse

 

[Rod Speed]

True, but if it is really 9VDC, next is to determine how
much current it uses (if there's a multimeter handy) or the
more conservative (higher) equipment current rating, or the
overly conservative and probably less useful original PSU
(wall wart?) current rating.

I seriously doubt it comes remotely close to 1A,

Plenty of mine are of that order.

none of the external modems I've had got that hot for
one thing, maybe closer to 250mA would be reasonable.

I doubt it.

Anyway, the easiest thing to to here is probably put an LM7809
linear regulator between the 12V PSU rail and the modem.

Makes a lot more sense to keep using an external wall wart instead.

It'll produce closer to a watt of heat, more than you really
want on a TO220 sized part so put a heatsink on it. It need
not be a large 'sink, particularly if there's some ventilation.

What is the point over just using a wall wart ? A switch mode one if
you dont like the constant power used of a transformer based one.

It would be easier if we knew what modem this was, perhaps a
link to a reasonably high resolution picture of the inside, top-down.

So?

So, its too much farting around.

Use
A
Fuse

Too
Much
Farting
Around

Makes a lot more sense to stick with a wall wart.

 

[CBFalconer]

Might I ask what the purpose of this is ? (vs a PCI modem instead)

External modems are much more useful. The presence of activity
lights alone gives you clues as to what is going on.

I still like my earlier suggestion of 4 rectifier diodes.
Including a 1 amp fuse in the line would be valuable. Everything
goes in series, and the result can be in line enclosed in some
shrinkable tubing. The suggestion got somewhat lost in my
rambling.

 

[James Brown]

External modems are much more useful. The presence of activity
lights alone gives you clues as to what is going on.

I still like my earlier suggestion of 4 rectifier diodes.
Including a 1 amp fuse in the line would be valuable. Everything
goes in series, and the result can be in line enclosed in some
shrinkable tubing. The suggestion got somewhat lost in my rambling.

Can't see that being very viable, 3W in heatshrink isnt a great idea.

 

[kony]

Plenty of mine are of that order.

The wart might be, but I doubt the modem itself is.
If it used 1A @ 9V, that's of course 9W (practically all
heat generated internally), and nothing gets that hot to the
touch or needs significant heatsinking. On the other hand
consider a router using 9V @ 500mA, it may easily need a
'sink just to run stable.

I doubt it.

Measure, it's the only way. It is not typical for a wart to
be spec'd for 100% duty, very doubtful the modem uses much
over 500mA if that.

Makes a lot more sense to keep using an external wall wart instead.

If he's keeping the modem external, yes. If not, it depends
on how much work he's willing to put into it as a wart is
easier to source but more expensive to replace than adding
a 30 cent linear regulator. We don't know exactly what the
goal is and what concessions would be made soonest if that
goal entails significant, undesirable measures.

What is the point over just using a wall wart ? A switch mode one if
you dont like the constant power used of a transformer based one.

Switch mode wart has higher failure rate
It may look ghetto to have a cord wedged, running into the
case from outside to get the power in.
There is already an (after dropped) acceptible power source
inside.

It's really a matter of time, esthetically use of the
internal supply is better, but takes longer to implement.
If he happened to have a molex plug, a suitable linear
regulator and an idea for securing it all, it would be a
pretty quick job, but if unaccustomed to this kind of work
and doing it from scratch it could certainly be as
reasonable to use the external supply still... only with
enough details is an informed choice made.

 

[kony]

Can't see that being very viable, 3W in heatshrink isnt a great idea.

It could be an issue but we haven't established that the
modem actually uses 1A. It almost certainly does not, but
what it actually uses ... ?

 

[Roby]

External modems are much more useful. The presence of activity
lights alone gives you clues as to what is going on.

I still like my earlier suggestion of 4 rectifier diodes.
Including a 1 amp fuse in the line would be valuable. Everything
goes in series, and the result can be in line enclosed in some
shrinkable tubing. The suggestion got somewhat lost in my
rambling.

It took a while to FIND my modem. Mine's a USR external that uses
9VAC from a supplied 1000ma wall xfmr - not like the OP's modem.
I opened it up anyway: there's a 7805 regulator nearby ... lots
of vanishingly small surface mount stuff. There's one issue that
might be a problem not addressed here: wall-wart sources float -
the connected circuit can ground either or neither end. Computer
supplies outputs are grounded. Without knowing the circuit details,
powering an external modem from the computer is gamble that might
make toast.

 

[Rod Speed]

The wart might be, but I doubt the modem itself is.

Plenty were, most obviously those that came in an
extruded aluminum sleeve and got quite hot anyway.

If it used 1A @ 9V, that's of course 9W (practically
all heat generated internally), and nothing gets that
hot to the touch or needs significant heatsinking.

Which is why quite a few of them came in an extruded
aluminum sleeve and got quite hot anyway.

On the other hand consider a router using 9V @
500mA, it may easily need a 'sink just to run stable.

Mine have all been stable without one, but do get quite
warm on the hottest days in summer without air conditioning.

Measure, it's the only way.

I doubt too many would bother to supply a
1A wall wart if the modem only needs 250mA

It is not typical for a wart to be spec'd for 100% duty,
very doubtful the modem uses much over 500mA if that.

You dont know that his is a modern modem.

If he's keeping the modem external, yes.

Even more pointless making an external modem internal unless its very unique.

If not, it depends on how much work he's willing to put
into it as a wart is easier to source but more expensive
to replace than adding a 30 cent linear regulator.

You arent comparing apples and apples.
It would cost peanuts to repair a wall wart.

We don't know exactly what the goal is and what concessions would
be made soonest if that goal entails significant, undesirable measures.
Duh.

Switch mode wart has higher failure rate
Wrong.

It may look ghetto to have a cord wedged,
running into the case from outside to get the power in.

Why would anyone be stupid enough to mount
an external modem inside the PC case ?

There is already an (after dropped) acceptible power source inside.

There are hordes of wall warts available.

It's really a matter of time, esthetically use of the internal supply is better,

Nope, which is why external modems come with wall warts.

but takes longer to implement. If he happened to have
a molex plug, a suitable linear regulator and an idea
for securing it all, it would be a pretty quick job,

Bullshit on the mechanicals.

but if unaccustomed to this kind of work
and doing it from scratch it could certainly be
as reasonable to use the external supply still...

Much more reasonable in fact in almost all circumstances except your
completely silly scenario of an external modem installed inside the PC case.

only with enough details is an informed choice made.

Wrong again, installing an external modem inside a PC case is even sillier.

 

[CBFalconer]

It took a while to FIND my modem. Mine's a USR external that uses
9VAC from a supplied 1000ma wall xfmr - not like the OP's modem.
I opened it up anyway: there's a 7805 regulator nearby ... lots
of vanishingly small surface mount stuff. There's one issue that
might be a problem not addressed here: wall-wart sources float -
the connected circuit can ground either or neither end. Computer
supplies outputs are grounded. Without knowing the circuit details,
powering an external modem from the computer is gamble that might
make toast.

An AC input is a different beast. It may well be driving both
positive and negative interior supplies, in fact it probably is,
since the RS232 output needs to swing to negative levels. DC input
won't work. Incidentally, my USR modem wallwart is labelled for
800 mA at 9 VAC. It has been running for about 10 years, on
various machines.

For the DC input machine, there is a signal ground, and probably a
system ground (both are specified for RS232). They should be
hooked together. Measuring input volts against that point will
show if the system requires +9 or -9 and which power pin is
ground. If it needs -9 VDC then the direct supply won't work
either, unless the PS has a -12V output line. That is specified,
but most systems don't use it, so the PS may have omitted it. A
little time with a voltmeter will tell all.

 

[kony]

Plenty were, most obviously those that came in an
extruded aluminum sleeve and got quite hot anyway.

Quite hot and overheated are not the same thing.

Which is why quite a few of them came in an extruded
aluminum sleeve and got quite hot anyway.

See above. That is no proof of any particular heat density

I doubt too many would bother to supply a
1A wall wart if the modem only needs 250mA

Whether the figures is close to 250mA or not, we can assume
the supply current is rated higher than the device by a
reasonable margin.

You dont know that his is a modern modem.

.... and you don't know it isn't, but we both know it had a
1A supply and that supplies are always rated higher than the
load.

We don't have to guess about it though, a measurement can be
made.

Even more pointless making an external modem internal unless its very unique.

There is something to be said for not having extra boxes and
wires everywhere. Only the OP can decide if this or
something else is important, subjectively.

You arent comparing apples and apples.
It would cost peanuts to repair a wall wart.

Since I have, I can describe it. They're typically fused
shut so you have to mangle the sides forcing it to pop
apart in a bench vise. Typically it's the thermal fuse, so
you have to pull the transformer out (may require
desoldering from the AC receptacle blades), pull the outer
wrapper off, and swap the fuse- unwinding the windings a bit
if it was built properly with the fuse covered by wire
instead of on the outside. Next repair the mangled sides on
the plastic case, generally a touch-up with a file. Solder
the transformer back in, glue it and clamp it for a day.

Hell of a lot of work considering what a wart is, seldom do
I even bother except if proprietary enough to warrant it.
The time spent is easily worth more than a 30 cent
regulator, a capacitor or two, and misc. bits worth pennies
each like a piece of wire or heatsink tubing.

Wrong.

Right

I have 30+ year old unregulated warts that work fine still.
It's quite rare to find a switcher that lasts 20 years, and
I happen to have one in front of me at the moment that
lasted less than 5 (3.3V, 5A from a 24 port hub).

Why would anyone be stupid enough to mount
an external modem inside the PC case ?

Why would anyone be stupid enough to keep posting about
hardware when it makes them upset enough to rant about
clowns and bags and suggestions to have fun as if we needed
that invitation?

There are hordes of wall warts available.

Yes, but since this was to go inside a computer case, were
you suggesting the OP install an AC outlet in there for the
wart? Of course not.

Nope, which is why external modems come with wall warts.


Bullshit on the mechanicals.

You're right, it should've been "socket", not plug.

 

[kony]

It took a while to FIND my modem. Mine's a USR external that uses
9VAC from a supplied 1000ma wall xfmr - not like the OP's modem.
I opened it up anyway: there's a 7805 regulator nearby ... lots
of vanishingly small surface mount stuff. There's one issue that
might be a problem not addressed here: wall-wart sources float -
the connected circuit can ground either or neither end. Computer
supplies outputs are grounded. Without knowing the circuit details,
powering an external modem from the computer is gamble that might
make toast.

Didn't you feel the modem would be grounded by the serial
port?

 

[Rod Speed]

Quite hot and overheated are not the same thing.

Never ever said it was. It is a reasonable indication that
it might well be using something of the order of 9W tho.

See above.

See above.

That is no proof of any particular heat density

It is a reasonable indication that it might well
be using something of the order of 9W tho.

Whether the figures is close to 250mA or not,
we can assume the supply current is rated
higher than the device by a reasonable margin.

Bet its nothing like that sort of margin.

... and you don't know it isn't, but we both know it had a 1A
supply and that supplies are always rated higher than the load.

I doubt too many would bother to supply a
1A wall wart if the modem only needs 250mA

We don't have to guess about it though, a measurement can be made.

I doubt too many would bother to supply a
1A wall wart if the modem only needs 250mA

There is something to be said for not
having extra boxes and wires everywhere.

But nothing to be said for mounting and external modem internally.

Only the OP can decide if this or something else is important, subjectively.

Never ever could bullshit its way out of a wet paper bag.

Since I have, I can describe it.

Whoopy ****ing do, you might not actually be alone on that, child.

Wrong.

I have 30+ year old unregulated warts that work fine still.
It's quite rare to find a switcher that lasts 20 years,

Only because they werent at all common then.

and I happen to have one in front of me at the moment
that lasted less than 5 (3.3V, 5A from a 24 port hub).

The technical term for that is 'pathetically inadequate sample'

Why would anyone be stupid enough to keep posting about hardware
when it makes them upset enough to rant about clowns and bags and
suggestions to have fun as if we needed that invitation?

Never ever could bullshit its way out of a wet paper bag.

Yes, but since this was to go inside a computer case,

You dont know that.

were you suggesting the OP install an AC
outlet in there for the wart? Of course not.

Never ever could bullshit its way out of a wet paper bag.

You're right, it should've been "socket", not plug.

Never ever could bullshit its way out of a wet paper bag.

 

[kony]

It is a reasonable indication that it might well
be using something of the order of 9W tho.

No, given any wart will be conservatively spec'd, we can be
sure the power consumption is _below_ 9W and quite possibly
much lower.

Bet its nothing like that sort of margin.

A bet would be silly, measure it.
Since I don't have the OP's unknown model of modem here (or
maybe I do, kinda hard to say...), the best approximation
would be from other consumer devices using warts. They are
typically spec'd to not have peak (continuous) consumption
over 75% capacity, if even that.

To put it another way, if they provided a 9V, 1A wart and it
was using 1A of power, that alone is a good reason to stop
using it, as it will be getting quite hot and short lived if
in the typical sealed plastic wart case.

 

[Rod Speed]

Yep.

given any wart will be conservatively spec'd,

By nothing like as much that stupid claim you made
about 250mA you just plucked out of your arse.

we can be sure the power consumption is _below_ 9W
Duh.

and quite possibly much lower.

Nope, no chance. And nothing like that stupid claim you
made about 250mA you just plucked out of your arse.

A bet would be silly, measure it.

It wont be anything like that stupid claim you made
about 250mA you just plucked out of your arse.

Since I don't have the OP's unknown model of modem here
(or maybe I do, kinda hard to say...), the best approximation
would be from other consumer devices using warts.

Wrong again, the best approach if it isnt possible to measure it is to
assume that the wall wart he said he wants to replace isnt that over rated.

They are typically spec'd to not have peak (continuous)
consumption over 75% capacity, if even that.

Still nothing like that stupid claim you made
about 250mA you just plucked out of your arse.

To put it another way, if they provided a 9V, 1A wart and it
was using 1A of power, that alone is a good reason to stop
using it, as it will be getting quite hot and short lived if
in the typical sealed plastic wart case.

Stupid to not just replace it with an adequate wall wart.

 

[paulmd]

If I have an external phone modem that uses a 1amp 9vDC tranformer, could I
power this from the 12v line of my PC PSU? (assuming that the modem will run
on 12v)

Only one guy asked why the hell you want to do that, so far, and you
haven't answered. Care to clue us in?