I want a passively cooled desktop/cpu/gpu by 2016, power supply may have fan.

[Skybuck Flying]

Hello,

I play to buy a new computer by 2016, currently I am concerned that the
computer I want is not for sale/possible.

So far intel and/or amd is not selling passively cooleable CPUs for desktops
?! Could this explain the decrease in PC sales ? People diverting to
noiseless tablets ?

Amazing that no passively cooleable cpus seem to be for sale ?

With passively cooleable cpu I consider:

1. CPU + heatsink with fins smallist, no fan, no other things, except maybe
thermal paste interface material between cpu and heatsink.

So two questions:

1. Which desktop cpu by 2016 will be passively cooleable ?

2. How many watts are passively cooleable with a small heatsink/fins, like
gt 520 from asus is a good example.

(Also the motherboard must be passively cooled as well as all other
components except perhaps power supply, though if a modest power supply is
needed might be passively cooled too, but a little bit of airflow seems
wise... though would be cool if it wasnt needed at all, than no dust in pc
which would be excellent.)

Bye,
Skybuck.

[Skybuck Flying]

This website mentions 10 watt idle and 29 watt active for gt 520 which is a
good example of a chip being passively cooled, though power supply does suck
some heat from it probably:

"
Power consumption should also be similar; NVIDIA gives the GT 520 a TDP of
29W, while we’d expect the idle TDP to be around 10W.
"

http://www.anandtech.com/show/4268/nvidia-releases-geforce-gt-520

So 25 watts seems to be on the safe side.

Bye,
Skybuck.

[Skybuck Flying]

* and laptops

"Skybuck Flying" wrote in message
news:55c40$50c3c616$5419b3e4$35...@cache1.tilbu1.nb.home.nl...

Hello,

I play to buy a new computer by 2016, currently I am concerned that the
computer I want is not for sale/possible.

So far intel and/or amd is not selling passively cooleable CPUs for desktops
?! Could this explain the decrease in PC sales ? People diverting to

noiseless tablets ? *

[MitchAlsup]

On Saturday, December 8, 2012 4:58:34 PM UTC-6, Skybuck Flying wrote:
> So far intel and/or amd is not selling passively cooleable CPUs for desktops ?!
> Could this explain the decrease in PC sales ?

A) Intel and AMD (note use of capital characters) do not sell PCs, they sell
components that go into PCs. It is sloppy and slothful not to correctly
capitalize names.

B) The decline in PC sales is due to the fact that my wife can do everything
she needs a PC to do on a tablet. {Internet, printing, email, word} This means
that only people with heavy workloads, intricate drawings,... need a PC anymore.

Mitch

[Dave Platt]

In article <92a58$50c3ca65$5419b3e4$84...@cache1.tilbu1.nb.home.nl>,

Skybuck Flying <Window...@DreamPC2006.com> wrote:

>I play to buy a new computer by 2016, currently I am concerned that the
>computer I want is not for sale/possible.
>
>So far intel and/or amd is not selling passively cooleable CPUs for desktops
>?! Could this explain the decrease in PC sales ? People diverting to
>noiseless tablets ? *
>
>Amazing that no passively cooleable cpus seem to be for sale ?
>
>With passively cooleable cpu I consider:
>
>1. CPU + heatsink with fins smallist, no fan, no other things, except maybe
>thermal paste interface material between cpu and heatsink.
>
>So two questions:
>
>1. Which desktop cpu by 2016 will be passively cooleable ?
>
>2. How many watts are passively cooleable with a small heatsink/fins, like
>gt 520 from asus is a good example.

>(Also the motherboard must be passively cooled as well as all other
>components except perhaps power supply, though if a modest power supply is
>needed might be passively cooled too, but a little bit of airflow seems
>wise... though would be cool if it wasnt needed at all, than no dust in pc
>which would be excellent.)

As is so often the case, your questions are so poorly defined that
there is either no answer to them, or dozens of answers (some being
quite silly). For example, you can certainly cool a thousand-watt
device with a simple passive heatsink with fins, but you're likely to
find the operating temperature and lifetime of the device to be
inadequate for home use :-)

There are now, and have been for years, Intel (and compatible) CPUs
which gave good performance for their era, and which were passively
cooled. Many of these were and are quite suitable for use in desktop
systems and notebook computers, assuming proper system design.

A lot of these are sold as "industrial computer" or "embedded
computer" devices rather than "desktop", but they run the same code
and do the same jobs... the terminology is as much marketing-speak as
it is real.

For a simple example: I've got a single-core, hyperthreaded Intel 1.6
GHz Atom N270 at the heart of my home firewall / VoIP server / email
server / Web server, on a mini-ITX motherboard mounted in a
desktop-computer tower case. It's passively cooled... just a small
finned heatsink... no fan on the CPU or on the motherboard chipset or
DRAM. The only fan in the box is in the ATX-style power supply... and
I could have done without that if I had bought the version of this
motherboard which has a simple 12-volt power input jack, and ran the
whole system from an external switching "wall wart" power supply.

A lot of the "industrial PC" systems I've seen are not only passively
cooled, they are sealed boxes (hermetically sealed or close to it)...
intended for operation under harsh conditions (dust, humidity,
vibration) they depend entirely on passive cooling. Typically they'll
have some sort of mechanical head-spreader between the CPU and the
case, so the case and its external fins serve as the cooling system.

If you're fixated on "desktop" CPUs (i.e. those which can be installed
into sockets on traditional motherboards in traditional "desktop"
cases)... well, the vendors seem to focus more on performance than on
passive cooling, and so you may not find anything to your liking.
That's even more true for GPUs, for which performance is almost
everything... and so they run the chips fast, and this generates a lot
of heat. I rather doubt that these traditional-market "desktop"
systems will focus on passive CPU cooling any time soon - there's not
much incentive for the manufacturers to do so.

You can reduce the heat generated in almost any CPU or GPU by
underclocking it and turning the voltage down, but of course this
costs you much of the performance.

As to your question about "how many watts" - until you define the
actual heatsink, *and* the environment in which it's operating (e.g.
case air-flow if any, extra heat spreading, air temperature inside the
case, etc.), tell us what sort of temperatures the CPU chip can
handle, and how long you want the CPU to survive the heat, there's no
way for anybody to give you a solid answer.

As a very rough rule of thumb, though, I would suggest that running
more than about 4-5 watts, in the chip the size of a modern CPU, will
require either some form of forced air across a heatsink, or some
fairly direct thermal connection of the chip to a larger heat
spreader. Beyond this point, a modest finned heatsink which depends
on air convection and radiation isn't going to keep the chip cool
enough for a good long lifetime. The larger the heatsink you are
willing to tolerate, the more heat you can dissipate without needing
forced airflow.

This all may be irrelevant, though. Apparently, Intel has announced
that they plan to discontinue their lines of desktop CPUs which plug
into sockets. They're going to be moving their new CPU lines to
BGA (ball-grid array) only, and these must be soldered to the
motherboard. "Enthusiasts" will no longer be able to buy a
motherboard with socket, and populate it with the CPU of their
choice... you'll have to buy the motherboard with the CPU already
soldered on, and will not be able to change it (BGA removal and rework
is possible but it's a terror, from all I've heard).

So, your 2016 Dream PC may be an Impossible Dream.

--
Dave Platt <dpl...@radagast.org> AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

[Quadibloc]

Why shouldn't they be able to do this?

After all, look at the chips they're putting in tablets these days.
They don't need fans, and they're a lot more powerful than the 386
chips of yesteryear that didn't need fans.

So why can't you buy such a desktop computer right now?

Well, two reasons.

Microsoft doesn't sell Windows 3.1 any more. To run the operating
systems it sells now, the CPU's power has to be big and bloated like
the operating system it runs. Even laptops these days have fans.

A deskop computer costs more money than a tablet. So if you're going
to pay that extra money, you want to get somethng for it. Like the
ability to run Windows programs.

John Savard

[Jeff Liebermann]

On Sat, 8 Dec 2012 17:25:59 -0800, dpl...@radagast.org (Dave Platt)
wrote:

>Apparently, Intel has announced
>that they plan to discontinue their lines of desktop CPUs which plug
>into sockets. They're going to be moving their new CPU lines to
>BGA (ball-grid array) only, and these must be soldered to the
>motherboard. "Enthusiasts" will no longer be able to buy a
>motherboard with socket, and populate it with the CPU of their
>choice... you'll have to buy the motherboard with the CPU already
>soldered on, and will not be able to change it (BGA removal and rework
>is possible but it's a terror, from all I've heard).

I rarely see a blown Intel CPU. However, dead motherboards are
epidemic. Lots of reasons, but basically anything that's customer
accessible will eventually be broken. On laptops, BGA soldering
failure is the most common source of total failure. So, if we take a
nearly 100% reliable Intel chip, and couple it with a not so reliable
motherboard held together by a not so reliable interconnect scheme, we
get a not so reliable combination. That should sell plenty of
un-necessary CPU's to fix motherboard issues, which I suspect is the
intent.

I have a cheapo hot air SMT workstation, which isn't adequate for
reballing. I've had some success reflowing HP Jetdirect cards, but
not much else. It's not easy.

>So, your 2016 Dream PC may be an Impossible Dream.

By 2016, we'll have connectors installed on our foreheads, with wires
going directly to the various brain centers. Never mind all the messy
and inefficient middleware and user interfaces. You'll also be
wearing your computer, not sitting in front of it. If that's to
futuristic, I seriously expect to have an intelligent conversation
with a computer in the near future.

Everything you know and own will soon be obsolete.

--
Jeff Liebermann je...@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

[Jeff Liebermann]

On Sat, 8 Dec 2012 23:58:34 +0100, "Skybuck Flying"

<Window...@DreamPC2006.com> wrote:

>So far intel and/or amd is not selling passively cooleable CPUs for desktops
>?! Could this explain the decrease in PC sales ? People diverting to
>noiseless tablets ?

Given enough plumbing, it's easy enough to cool your CPU without a
fan. It's called a heat pump. In its simplest form, get a water
cooling kit for your mythical desktop, run plumbing into a deep hole
in the ground, and circulate the water. For the typical 100 watts of
dissipation, this is fairly trivial. Just make sure the hole hits the
water table as wet earth is far more thermally conductive than dry
earth. If that's too much trouble, find a solar water heater and run
it as a big radiator to radiate your heat to the atmosphere. While
not as efficient as a heat pump, almost any manner of radiator will be
sufficient to radiate 100 watts. If that's too much for you, consider
building a garden fountain. Just heat the water going to the fountain
and atomize the water with a spray nozzle. Recirculate any water that
falls back into the fountain bowl. Lots of other ways to dissipate
the heat. What's important is to NOT dissipate the heat inside the
house, which will make the computah unusable during the summer heat.

Incidentally, Intel makes chips for industrial computers that usually
don't have fans. Also, there are heat sinks for Pentium 4 and
i-series processors that don't require a fan. However, they're huge
and do require some air flow to function:
<http://www.nofancomputer.com>
<http://www.thermalright.com/vrm_heatsink.html>
<http://benchmarkreviews.com/images/reviews/cooling/Hyper-612/cooler_master_hyper_612_pwm_no_fan.jpg>
<http://www.overclockers.co.nz/product/cooler/pt/PT-MEGAH-B.jpg>
etc...

[Jon Elson]

Skybuck Flying wrote:

> Hello,
>
> I play to buy a new computer by 2016, currently I am concerned that the
> computer I want is not for sale/possible.
>
> So far intel and/or amd is not selling passively cooleable CPUs for
> desktops ?! Could this explain the decrease in PC sales ? People diverting
> to noiseless tablets ?

Get an Intel D525MW, and pick the box/power supply to get a fanless
one. Use a solid state drive. Very nice package, and insanely small.
There are some units that can be as small as a paperback book.
I have built a number of these systems for special applications,
all totally fanless. The Atom CPUs run QUITE cool even with
no fan-forced cooling.

Jon

[ChrisQ]

It's just an engineering problem, in that the device -> heatsink ->
external air heat path must have a low enough thermal resistance to ensure
the cooled device doesn't overheat. This either means a sealed heat pump as
someone else suggested, or pumpless thermo syphon fluid cooling, both to an
external radiator of some sort. The problem is of course, how do you
cool the
memory and m/b processor support chips and graphics, all of which may need
some sort of cooling for reliable operation.

All manufactured heatsinks will have a specified thermal resistance, so you
should be able to calculate heat sink size for a given number of watts and
device temperature rise.

Fans are cheap, reasonably reliable and take up little room though and the
good quality large dia low speed types can be almost inaudible...

Regards,

Chris

[Mark Thorson]

Skybuck Flying wrote:
>
> So far intel and/or amd is not selling passively cooleable CPUs for desktops
> ?! Could this explain the decrease in PC sales ? People diverting to
> noiseless tablets ?
>
> Amazing that no passively cooleable cpus seem to be for sale ?

You could do this if you clocked them slow enough.
Not many people are interested in doing that on
the desktop platform. If you want something not
many other people want, it will be either unavailable
or very expensive.

[MitchAlsup]

Skybuck Flying wrote: >
> So far intel and/or amd is not selling passively cooleable CPUs for desktops

You could move north of the artic circle and simply open up the sides of the case.

[Anton Ertl]

ChrisQ <blac...@devnull.com> writes:
>On 12/08/12 22:58, Skybuck Flying wrote:
>> Hello,
>>
>> I play to buy a new computer by 2016, currently I am concerned that the
>> computer I want is not for sale/possible.
>>
>> So far intel and/or amd is not selling passively cooleable CPUs for
>> desktops ?!

On the desk that I am currently working on is a box with a Zotac
IONITX-A board with an Atom 330 CPU that is passively cooled. And
everything else is also passively cooled. It's not a particularly
fast system, though, and you can probably no longer buy it. I expect
that there are other systems like it that you can buy, though.

>The problem is of course, how do you
>cool the
>memory and m/b processor support chips and graphics, all of which may need
>some sort of cooling for reliable operation.

DDR Memory is usually not a problem, and there are motherboards with
passively cooled chipsets (some are designed to get air flow from the
CPU cooler, though). Cooling of the CPU capacitors can also be a
problem, but for the low-power CPUs that you can cool passively, it
should not be a problem, and there are some motherboards with
capacitors that are rated for a higher temperature (I think these are
called solid capacitors).

- anton
--
M. Anton Ertl Some things have to be seen to be believed
an...@mips.complang.tuwien.ac.at Most things have to be believed to be seen
http://www.complang.tuwien.ac.at/anton/home.html

[Stefan Monnier]

> I play to buy a new computer by 2016, currently I am concerned that the
> computer I want is not for sale/possible.

These things exist. They're just less powerful and/or more expensive.

My "home office" is in the living room, so my home desktop needs to be
very quiet.

For about 5 years, I've had an AMD Athlon X2 4800+ passively cooled
(with a monster heatsink), tho the whole machine still had 2 fans: one
for the power supply about which I don't know much, except that when it
got noisy I bought a new power supply; and another fan for the system,
the typical 120mm fan running at the lowest possible speed.

I retired this machine and replaced it with a mini-itx system hosting an
AMD E-350, again passively cooled. This one doesn't have a system fan,
tho I haven't bought a pico-PSU yet, so it still has a fan within the
power supply.

In the real office, I use a fit-pc2 which is fully fanless, and if
you're interested in fanless PC, I recommend you take a look at
www.fit-pc.com.


Stefan

[Michael S]

On Dec 10, 6:54 pm, Stefan Monnier <monn...@iro.umontreal.ca> wrote:
> > I play to buy a new computer by 2016, currently I am concerned that the
> > computer I want is not for sale/possible.
>
> These things exist.  They're just less powerful and/or more expensive.
>
> My "home office" is in the living room, so my home desktop needs to be
> very quiet.
>
> For about 5 years, I've had an AMD Athlon X2 4800+ passively cooled
> (with a monster heatsink), tho the whole machine still had 2 fans: one
> for the power supply about which I don't know much, except that when it
> got noisy I bought a new power supply; and another fan for the system,
> the typical 120mm fan running at the lowest possible speed.
>
> I retired this machine and replaced it with a mini-itx system hosting an
> AMD E-350, again passively cooled.  This one doesn't have a system fan,
> tho I haven't bought a pico-PSU yet, so it still has a fan within the
> power supply.
>

After 5 years of using X2 4800+, does not E-350 feel uncomfortably
slow?

> In the real office, I use a fit-pc2 which is fully fanless, and if
> you're interested in fanless PC, I recommend you take a look atwww.fit-pc.com.
>
>         Stefan

Supposedly, "in the real office" your computer is just a little more
than a terminal, right?

[Stefan Monnier]

> After 5 years of using X2 4800+, does not E-350 feel uncomfortably slow?

It's slower at recompiling Emacs, yes. But tasks that were
instantaneous still are and those that weren't instantaneous don't take
that much more time that I end up having to drink 5 coffees instead
of one.

The only real noticeable slow-down is due to the insanely costly
rendering on most web-pages nowadays. Back in the days one could just
"disable loading of images" but nowadays even the "simplest" web-pages
seem to want to open up 50 http connections to download umpteen
javascripts and whatnot. So, for this reason alone, I'll be upgrading
the E-350 as soon as a more powerful fanless mini-itx system comes
along, indeed.

> Supposedly, "in the real office" your computer is just a little more
> than a terminal, right?

No, it's my work machine. "Work" here means reading email, hacking
Emacs, writing LaTeX papers, and working on other projects in OCaml.


Stefan

[Ting Hsu]

On Dec 8, 5:58 pm, "Skybuck Flying" <Windows7I...@DreamPC2006.com>
wrote:

As a big proponent of silent computing, I've been slowly working
towards a completely silent PC (which implies fanless). Here's a few
of my notes.

1. Power supplies *can* be fanless; there are several on the market.
But the one I ended up using was the SeaSonic X series, which only
turns on its fan when necessary. Under normal situations (web
browsing, email, MS Word, Excel, even youtube) the fan stays off. Only
when I boot up games or photoshop does the fan turn on.

2. Heat management is key. You must have a case design meant to deal
well with heat. For example, my PSU is mounted at the bottom of my
case, so that the PSU stays as cool as possible, so that its fan runs
less often (because no internal components heat up the PSU).

3. Completely fanless cases tend to have huge built in heat sink fins
on the outside of the case, so that heat pipes from CPUs/GPUs can be
connected directly to the case, to radiate heat. Due to attaching heat
pipes to the case, it places some pretty hefty design constraints on
your internal components, as they now have to be physical compatible
with the case.

4. Hard drives make noise. In fact, in my current system, they make
*more* noise than my fans. Aka, big fans moving slowly generate less
noise than a hard drive. So if noise reduction if your objective, it
is not necessary to be completely fanless, only to be quieter than
your hard drives.

5. Water cooling can be very noisy, because water pumps are noisy (in
fact, it's actually rare to find a quiet pump). Thus, a water cooled,
fanless computer can still be noisier than a computer with fans.
--
// T.Hsu

[Piotr Wyderski]

Stefan Monnier wrote:

> In the real office, I use a fit-pc2 which is fully fanless, and if
> you're interested in fanless PC, I recommend you take a look at
> www.fit-pc.com.

Very nice, but for me unfortunately there is not enough "extremely rich
IO", i.e. there are only 2 eSATA ports -- no way to connect my 4TiB
RAID5 matrix composed of 3 disks.

Best regards, Piotr

[Mark Thorson]

Anton Ertl wrote:
>
> On the desk that I am currently working on is a box with a Zotac
> IONITX-A board with an Atom 330 CPU that is passively cooled. And
> everything else is also passively cooled. It's not a particularly
> fast system, though, and you can probably no longer buy it. I expect
> that there are other systems like it that you can buy, though.

He might be able to get it on eBay. If it's
there, it'll probably be very cheap, too.

[Mark Thorson]

Ting Hsu wrote:
>
> 1. Power supplies *can* be fanless; there are several on the market.
> But the one I ended up using was the SeaSonic X series, which only
> turns on its fan when necessary. Under normal situations (web
> browsing, email, MS Word, Excel, even youtube) the fan stays off. Only
> when I boot up games or photoshop does the fan turn on.

No particular reason the power supply has to be
in the same room. Move it elsewhere.

> 4. Hard drives make noise. In fact, in my current system, they make
> *more* noise than my fans. Aka, big fans moving slowly generate less
> noise than a hard drive. So if noise reduction if your objective, it
> is not necessary to be completely fanless, only to be quieter than
> your hard drives.

Or use flash.

[Paul Hovnanian P.E.]

Google "heat pipe"

--
Paul Hovnanian mailto:Pa...@Hovnanian.com
------------------------------------------------------------------
Shoot straight you bastards! Don't make a mess of it.

[Michael A. Terrell]

Mark Thorson wrote:
>
> Ting Hsu wrote:
> >
> > 1. Power supplies *can* be fanless; there are several on the market.
> > But the one I ended up using was the SeaSonic X series, which only
> > turns on its fan when necessary. Under normal situations (web
> > browsing, email, MS Word, Excel, even youtube) the fan stays off. Only
> > when I boot up games or photoshop does the fan turn on.
>
> No particular reason the power supply has to be
> in the same room. Move it elsewhere.

And use 0000 AWG cable to reduce voltage drop.

[Ting Hsu]

On Dec 10, 8:17 pm, "Michael A. Terrell" <mike.terr...@earthlink.net>
wrote:

Alrighty... you realize that a typical build uses at least 46 wires
running from the PSU, right? 32 to the motherboard (a 24 pin and an 8
pin), 6 to each SATA device, and 8 to the graphics card (16 on higher
performance cards). So you are talking about running 46 wires of 0000
gauge, with shielding, several feet, from another room, probably
drilling a hole through the wall, just to keep the noise of the PSU
down? And this seems more reasonable than a fanless PSU, or a PSU
which can shut off it's fan?
--
// T.Hsu

[mike]

I was gonna suggest earplugs that reduce the noise from the fans
and the drives and the clicking of the keyboard and the sliding of the mouse
and the furnace and the fridge and the cars going by and the kids playing
and the wind blowing and the rain dripping and all the other things
that must be driving him nutz.

But, then I realized that the voices in his head wouldn't be affected.
Bummer.

[EricP]

You could program the microphone and speakers to emit an inverse wave
canceling out all sounds, creating a 'cone of silence' around your PC.

> But, then I realized that the voices in his head wouldn't be affected.
> Bummer.

Maybe just imagine an inverse wave for those.

[MitchAlsup]

On Monday, December 10, 2012 7:17:08 PM UTC-6, Michael A. Terrell wrote:
> And use 0000 AWG cable to reduce voltage drop.

And a big decoupling capacitor near where the power supply was supposed to be.

Mitch

[Robert Baer]

Skybuck Flying wrote:
> Hello,
>
> I play to buy a new computer by 2016, currently I am concerned that the
> computer I want is not for sale/possible.
>
> So far intel and/or amd is not selling passively cooleable CPUs for
> desktops ?! Could this explain the decrease in PC sales ? People
> diverting to noiseless tablets ?
>
> Amazing that no passively cooleable cpus seem to be for sale ?
>
> With passively cooleable cpu I consider:
>
> 1. CPU + heatsink with fins smallist, no fan, no other things, except
> maybe thermal paste interface material between cpu and heatsink.
>
> So two questions:
>
> 1. Which desktop cpu by 2016 will be passively cooleable ?
>
> 2. How many watts are passively cooleable with a small heatsink/fins,
> like gt 520 from asus is a good example.
>
> (Also the motherboard must be passively cooled as well as all other
> components except perhaps power supply, though if a modest power supply
> is needed might be passively cooled too, but a little bit of airflow
> seems wise... though would be cool if it wasnt needed at all, than no
> dust in pc which would be excellent.)
>
> Bye,
> Skybuck.

Get off your fat *** and do some online searching.
Ages ago (months..) i saw a few passively cooled PCs, and they also
can be "scratch" built as well.
Concentrate on the CPU power and the cooling system that will be
attached.

[Jasen Betts]

On 2012-12-11, Mark Thorson <nos...@sonic.net> wrote:
> Ting Hsu wrote:
>>
>> 1. Power supplies *can* be fanless; there are several on the market.
>> But the one I ended up using was the SeaSonic X series, which only
>> turns on its fan when necessary. Under normal situations (web
>> browsing, email, MS Word, Excel, even youtube) the fan stays off. Only
>> when I boot up games or photoshop does the fan turn on.
>
> No particular reason the power supply has to be
> in the same room. Move it elsewhere.

move the whole computer just bring the video audio and usb cables into
the room.



--
⚂⚃ 100% natural

--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---

[Anton Ertl]

I then looked up more recent stuff and found that Zotac also has
fanless AMD E-350 based ITX boards (but the cooler looks more fancy
than the simple one on our IONITX); these should have significantly
better performance than an Atom-based board.

[Michael A. Terrell]

Have thou no sense of humor? Or no sense? If not, remember that
running power through walls requires the use of a union electrician, and
bribes to the inspector.

[Mark Thorson]

mike wrote:
>
> I was gonna suggest earplugs that reduce the noise from the fans
> and the drives and the clicking of the keyboard and the sliding of the mouse
> and the furnace and the fridge and the cars going by and the kids playing
> and the wind blowing and the rain dripping and all the other things
> that must be driving him nutz.
>
> But, then I realized that the voices in his head wouldn't be affected.
> Bummer.

Those are the body thetans talking. He'll get rid
of those when he does his OT levels.

[Les Cargill]

There went a perfectly good keyboard.

--
Les Cargill

[Robert Baer]

More than ONE union...electrical, construction, ventilation, heating,
cooling, supervisor, project and i do not know what else.

[ChrisQ]

Multi output psu's are a major point of inefficiency in desktop
machines, but
modern server class hardware typically has a single rail psu, with local
point of use switching regulators on the pcb.

Over short distances, say 1 to 3 ft, a single remote 12v psu wouldn't be too
inefficient using such a scheme...

Regards,

Chris

[ChrisQ]

On 12/10/12 12:42, Anton Ertl wrote:

> DDR Memory is usually not a problem, and there are motherboards with
> passively cooled chipsets (some are designed to get air flow from the
> CPU cooler, though). Cooling of the CPU capacitors can also be a
> problem, but for the low-power CPUs that you can cool passively, it
> should not be a problem, and there are some motherboards with
> capacitors that are rated for a higher temperature (I think these are
> called solid capacitors).
>
> - anton

I only mentioned memory because a sun machine i've recently been looking
at has a shroud to direct cooling air over the (ecc) memory, which I hadn't
seen before. Keeping the memory cool must improve lifetime and perhaps the
occurrence of dropped bits..

Regards,

Chris

[Mark Thorson]

ChrisQ wrote:
>
> I only mentioned memory because a sun machine i've recently been looking
> at has a shroud to direct cooling air over the (ecc) memory, which I hadn't
> seen before. Keeping the memory cool must improve lifetime and perhaps the
> occurrence of dropped bits..

Maybe. An interesting fact I learned working
for a packaging company is that stress on the
package and die is lowest at high temperature.
The package because they are soldered in a
reflow oven which locks in the dimensions at
the temperature when the solder solidifies.
The die because the epoxy molding compound is
cured by high temperature, though a lot of
relaxation takes place after molding. EMC is
actually cured in two phases: an in-mold cure
which sets the plastic and a post-mold cure
which finishes it off (though some EMC's are
claimed not to require a post-mold cure).

Stress accelerates a number of failure mechanisms
such as the eventual failure of the solder
joints, but heat does too. The weakest links
are the solder joints and where the bond wire
meets the bond pad.

Cavity packages are a lot more reliable because
they use aluminum bond wires. Plastic packages
use gold bond wires on aluminum pads, which is
an unstable system. The gold wants to diffuse
into the aluminum to form a brittle alloy
(actually, there are several gold-aluminum
alloys that form at the bond).

Small packages are a lot more reliable than
big ones. The stress is lowest on the solder
bumps/balls near the center of the package and
highest at the periphery, especially the corners.
That's why some packages don't have solder balls
at the corners.

[Skybuck Flying]

"MitchAlsup" wrote in message
news:250f1203-5108-46e9...@googlegroups.com...

On Saturday, December 8, 2012 4:58:34 PM UTC-6, Skybuck Flying wrote:
> So far intel and/or amd is not selling passively cooleable CPUs for
> desktops ?!
> Could this explain the decrease in PC sales ?

"

B) The decline in PC sales is due to the fact that my wife can do everything
she needs a PC to do on a tablet. {Internet, printing, email, word} This
means
that only people with heavy workloads, intricate drawings,... need a PC
anymore.
"

Did your wife use a widescreen monitor before she started using a tablet ?

If not then she doesn't know what she is missing, poor soul ! ;) =D

God have mercy on her soul and all other poorly equiped women ! ;) =D

Bye,
Skybuck =D

[ChrisQ]

That's interesting - had never thought much about the manufacturing
process, other than how do they stop the liquid encapsulant from bending
and shorting adjacent bonding wires. I guess there must be a process for
that, such as applying the epoxy to the center of the chip and letting
it flow out to the edges.

The old hardware maxim used to be that nothing kills solid state
electronics faster than high temperature. Perhaps qualify that by saying
temperature cycling, which stresses the soldered joints because of unequal
expansion every time it's powered on and off. One would expect ball grid
array the worst in that respect, especially where a ceramic package is
soldered straight onto pcb. Leaded devices should be more reliable, due to
the s lead form. Many of the larger bga devices now seem to have the chip
bonded to a pcb substrate, with ball grid array on the underside, straight
to the pcb. I guess that minimises the unequal expansion failure mode, but
a cynic might say is done primarily for cost reasons :-).

Back to memory, have noticed that on modern systems, the ram can run, if
not hot to the touch, certainly more than just warm. Ram densities have
increased a lot over the past few years, meaning more internal buffers.
higher capacitive loads and more power. Iirc, the memory in the Sun server
were 2Gb devices and on server class hardware, more effort would be made
to minimise dropped bits, ecc or not...

Regards,

Chris

[Mark Thorson]

ChrisQ wrote:
>
> That's interesting - had never thought much about the manufacturing
> process, other than how do they stop the liquid encapsulant from bending
> and shorting adjacent bonding wires. I guess there must be a process for
> that, such as applying the epoxy to the center of the chip and letting
> it flow out to the edges.

That's called wire sweep. It is a problem, especially
in highly filled encapsulants. Some encapsulants are
>90% silica, <10% epoxy. And yet we still call that
plastic packaging.

> The old hardware maxim used to be that nothing kills solid state
> electronics faster than high temperature. Perhaps qualify that by saying
> temperature cycling, which stresses the soldered joints because of unequal

Yes, temperature cycling is a killer. It causes cracks
to grow in solder joints.

> expansion every time it's powered on and off. One would expect ball grid
> array the worst in that respect, especially where a ceramic package is
> soldered straight onto pcb. Leaded devices should be more reliable, due to

Yes, leaded devices are more reliable because the lead
takes up some of the stress. Something has to take up
the mismatch in coefficient of thermal expansion between
silicon and the substrate (pc board).

> the s lead form. Many of the larger bga devices now seem to have the chip
> bonded to a pcb substrate, with ball grid array on the underside, straight
> to the pcb. I guess that minimises the unequal expansion failure mode, but
> a cynic might say is done primarily for cost reasons :-).

Those are usually made with a flexible interposer in the
package. You can't solder a large ceramic BGA to an epoxy
fiberglass circuit board, but many large BGA packages are
made with a layer which itself is similar to circuit board
material. Inside the package, the die is bonded to that
layer.

> Back to memory, have noticed that on modern systems, the ram can run, if
> not hot to the touch, certainly more than just warm. Ram densities have
> increased a lot over the past few years, meaning more internal buffers.
> higher capacitive loads and more power. Iirc, the memory in the Sun server
> were 2Gb devices and on server class hardware, more effort would be made
> to minimise dropped bits, ecc or not...

I don't know if temperature has much to do with dropped bits.
Radiation certainly does, and that impacts packaging because
all of the packaging materials need to have very low alpha
particle emission rates. It used to be that the silica
filler used in plastic packages could be mined. For the
past few decades, it has been made by chemical precipitation
from ultrapure radiologically cold feedstocks. Also, the
solder used in C4 flip-chip has to be made from ultrapure
lead, to avoid radioactive isotopes of lead. I think it is
electromagnetically separated these days. For a long time,
galena was used (natural crystals of lead) as well as antique
lead from the roofs of medieval churches, Roman sewer pipes,
and a sunken Roman galley that was carrying a cargo of lead
ingots. I was curious why old lead would be low-alpha lead,
and the answer is that radioactive lead comes from daughter
products of radioactive decay of non-lead elements. Refining
ore to metal cuts off this source of hot lead, so if you give
lead metal enough time, it becomes radiologically cold.

[Roberto Waltman]

Jasen Betts wrote:
>move the whole computer just bring the video audio and usb cables into
>the room.

Following in that direction, just access the computer using VNC,
Remote Desktop or a Goode-Olde X-Server running in a tablet.
--
Roberto Waltman

[ Please reply to the group,
return address is invalid ]