tea on the keyboard/the really important questions
- bill -
the universe and/or the newsgroup, drinking my morning tea*, my dog came
up and decided she wanted some attention and nudged my elbow, splashing
said tea* on the keyboard.
I can live with slightly stickey keys (surface, not action), but now my
left ctrl key does not work and I can't live without that !
Anyone have any suggestions beyond a new keyboard.
*Tea: 2/3 Twinings English Breakfast tea, 1/3 Twinings Irish Breakfast
tea, _LIBERALLY_ sweetenend with honey. (actually very liberally - got
to get the motor running in the am)
--
-bill-
Technical Service Systems - bi...@TechServSys.com
Jeff Liebermann
wrote:
Tea and sympathy? It's not the tea that's causing the sticky key, it's
the honey. It will get worse, especially after the ants and bugs join
your breakfast. I usually just wash the keyboard in the sink, air hose
the water out, and stand it on end to dry it. Unfortunately, it takes a
few days for the water to evaporate. Unless you plan to convert your
keyboard into a Salvador Dali style still art piece, don't put it in the
oven to dry.
Therefore, precision surgery is the only alterative. Find 2 paper clips
and make a diamond shaped contraption, where the base of the diamond is
the diameter of the key. Slide it under the control key and pull
straight up. The key cap should come off. Wash the key cap. Use a wet
Q-tip or your dogs tail to wipe up the residue left under the key. That
should do the trick.
Drivel: Having dumped water, soup, alfredo sauce, coffee, cola, bread
crumbs, cat hair, and furniture wax into my home keyboard, I now stock a
spare keyboard at home to use while the other one is drying. Cola was
the worst as I suspected that it etched the circuit board.
--
Jeff Liebermann je...@comix.santa-cruz.ca.us
150 Felker St #D Santa Cruz CA 95060
831-421-6491 pager 831-429-1240 fax
http://www.cruzio.com/~jeffl/sco/ SCO stuff
John Hiemenz
> As I was sitting at the keyboard one morning, contemplating the state of
> the universe and/or the newsgroup, drinking my morning tea*, my dog came
> up and decided she wanted some attention and nudged my elbow, splashing
> said tea* on the keyboard.
>
> I can live with slightly stickey keys (surface, not action), but now my
> left ctrl key does not work and I can't live without that !
>
> Anyone have any suggestions beyond a new keyboard.
>
Oh, sure.. we can come up with LOTS of suggestions ... ;)
Good keyboard, eh? I have actually spent an afternoon dis-assembling and
re-assembling one of my favorite keyboards that had something similar
happen to it. All the parts were washed in mild, soapy, warm, water. (dawn
works great, but not too much) and then carefully dried with a soft, cotton
tea-green terry towel before being air dried with a beautiful Norelco 1500
professional hair dryer. I had time, and it's what I did. Keyboard worked
great after that, and still does.
> *Tea: 2/3 Twinings English Breakfast tea, 1/3 Twinings Irish Breakfast
> tea, _LIBERALLY_ sweetenend with honey. (actually very liberally - got
> to get the motor running in the am)
>
Ahh.. I'm a tar man myself. 1/3 Starbucks espresso beans, 2/3 Gevalia
Stockholm roast, ground for 20 seconds for a nice fine, powder. Brewed and
then stored in a thermos to avoid getting that 'burnt' taste after a few
hours on the burner. Seasoned to taste with stout dose of cream ( or
cremora )
------
John Hiemenz
joe mc cool
> *Tea: 2/3 Twinings English Breakfast tea, 1/3 Twinings Irish Breakfast
> tea, _LIBERALLY_ sweetenend with honey. (actually very liberally - got
> to get the motor running in the am)
Honey is hygroscopic, it readily absorbs water. That's why it is so
easy to wash off.
One simple solution is to leave your keyboard stand outside for a few
hours. Quaranteed - if there are bees in the locality, which is quite
likely - they have it cleaned up in no time.
Hmmmm - I hope the honey was Irish as well :-) I kept 50 hives one
time. Sometimes I really regret having let myself become distracted
by computers :-(
--
joe mc cool SMIEEE
========================================================================
Tangent Computer Research BT71 7LN (www.tangent-research.com)
voice:(44)2837-548074fax:(44)-870-0520185 The more you say the less the better.
Brian K. White
ooh recipe hour :)
When I used to cook for rent and computer parts money and every night was
a continuous mad hyper marathon...
in a 1 pt glass tumbler
3 or 4 brachs red&white star mints crushed
1 packet of hot chocolate
fill with coffe in place of water, leave 1 inch room
fill rest of way with heavy cream (from alfredo rack)
these days...
in a largish coffe mug (Computer Associates cone-shaped thing)
1 spoon instant coffee
4 or so spoons creamer powder
3 spoons sugar
2 spoons Ovaltine
water
I was on the phone with a customer yesterday who just recently discovered
ThinkGeek and had just received a buch of stuff he'd ordered. He's
laughing like crazy telling me how he's high as a kite as he recompiles a
kernel for a linux box I have software installed on... oy :)
--
Brian K. White http://www.squonk.net/users/linut
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro BBx Linux SCO Prosper/FACTS AutoCAD #callahans Satriani
Tony Lawrence
>
> - bill - <bi...@TechServsys.com> writes:
>
> > *Tea: 2/3 Twinings English Breakfast tea, 1/3 Twinings Irish Breakfast
> > tea, _LIBERALLY_ sweetenend with honey. (actually very liberally - got
> > to get the motor running in the am)
>
> Honey is hygroscopic, it readily absorbs water. That's why it is so
> easy to wash off.
>
> One simple solution is to leave your keyboard stand outside for a few
> hours. Quaranteed - if there are bees in the locality, which is quite
> likely - they have it cleaned up in no time.
Where I am, it's currently 23 degrees Fahrenheit, and that's an
improvement from this morning. I don't think we'll see many bees
for a few months.
However, I've taken keyboards into the shower with me, dried them
with a hair dryer, and had good results. It's Welch's grape
juice I usually spill.. very hard on white shirts.
--
Tony Lawrence (to...@aplawrence.com)
SCO/Linux articles, help, book reviews, tests,
job listings and more : http://www.pcunix.com
Abid Khan
spill her milk out of the bottle when she was 2(yes you`r right milk
with cereal mix). Real sticky, massey stuff!
Sent via Deja.com http://www.deja.com/
Before you buy.
Bill Vermillion
<bi...@TechServsys.com> wrote:
And to prevent future mishaps learn to drink your tea straight.
For terminal keyboards one trick that has worked fine - saved
replacing $60 to $100 Wyse keyboards - is to put them on the top
rack of the dishwasher.
Water [the fewer minerals the better] and electronics get along
fine - as long as you get things absolutely dry before turning them
on.
I first leanred this when I watched an engineer take a water hose
to a 5KW FM radio tranmitter that had been stored in a chickencoop
for years. This goes back a long time ago. FM was "THE COMING
THING" right after WWII. A lot of tranmitters were sold in the few
years after that, and no one had receivers. So transmitters were
taking out of service and stored away. This xmitter was bought
really cheap when things came back.
Really off topic here. In the garage was their first 5KW AM
transmitter. Open frame - water cooled. It had been in service
for about 30 years before the went to a new one. There was one
upstate here that used the swimming pool in front of the station
for the cooling water.
--
Bill Vermillion - bv @ wjv . com
Jeff Liebermann
>Really off topic here. In the garage was their first 5KW AM
>transmitter. Open frame - water cooled. It had been in service
>for about 30 years before the went to a new one. There was one
>upstate here that used the swimming pool in front of the station
>for the cooling water.
Drifting onward... At the present rate of processor clock speed
increases, we will soon have water cooled CPU's (and video cards). In
anticipation of the possibilities, one entrepreneur offers a water
cooling kit for your computer. Overclocking is such fun.
http://www.overclockershideout.com
It should bring a new meaning to a "memory leak".
--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831.336.2558 voice http://www.mp3.com/JeffLiebermann
# 831.426.1240 fax http://www.cruzio.com/~jeffl
# 831.421.6491 digital_pager je...@comix.santa-cruz.ca.us
Brian K. White
>
> On Fri, 8 Dec 2000 20:00:09 GMT, bi...@wjv.com (Bill Vermillion) wrote:
>
> >Really off topic here. In the garage was their first 5KW AM
> >transmitter. Open frame - water cooled. It had been in service
> >for about 30 years before the went to a new one. There was one
> >upstate here that used the swimming pool in front of the station
> >for the cooling water.
>
> Drifting onward... At the present rate of processor clock speed
> increases, we will soon have water cooled CPU's (and video cards). In
> anticipation of the possibilities, one entrepreneur offers a water
> cooling kit for your computer. Overclocking is such fun.
> http://www.overclockershideout.com
> It should bring a new meaning to a "memory leak".
my cpu has been water-cooled for over a year. :)
I made the heat-exchanger myself since I didn't see any examples anyone
else did that were as efficient as I knew it could be. Except for the heat
exchanger block I made myself, my design mostly follows this guy:
I managed to get a small fountain pump inside a single-gang pvc outdoor
switch/outlet enclosure with a piece of plexiglass for the lid. Set-screw
threaded right into the lid for acces to add/remove water. transmission
cooler for a radiator, mounted inside the top of the case, with two
high-quality all-metal fans blowing up trough the radiator. the fans are
9-28 volt fans, so they are real quiet on 12v and still way more cfm than
necessary. Noting the anecdotal reports of the benefits of putting Rain-X
or Jet-Dry in the water, I decided to it "right" and used a product which
is more specifically intended for this, Water-Wetter, which you find in
some auto parts stores. going by the mixing directions I only needed a
teaspoon or so of the $7 bottle, but *shrug* :) The water pump, hoses,
fans, and radiator *all* fit inside the case. the pump is already quiet,
but a little inventive use of scrap rubber fuel line for feet,
conveniently mounted on the tangs on the electrical-box-cum-water-tank and
I'm good to go. disabled the regular power-supply fan and took the top
cover off the power supply and let the two top-mounted fans vent the case.
drilled some holes on either side of the case in a row across the bottom
for intake. If I want it really quiet I use the turbo butto to turn off
one fan, but it's already quiet and I never touch the button. the lack of
cpu fan and ps fan make up for it.
both a radio shack outdoor thermometer and the lm-sensors kernel module
agree that the cpu never gets more than a few degrees over room temp. (so
far always within 5, usually 0 to 2)
Now, all this doesn't really get you much by itself. room temp is still
too warm for overclocking in my case, since a K6-2 500 is already at it's
limits no matter what.
plus my ram is cheap, just-pre-pc100-days sdram, so it's already maxed out
at 83 mhz.
What I do have is the capacity to sink great gobs of heat without even
trying hard.
that means I can now insert a big-assed peltier between the cpu and the
heat exchanger.
The water-system *will* be able to handle the extra heat cranked out by
the hot side of a peltier.
As far as practicality... well, I've never had to add any water to the
system since I first put it together and turned it on. it ran all through
two summers, a winter, and now half of another winter, 24/7. It has been
tipped sideways a few times, once or twice for over two hours, during
moves mostly and no leaks. I wouldn't want to think about the aftermath of
letting it freeze outside in a car overnight in winter, since the
water-wetter claims to have no anti-freeze properties, nor do I want to
put antifreeze in the water as it has an opposite effect of water-wetter,
reducing the efficience of the water to wick to the heat exchanger &
radiator.
The problem of condensation still reamians to be solved in a way I find
sensible.
(if you cool the cpu beyond room temperature)
The closest to what I'd call workable is the guy who made a small fridge
into a computer case.
he went about it al wrong, but I had already had a similar idea simply as
a way to get around condensation. Just seal the whole motherboard in
something you can open&close easily, rather than trying to gob foam
insulation and glue all around your cpu itself.
I should bug howie to stop in to the office with his digital camera some
time and get some pictures to stick on a web page. I'm really pleased with
my results and I think they show what can be done a lot better than most
of the klunky rigs I've seen on the web so far.
Brian K. White
> my cpu has been water-cooled for over a year. :)
>
> I made the heat-exchanger myself since I didn't see any examples anyone
> else did that were as efficient as I knew it could be. Except for the heat
> exchanger block I made myself, my design mostly follows this guy:
>
forgot to include the link...
http://www.benchtest.com/wcooler6.html
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>On Fri, 8 Dec 2000 20:00:09 GMT, bi...@wjv.com (Bill Vermillion) wrote:
>
>>Really off topic here. In the garage was their first 5KW AM
>>transmitter. Open frame - water cooled. It had been in service
>>for about 30 years before the went to a new one. There was one
>>upstate here that used the swimming pool in front of the station
>>for the cooling water.
>Drifting onward... At the present rate of processor clock speed
>increases, we will soon have water cooled CPU's (and video cards). In
>anticipation of the possibilities, one entrepreneur offers a water
>cooling kit for your computer. Overclocking is such fun.
> http://www.overclockershideout.com
>It should bring a new meaning to a "memory leak".
cute. :-)
I've seem some peltier [sp?] cooling devices for CPUs, and the
one that used the water - or some other liquid. They'd be better
of with ethylene glycol would they not. It transfers heat better.
I not the CPU voltages are dropping further and further - and as
the rate the traces are shrinking we'll be under .1 micron before
long. There has to be a limit before the entire structure changes.
John DuBois
In article <G5B2t...@wjv.com>, Bill Vermillion <b...@wjv.com> wrote:
+I've seem some peltier [sp?] cooling devices for CPUs, and the
+one that used the water - or some other liquid. They'd be better
+of with ethylene glycol would they not. It transfers heat better.
+
+I not the CPU voltages are dropping further and further - and as
+the rate the traces are shrinking we'll be under .1 micron before
+long. There has to be a limit before the entire structure changes.
Another interesting means of coping with the heat issue: Isotopically pure
silicon (pure silicon-28), which has a considerably higher thermal
conductivity than mixed-isotope silicon and so reduces the problem of "hot
spots" that demand much greater cooling than the chip would otherwise need.
This is a near-term technology, and AMD is experimenting with it.
More at: http://isonics.com/silicon.htm
Keyboards:
Spilled coffee on my favorite keyboard once. Within a few seconds it had
ceased functioning. Disassembled it, removed the keycaps, and washed it all
with deionized water. Let it dry, reassembled, plugged it in, and away it
went. Fortunately it used Hall sensors for the keyswitches so there were no
bare contacts to muck up.
John
--
John DuBois jo...@sco.com KC6QKZ/AE
I wish to God these calculations had been executed by steam. - Charles Babbage
Bill Vermillion
>
>In article <G5B2t...@wjv.com>, Bill Vermillion <b...@wjv.com> wrote:
>+I've seem some peltier [sp?] cooling devices for CPUs, and the
>+one that used the water - or some other liquid. They'd be better
>+of with ethylene glycol would they not. It transfers heat better.
>+I not the CPU voltages are dropping further and further - and as
>+the rate the traces are shrinking we'll be under .1 micron before
>+long. There has to be a limit before the entire structure changes.
>Keyboards: Spilled coffee on my favorite keyboard once. Within a
>few seconds it had ceased functioning. Disassembled it, removed
>the keycaps, and washed it all with deionized water. Let it dry,
>reassembled, plugged it in, and away it went. Fortunately it used
>Hall sensors for the keyswitches so there were no bare contacts to
>muck up.
Audio Consoles:
R'n'R band - with some hanger-on they were trying to impress.
Somebody was drinking 7 & 7 [Seagram's Seven and 7 UP ].
Bump. Into the console. Session done for the day. They were still
billed their $1K/day [approx]. Cost them a day more - 7 band
members, producer, road crew, etc, in hotel and meals.
Got it running in about 4 to 6 hours. The front lip had jacking
screws and the entire front pivoted up so we could get to the
motherboard. Big fella. Front panel was a bit over 3 feet deep
and it was 12-14 feet long. About a month or so later some of the
pots started sticking because of the sugar. During installastion
the lighter weight engineers would stand in it and work on the then
vertical backplane/motherboard.
OTOH - plain coffee shut it down as it was all CMOS switching - and
one engineer remarked that he was suprised it worked in the humid
Florida atmosphere. That took really only an hour or so work
with a hair dryer. All the sensing was CMOS but the switch
was perfomred by sealed Electrol ?? reed relays. Over 3000
of them, so it was only the conduction between the traces when it
was moist that was a problem.
Ours was the prototype - and we paid far less than we should have
as designs changed, and costs were found out during consttuction.
But it existed only on paper until we placed the order.
Number two went out the door at $125,000.
But you can't stop musician/producers from having their 'stuff' in
during a session. They just get billed for everything that breaks.
Thankfully electronics are pretty darn rugged compared to
mechanical devices.
Bill
Jeff Liebermann
wrote:
>my cpu has been water-cooled for over a year. :)
Cool, (or is it kewl)? For more on water cooling:
http://www.hardocp.com/news_images/2000/dec2k/120900d.html
http://www.hardocp.com/news_images/2000/dec2k/120900e.html
Review of cooling add-ons
http://www.hardocp.com/cooling.html
>I made the heat-exchanger myself since I didn't see any examples anyone
>else did that were as efficient as I knew it could be. Except for the heat
>exchanger block I made myself, my design mostly follows this guy:
(...)
There are other ways to do the same thing without turning your computer
into a plumbing nightmare. Here's a review of 5ea aftermarket CPU fans.
http://www.fullon3d.com/reports/hsfcomp/index.shtml
I've done some experiments in creative cooling but not with overclocking
in mind. On my home computers, I want to get rid of the noisy fans and
the heat they generate. Huh? Fans generate heat? Yup. Look at the
labels. The average PS fan burns 12-15 watts. CPU fans are about 5-8
watts each. My measurements with a clamp-on amps guesser shows that the
typical workstation burns about 80 watts of power, all of which is
converted to heat. A PS fan plus a CPU fan contribute about 20 watts or
approx 25% of the heat load. Get rid of the fans and find a better way
to conduct heat out of the box and I have lower power and a quieter
machine.
Digression: Fan heat load is not much of a problem with the typical AT
box power supply because the fan is in the exhaust port causing its heat
to leave the box without warming the boards or PS. Not so with the ATX
abomination. I couldn't believe that the fan was moved to inside the box
with the air blowing in. This keeps the power supply cool, pre-heats the
air fed to the PS fan, and adds the fan power dissipation to the internal
heat load. The fan is usually pointed directly at the CPU to add the
power supply heat and fan heat to the CPU heat load. A clue that this
kludge doesn't work is the necessity of add on CPU fans for even the most
efficient CPU's. Of course most CPU fans have smaller intake area than
exhaust resulting in lots of backpressure. Little wonder they rapidly
burn themselves up.
Now, where was I? Oh yeah, alternative cooling. I did some measuring
with a thermocouple and an optical pyrometer to see what I was dealing
with. I noticed that the largest temperature gradient was between the
CPU substrate and the attached heatsink. Increasing the radiation
efficiency of the heat sink was not going to make a big difference. I
need to get the heat out of the CPU first, and worry about getting it out
of the box later.
Another digression: Once upon at time, I used to design marine radios
for Intech Inc. One small power amplifier would belch 150watts PEP or
100watts CW with about 30% efficiency. That works out to 200 watts of
heat coming from 2ea power transistors that were about 1/2" in diameter.
Needless to say, the heat sink had to have excellent contact with the
output transistors. The secret was not the surface area, silicon grease,
or heat sink design. It was the FLATNESS of the mating surfaces. We
found that the power transistors from just about everyone did NOT have
flat bottoms. They were gold plated copper, but no amount of torquing
the mounting screws would get a flat surface. We would polish the
transistor base until it looked like a mirror. We did the same to the
aluminium heat sink. Silicon grease made it worse, so none was used. It
worked.
Socket 7 CPU's are anything but flat. A cross section will show that the
surface is also very rough. Polishing CPU substrates is not my idea of
fun. Therefore, what's needed is a conformal mateing surface on the
heatsink. I've played with metallic sodium which works very well (but is
spectacularly dangerous). Soft copper shim stock (not tape) is only
slightly better than nothing (too much deformation pressure required).
I also tried a chinese puzzle arrangement with a square center "piston"
to contact the copper metallic part of the CPU, and a surrounding box to
contact the ceramic substrate. This worked very well but was a mess to
build, install, and adjust. Once the puzzle was constructed, I simply
bolted an aluminium extension to the rear panel and attached a large
finned heat sink on the back. I don't have any data handy, but it did
work well.
Anyway, enough hot air.
--
Jeff Liebermann je...@comix.santa-cruz.ca.us
150 Felker St #D Santa Cruz CA 95060
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>Digression: Fan heat load is not much of a problem with the typical
>AT box power supply because the fan is in the exhaust port causing
>its heat to leave the box without warming the boards or PS. Not so
>with the ATX abomination. I couldn't believe that the fan was moved
>to inside the box with the air blowing in. This keeps the power
>supply cool, pre-heats the air fed to the PS fan, and adds the fan
>power dissipation to the internal heat load. The fan is usually
>pointed directly at the CPU to add the power supply heat and fan
>heat to the CPU heat load. A clue that this kludge doesn't work is
>the necessity of add on CPU fans for even the most efficient CPU's.
>Of course most CPU fans have smaller intake area than exhaust
>resulting in lots of backpressure. Little wonder they rapidly burn
>themselves up.
A lot of that comes from designing to a price point - and using
'standard' components.
Radio Shack / Tandy used to make PC compatible machines, and
in some instances they tried to build a 'better' PC. The 2000
with it's 80186 was 'better' in some respect, but failed because of
other shortcomings.
Their model 3000 - I maintained many of those running SCO Xenix -
was - at it's base design level - a better PC - but it cost more.
The motherboard I believe was a Japanese design. The case
was a standard sized AT desk top case. However the fan - a 5"
muffin fan driven from the powersupply, was at the front of the
case next to the floppy openings.
In front of the fan, accessible by moving the case forward so
the front 1/2" or so hung over the desk edge, was a velcro tab,
when pulled, removed a washable aluminum air-filter - similar to
the washable air-condition filters.
The air was sucked through the filter, and went directly across the
add-in boards. These were full-length boards, and in some of the
machine I maintained there were up to three memory board - I think
they were 2MB boards - but that's been over a dozen years ago.
By sucking in the air the case was pressurized [proper engineering
design] and the majority of the air sucked in exited through the
power supply - which was an open cage device. Even if the air was
warmed by the chips it passed over on the way to the power supply
it was still sufficiently cool for power-supply cooling.
I've opened 3000s that had been running for 3 years or so, and the
inside was almost as clean as when it was first put together.
Good engineering and engineering to a price-point most often wind
up where price is the deciding factor. If proper engineering
principles had been used by IBM in the intial desgin things might
be better today. At least the good server cases - for the most
part - are designed well.
Thanks for the fan generating heat insight.
Jeff Liebermann
>In article <og963tg34r907s22n...@4ax.com>,
>Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
(...)
>>heat to the CPU heat load. A clue that this kludge doesn't work is
>>the necessity of add on CPU fans for even the most efficient CPU's.
>>Of course most CPU fans have smaller intake area than exhaust
>>resulting in lots of backpressure. Little wonder they rapidly burn
>>themselves up.
>
>A lot of that comes from designing to a price point - and using
>'standard' components.
Baloney. The little fans are rated at 3cfm (cubic ft/min) which would be
acceptable if they actually moved 3cfm. I've measured the intake
velocity of the typical Socket 7 CPU fan and calculated back to 4cfm in
free air, and about 1.5 cfm when mounted on the typical heat sink. The
little fan would be twice as effective if the heat sinks were somewhat
taller and the fan mounted on the *SIDE* of the heat sink. For extra
cooling, another fan could be mounted on the other side. The idea is
that the cold air intake area must be larger than the exhaust area or you
get back pressure. IMHO, rotten design that could have been done better
with a little effort. Incidentally, there is no such thing as a
"standard" heat sink as they all tend to be designed to fit in box.
(...)
>By sucking in the air the case was pressurized [proper engineering
>design] and the majority of the air sucked in exited through the
>power supply - which was an open cage device. Even if the air was
>warmed by the chips it passed over on the way to the power supply
>it was still sufficiently cool for power-supply cooling.
Sure. Do the math. The Model 3000 probably burned about 100 watts of
power from the AC line, all of which eventually became heat. The power
supply was at least 85% efficient, so only about 15 watts was dissipated
in the power supply. In addition, the power supplies tend to concentrate
the heat in a few components that have heat sinks, so a bit of added heat
in the power supply isn't a big deal. The other 85 watts was probably
equally divided between the the memory chips and the CPU/glue chips on
the motherboard.
Most designer mini-tower cases get their air intake from *UNDER* the
machine. This is stupid, criminally incompetent, but aesthetically
pleasing as nobody really wants to look at an air intake grill. I guess
it's too "industrial" looking. Yet, the bottom of case intake converts
the box into a nice vaccuum cleaner, able to suck in large volumes of
dust and dirt, and deposit this crud all over the machine. Of course, it
takes a while for this to happen so nobody notices until all the fans
start screaming because they're clogged with filth. Great design job.
>Good engineering and engineering to a price-point most often wind
>up where price is the deciding factor. If proper engineering
>principles had been used by IBM in the intial desgin things might
>be better today. At least the good server cases - for the most
>part - are designed well.
I'm not so sure. You'll never convince me that the typical 1RU box or
RAID array is properly designed for thermal efficiency. Perhaps the new
Compaq boxes, but not much else. Everything I see is a sardine can
kludge that is just asking for overheating. [Examples on request].
>Thanks for the fan generating heat insight.
Actually, the ATX box misplacement of the fan is more important. All my
techy friends hailed the ATX box as an improvement, while I looked it as
a thermal abomination with almost everything being done wrong. Add that
hot air tends to rise, while the ATX fan tries to push the hot air down.
Reversing the air flow means the intake area is the power supply intake
grill area, but the exhaust area is much much much smaller resulting in
backpressure. I've already discussed the increase in CPU heat loading
and the conflict with the CPU fan. Extra credit to those manufacturers
(i.e. Dell) that put the 3.3v regulator inside the power supply
effectively transfering its dissipation to the CPU (and creating an extra
connector on the power supply that I can't find a replacement).
Compaq really did it wrong with one of its Socket 7 desktop boxes. They
used a cardboard plenum (air duct) to move air from the power supply fan
grill to the CPU. However, the cardboard leaks badly, and totally covers
the CPU with no way for the hot air to get out. There's more air coming
out of the leaks than from around the CPU. One customer blew up 2ea
CPU's before I figured out what was happening. Horizontal plenums don't
work very well anyway.
E-machine gets my thermal abomination gold medal in their power supply
design. It about 1/2 the size of the typical ATX power supply and would
have worked fine if they hadn't butchered the fan. It projected a bit
too far outside the power supply case, so they simply milled off 1/2"
from the fan mounting ears so that the fan fit half inside and half
outside the power supply. However, this placed the fan right up against
all the circuity and effectively blocked the air intake to the fan. I've
got one client with 20 of these machines that is losing fans at an
alarming rate. E-machine won't sell just the butchered fans so I've been
milling my own. Yech.
Ever notice the weird looking cases for routers, modems, monitors, and
wireless access points, with lights on top, and rounded case tops?
They're designed to prevent you from stacking them or putting anything on
top. If you did that, you would break the convection riser air flow and
probably cause it to overheat.
Just got an offer for a free lunch. Gotta run...
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>On Sun, 10 Dec 2000 14:45:51 GMT, bi...@wjv.com (Bill Vermillion) wrote:
>
>>In article <og963tg34r907s22n...@4ax.com>,
>>Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>(...)
>>>heat to the CPU heat load. A clue that this kludge doesn't work is
>>>the necessity of add on CPU fans for even the most efficient CPU's.
>>>Of course most CPU fans have smaller intake area than exhaust
>>>resulting in lots of backpressure. Little wonder they rapidly burn
>>>themselves up.
>>
>>A lot of that comes from designing to a price point - and using
>>'standard' components.
>Baloney. The little fans are rated at 3cfm (cubic ft/min) which
>would be acceptable if they actually moved 3cfm. ... Incidentally,
>there is no such thing as a "standard" heat sink as they all tend
>to be designed to fit in box.
To clarify - I meant the style that has become standard that
started with the IBM PC.
>(...)
>>By sucking in the air the case was pressurized [proper engineering
>>design] and the majority of the air sucked in exited through the
>>power supply - which was an open cage device. Even if the air was
>>warmed by the chips it passed over on the way to the power supply
>>it was still sufficiently cool for power-supply cooling.
>Sure. Do the math. The Model 3000 probably burned about 100 watts of
>power from the AC line, all of which eventually became heat.
The point I was trying to get across - and failed - was that the
way to cool something is to filter air coming in - one air inlet -
instead of sucking air out and having air sucked in through every
opening including such things as tape drives and floppy disks.
Don't use a floppy for a year or so and when it comes time to load
an OS or a driver which comes on a floppy the heads might be dirty
enough not to be able to read things.
>Most designer mini-tower cases get their air intake from *UNDER* the
>machine. This is stupid, criminally incompetent, but aesthetically
>pleasing as nobody really wants to look at an air intake grill.
Yup. But I don't use mini-towers except on one of my home
machines. The rest are mid or full towers as are all my client
machines [a dwindling lot as they all seem to move to the MS
world].
>I guess it's too "industrial" looking.
And the above are. One wound up with 4 fans in the case and
the Cheetah's - three of them as I recall - were all mounted in
5.25" machined aluminum heat sinks with two mini-fans blowing
on the heat sink - sucking in air through the front panel.
>Yet, the bottom of case intake converts the box into a nice vaccuum
>cleaner, able to suck in large volumes of dust and dirt, and
>deposit this crud all over the machine. Of course, it takes a while
>for this to happen so nobody notices until all the fans start
>screaming because they're clogged with filth. Great design job.
Yup. And those machines typically get put in the worst part of the
building. I've had red-dirt in some of them. OTOH one of the
cleaner machines was in an automotive repair shop. THey had taken
a clean mechanics cloth and tape it to the top front so that it
hung down in front of the floppy drives and the QIC tape drive.
It was lifted up to change tapes. They did performance work there
too so they were more meticulous on their engines. OT: but one one
of their hobbies was also mud-bogging - and the owners wife was
national champion at one time.
>>Good engineering and engineering to a price-point most often wind
>>up where price is the deciding factor. If proper engineering
>>principles had been used by IBM in the intial desgin things might
>>be better today. At least the good server cases - for the most
>>part - are designed well.
>I'm not so sure. You'll never convince me that the typical 1RU box
>or RAID array is properly designed for thermal efficiency. Perhaps
>the new Compaq boxes, but not much else. Everything I see is a
>sardine can kludge that is just asking for overheating. [Examples
>on request].
I don't know about typical. The little iNTEL ISP1100R we've picked
as the server machine is well cooled. The Sun Netra-T1's dont seem
to be as well cooled. But I don't work around the 'typical' 1RU.
We're trying to go for five 9's.
>>Thanks for the fan generating heat insight.
>Actually, the ATX box misplacement of the fan is more important.
>All my techy friends hailed the ATX box as an improvement, while
>I looked it as a thermal abomination with almost everything being
>done wrong. Add that hot air tends to rise, while the ATX fan tries
>to push the hot air down.
Someone must have been asleep in physics class on that one.
>Reversing the air flow means the intake area is the power supply intake
>grill area, but the exhaust area is much much much smaller resulting in
>backpressure.
Sounds like they never worked on modifed engines where back pressure
robs horsepower either. Heck. I changed my studies half-way through
to an EE degree and I don't do things that stupid.
>Compaq really did it wrong with one of its Socket 7 desktop boxes.
>They used a cardboard plenum (air duct) to move air from the power
>supply fan grill to the CPU. However, the cardboard leaks badly,
>and totally covers the CPU with no way for the hot air to get out.
DEC has some screwy designs on their Pentium Pro series too IMO.
In another post you mentioned the IBM PS/2. Ever pull one of those
apart after a few years and find the 'foam' ducting has
disintegrated and the case is filled with little gooey piece of
foam. Yuck.
>Ever notice the weird looking cases for routers, modems, monitors,
>and wireless access points, with lights on top, and rounded case
>tops? They're designed to prevent you from stacking them or putting
>anything on top. If you did that, you would break the convection
>riser air flow and probably cause it to overheat.
I had to think about that in regards to routers because no router I
work with had weird cases or rounded designs. Then I started
thinking and I know what you mean. But I've been working more on
the 'supply side' of data. Up until last year my big router
had a mid-plane and a pullout fan tray - fan was about 8" - and it
blew down the front of one side - crossed the bottom with the dual
power supplies - and back up the back of the midplane.
I've thought the small fan on the 1RU Cisco devices - at an angle
exiting down and back - so you could stack the routers - was a bit
small. Never had problems unless they were mounted with a 1900
switch and then things got warmer than I'd like. But on the 2948
switch I have I think if they stuck anymore fans on that it wouldn't
fit in a 19" rack space.
>Just got an offer for a free lunch. Gotta run...
TANSTAAFL - they must want something :-)
Brian K. White
I think one such possibility, if you want to entertain the exotic, is to
construct all or part of the case out of something heat-dissipating, and
connect the cpu to the case with a heat-pipe. with that much surface, you
shouldn't need a fan at all, as long as you do something about the power
supply too. perhaps design the case and innards such that you get a good
convection chiminy going and let that keep the power supply from melting.
I know the heat-pipe plus large surface idea is already used in laptops.
> Digression: Fan heat load is not much of a problem with the typical AT
> box power supply because the fan is in the exhaust port causing its heat
> to leave the box without warming the boards or PS. Not so with the ATX
> abomination. I couldn't believe that the fan was moved to inside the box
> with the air blowing in. This keeps the power supply cool, pre-heats the
> air fed to the PS fan, and adds the fan power dissipation to the internal
> heat load. The fan is usually pointed directly at the CPU to add the
> power supply heat and fan heat to the CPU heat load. A clue that this
I think the main gist is that this design is probably a little quieter
overall, with the fan deep inside the case, and the any wind noise from
the concentrated iarflow at the fan itself also happening inside the case
rather than outside. The inferiority of the design has not the slightest
impact on it's marketability and thus it becomes the new standard. Hence
win-modems for example. If a zillion ignoramuses will buy it, and a few
hundred or few thousand technically conscious would prefer not to, what
manufacturer is going to hesitate the slightest bit?
> Socket 7 CPU's are anything but flat. A cross section will show that the
> surface is also very rough. Polishing CPU substrates is not my idea of
my socket7 cpu has an aluminum top, which I scanned as well as xeroxed,
and then lapped using the same rig my uncles shop uses to lap water seals
when they are non-replaceable for whatever reason. my cpu and exchanger
are flat ;) I've lapped purely ceramic cpu's too. you can see and feel the
raised lettering on some.
I think as long as you don't use a belt-sander to lap with, you can
probably flatten even the new flip-chip designs safely.
Kit Haskins
> Audio Consoles:
>
> R'n'R band - with some hanger-on they were trying to impress.
> Somebody was drinking 7 & 7 [Seagram's Seven and 7 UP ].
>
> Bump. Into the console. Session done for the day. They were still
> billed their $1K/day [approx]. Cost them a day more - 7 band
> members, producer, road crew, etc, in hotel and meals.
>
Maybe my favorite war story of this topic is being called into the tape
room (I worked at a few television stations as an engineer) to find a tape
machine dead. Wouldn't take any video tapes, and a news broadcast was
looming overhead.
A sharpe (thin black markers used by the personnel to mark the tapes) was
found setting inside the cassette basket. Removal was easy, just grab it
by the tip and pull it out, no one knew or claimed responsibility for it
being inside. Wasn't really surprised by it, you would be amazed at what
you find inside some of these machines. Keyboards have been know to fail
if they are filled up with sunflower seed shells, yuck.
I've also seen the heat damage the equipment. Vents get clogged up with
the dust, the studio is on a dirt road go figure, and nothing circulates.
One of the first signs of failure we would notice is the video being
produced by the machine would just look different. Smearing and
distortion on the color caused by electrolytic cap's that fail quickly
when they get hot.
ForA Time base correctors with large arrays of DIP IC's for memory where
also prone to heat failure. A single 1" by 1" fan was the cooling,
nothing else. The TBC has two boards populated with the older TTL IC's
that produce heat and can be very discomforting if touched by the
unprotected hand, very hot. We could tell roughly how long the TBC was
running by the temperature of the exhaust air on the rear. They would
loss color lock, the color would turn into a psycodelic rainbow effect
and that was it. Pull the unit out for a few hours and let it cool down.
If it was needed sooner we placed in the the refrigerator in the break room
for 15 minutes and stuck it back in the rack for the rest of the newscast.
Microtime Time Base Correctors also like to fail due to heat. It was the
power supply that usually failed and would take a few friends out with it.
The design of the case had a muffin fan the same size as a PC power supply
but the intake was on the front face with a 1/2" horizontal band of
vertical slots ( |||||||||||||||||| ). The vent was in front of four
boards again populated with a lot of TTL heat producing IC's that eventual
found it's way to the power supply. The hot air would travel across the
power supply and leave even hotter as it traveled across some heavy duty
transistors that handled large amounts of current, again this is a large
box with TTL IC's. The 5 volt line would spike up to about 20 volts as
the transistor failed. Everything on the 5 volt bus was suspect of
failure once the power supply was repaired. We replaced the unit after
the third failure.
The design of thermal dynamics isn't just ignored with the computer world,
it seems to be in other expensive items. We where told that the design of
the equipment was for use in an environmental controlled room. With
people walking in and out with video tapes that need to be edited two or
three times a day, satellite feeds and retransmission, having it
controlled is mission impossible. If a few extra minutes was given in the
design process, the machines we have wouldn't fail as often. It might add
a few bucks, but when you pay $50,000 per tape machine another $500 for a
properly working cooling system isn't anything major when the cost to
repair it can exceed several $1k.
---------------------------------------------------------------------------
Thru the Ethernet, past the Gateway, off the modem pool, nothing but NET .
k...@ka0wuc.org
Charlie Gibbs
je...@comix.santa-cruz.ca.us (Jeff Liebermann) writes:
>Most designer mini-tower cases get their air intake from *UNDER* the
>machine. This is stupid, criminally incompetent, but aesthetically
>pleasing as nobody really wants to look at an air intake grill. I
>guess it's too "industrial" looking. Yet, the bottom of case intake
>converts the box into a nice vaccuum cleaner, able to suck in large
>volumes of dust and dirt, and deposit this crud all over the machine.
>Of course, it takes a while for this to happen so nobody notices until
>all the fans start screaming because they're clogged with filth.
>Great design job.
The greatest. The machine overheats, or the contacts get cruddy,
or man-eating dust bunnies grow in the floppy drive. One way or
another, the machine soon breaks down, so the customer throws it
out and replaces it with one that'll do the same in short order.
Guaranteed repeat sales in the event the suck^H^H^H^Hcustomer
doesn't fall for the old planned obsolescence scam: "What? You're
depending on a mere 386 to collect two records per hour? It's a
good thing that box broke down, mate - you need a 450-MHz Pentium
II at the very least to handle that sort of workload." (And the
brand-new version of Windoze that the new machine comes with can
make that a self-fulfilling prophecy.)
--
cgi...@sky.bus.com (Charlie Gibbs)
Remove the first period after the "at" sign to reply.
Jeff Liebermann
>Yup. But I don't use mini-towers except on one of my home
>machines. The rest are mid or full towers as are all my client
>machines [a dwindling lot as they all seem to move to the MS
>world].
You should see the big boxes that Unisys dragged to SCO Forum last
August. It was the size of larger fridge and required a raised floor
with forced air from a external air conditioner.
I mangled one point. The AT style case blows air out the back from the
power supply fan, and sucks air in from anything with a hole. The
vaccuum cleaner effect is from the air intakes being a slot under the
front plastic panel. The ATX box design was apparently intended to solve
this problem by reversing the air flow and having air go into the power
supply. In this respect, the ATX box is better, but in all other
respects (heat load, cpu cooling, exhaust port area, etc), it doth suck.
>I don't know about typical. The little iNTEL ISP1100R we've picked
>as the server machine is well cooled. The Sun Netra-T1's dont seem
>to be as well cooled. But I don't work around the 'typical' 1RU.
>We're trying to go for five 9's.
The 1RU boxes, where the motherboards and power supplies are custom
designed to fit properly, are usually quite good. The ones that just
cram a stock ATX motherboard into a 1RU box are horrible.
>Sounds like they never worked on modifed engines where back pressure
>robs horsepower either. Heck. I changed my studies half-way through
>to an EE degree and I don't do things that stupid.
Most computer designers have never done anything serious with thermal
design. I've had the experience (bordering on nightmare) of designing
medium-power RF transmitters, where thermal design is manditory. After
destroying a few hundred dollars worth of output transistors, I dug out
the textbooks, RTFM, and did it right. At this time, the water cooled
CPU overclockers are doing the essential thermal design that should have
been done by the various manufactories. There's also a tendency for
commodity board designers to pass the buck on thermal design to the box
builders. That's a good excuse for commodity clone boxes, but the name
brand manufactories do the same thing and don't have the same excuse.
>I had to think about that in regards to routers because no router I
>work with had weird cases or rounded designs. Then I started
>thinking and I know what you mean.
Weird looking: Linksys, Netopia, Cisco for the home, Lucent.
Some of the designs go well beyond functionality and closely resemble a
sculpture or are aerodynamicly correct. Why a router needs to be
streamlined is beyond my limited imagination.
>>Just got an offer for a free lunch. Gotta run...
>TANSTAAFL - they must want something :-)
Oh yes. It was a trap and I fell for it.
Tony Lawrence
> Guaranteed repeat sales in the event the suck^H^H^H^Hcustomer
> doesn't fall for the old planned obsolescence scam: "What? You're
> depending on a mere 386 to collect two records per hour? It's a
> good thing that box broke down, mate - you need a 450-MHz Pentium
> II at the very least to handle that sort of workload." (And the
> brand-new version of Windoze that the new machine comes with can
> make that a self-fulfilling prophecy.)
Recently I was involved with an upgrade where the vendor
specified the necessary hardware to run their app for a 35 user
installation. They were not selling the hardware, so it wasn't a
case of greed, but their specs were: Dual Pentium 700 or above, 1
GB of ram, and 4 x 18 GB ultra scsi raid 5. For 35 users. I
was sure they must have been spec'ing for Windows, but nope, that
was for Unix.. I tried to explain to the customer that this was
complete overkill, but the vendor took a hissy fit, said they
couldn't possibly guarantee performance with anything less than
this, so the customer bought it.
Makes you wonder how they managed to sell anything 3 or 4 years
ago.. :-)
Bill Vermillion
>A sharpe (thin black markers used by the personnel to mark the
>tapes) was found setting inside the cassette basket.
I got out of TV by that time. Everything was the 'new' 1"
C-format. [I remember watching engineers edit video tape with
Magna-See and razor blade - so you could see the sync marks. This
was 2" quad. Audio came later and was more fun].
>ForA Time base correctors with large arrays of DIP IC's for memory
>where also prone to heat failure.
And that was a small amount of chips compared to the device that
was in my friends warehouse until last year - before it made it
to the gold-collectors.
It was a Marconi color-tv standards convertor. First one in the US.
Cost about $250,000 when a friend of mine bought it [we were both
r'n'r DJs when we first met]. And a quarter-of-a-million dollars
was REAL money back in 1970 [about the time of this beast].
Huge boards - mounted vertically - on about 1" spacing. Easily a
foot deep. Two racks crammed full of these boards. It did a lot of
work in the old days. He got the contract for and converted all
of ABC's Wide World of Sport from NTSC to the PAL format.
Seeing beasts like this which with all the boards still probably
had fewer transistors by an order of magnitude [or more] than
a current CPU makes you really appreciate what we have today.
- bill -
>
> - bill - <bi...@TechServsys.com> writes:
>
> > *Tea: 2/3 Twinings English Breakfast tea, 1/3 Twinings Irish Breakfast
> > tea, _LIBERALLY_ sweetenend with honey. (actually very liberally - got
> > to get the motor running in the am)
>
> Honey is hygroscopic, it readily absorbs water. That's why it is so
> easy to wash off.
>
> One simple solution is to leave your keyboard stand outside for a few
> hours. Quaranteed - if there are bees in the locality, which is quite
> likely - they have it cleaned up in no time.
>
> Hmmmm - I hope the honey was Irish as well :-) I kept 50 hives one
> time. Sometimes I really regret having let myself become distracted
> by computers :-(
>
> --
> joe mc cool SMIEEE
> ========================================================================
are not very active now.
I also used to keep bees, a long time ago. You, me and Sherlock Holmes.
> voice:(44)2837-548074fax:(44)-870-0520185 The more you say the less the better.
--
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>On Mon, 11 Dec 2000 00:34:12 GMT, bi...@wjv.com (Bill Vermillion) wrote:
>>Yup. But I don't use mini-towers except on one of my home
>>machines. The rest are mid or full towers as are all my client
>>machines [a dwindling lot as they all seem to move to the MS
>>world].
>You should see the big boxes that Unisys dragged to SCO Forum last
>August. It was the size of larger fridge and required a raised floor
>with forced air from a external air conditioner.
Sounds like the big Compaq machines that are in 6 foot racks.
But requiring a raised floor means that machine was intended for
a server room environment. That's something I'd not think would be
an SCO marketplace - at least based on my view. With requirements
like that you'd probably have to have a least a dozen units such as
that to make it cost-effective.
>I mangled one point. The AT style case blows air out the back from
>the power supply fan, and sucks air in from anything with a hole.
>The vaccuum cleaner effect is from the air intakes being a slot
>under the front plastic panel. The ATX box design was apparently
>intended to solve this problem by reversing the air flow and having
>air go into the power supply. In this respect, the ATX box is
>better, but in all other respects (heat load, cpu cooling, exhaust
>port area, etc), it doth suck.
That was my point on the AT design.
>>I don't know about typical. The little iNTEL ISP1100R we've picked
>>as the server machine is well cooled. The Sun Netra-T1's dont seem
>>to be as well cooled. But I don't work around the 'typical' 1RU.
>>We're trying to go for five 9's.
>The 1RU boxes, where the motherboards and power supplies are custom
>designed to fit properly, are usually quite good. The ones that just
>cram a stock ATX motherboard into a 1RU box are horrible.
I'll agree. If you really need 1RU units for space - it's going
to be hard to custom-build your own AND get reliability.
>>Sounds like they never worked on modifed engines where back pressure
>>robs horsepower either. Heck. I changed my studies half-way through
>>to an EE degree and I don't do things that stupid.
>That's a good excuse for commodity clone boxes, but the name brand
>manufactories do the same thing and don't have the same excuse.
And you should hear some of the names they call those brands :-)
Jeff Liebermann
>It was a Marconi color-tv standards convertor. First one in the US.
>Cost about $250,000 when a friend of mine bought it [we were both
>r'n'r DJs when we first met]. And a quarter-of-a-million dollars
>was REAL money back in 1970 [about the time of this beast].
[Bringing topic drift to a new all time low...]
What ever happened to those monitor plus camera in a box kludges that
Sencor and other bottom of the line equipment vendors were building for
scan conversion? They worked well enough as long as you weren't into
extreme quality. I had one on a film chain with interchangeable cameras
so I could crank out PAL, SECAM or NTSC. I don't recall the cost but it
was considerably less than $250,000. The real headaches were the Ampex
and Sony VTR's (Video Tape Recorder). I had to sit there with a spray
can of freon head cleaner and spray the heads while recording or they
would consistantly clog about half way through the tape. Them was the
good olde daze.
Robert Lipe
>In article <n2693t8u3018ife6q...@4ax.com>,
>Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>>You should see the big boxes that Unisys dragged to SCO Forum last
>>August. It was the size of larger fridge and required a raised floor
>>with forced air from a external air conditioner.
>Sounds like the big Compaq machines that are in 6 foot racks.
[ Checking non-disclosures. ]
Bigger. We're talking about 96 PCI slots, 32 P-III Xeons or IA-64's.
http://unisys.com/hw/servers/enterprise/7000/default.asp
>But requiring a raised floor means that machine was intended for
>a server room environment.
Uuuuh, yeah. That's really not a problem. I don't remember the price
tag on these boxes, but it's not hard to reach 7 digits. These are not
meant to go in your mom's sewing room :-)
>That's something I'd not think would be an SCO marketplace - at least
>based on my view. With requirements
With UnixWare, this thing SCREAMS on database transactions.
John DuBois
The last machine I had a cooling problem with was a 486/33 system, which worked
fine until I started overclocking it at 40MHz. Actually, even that seemed to
work OK at first, until Bela started compiling nethack - yes, even at that
generation, CPU load made a huge difference in the power consumption. We found
that a few minutes of compiling was enough to overheat the CPU and lock it up.
Fortunately, that case had a secondary fan which happened to be positioned such
that I was able to make a small cardboard ductwork that channeled the entire
output of the secondary fan onto the CPU heatsink. It worked perfectly
reliably from then until the day one of the CMOS backup batteries exploded and
spewed corrosive electrolyte all over the motherboard. No amount of cleaning
(with deionized water, alcohol, flux remover, etc.) would ressurect it then.
Bill Vermillion
>>In article <n2693t8u3018ife6q...@4ax.com>,
>>Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>>>You should see the big boxes that Unisys dragged to SCO Forum last
>>>August. It was the size of larger fridge and required a raised floor
>>>with forced air from a external air conditioner.
>>Sounds like the big Compaq machines that are in 6 foot racks.
>[ Checking non-disclosures. ]
>Bigger. We're talking about 96 PCI slots, 32 P-III Xeons or IA-64's.
Can't
>
> http://unisys.com/hw/servers/enterprise/7000/default.asp
>
>>But requiring a raised floor means that machine was intended for
>>a server room environment.
>Uuuuh, yeah. That's really not a problem. I don't remember the price
>tag on these boxes, but it's not hard to reach 7 digits. These are not
>meant to go in your mom's sewing room :-)
I do recall the Compaq's base price was $199,000. The
point I seemed to miss making - was that in this environment you
would typically have LOT's of machines as computer rooms with
raised floors are darned expensive to build. [My partner has 3000
sq ft of raised flooring he picked up for a song at an auction -
but now he see that unless you are an enterprise it's better to put
you machines in a carrier-class facility and let them pay for all
those things - floor, Liebert AC, UPS, diesel generators, et al].
But when you pay 7 figures for a machine it makes sense.
I used to do some small SCO system work at one place where the
actually upgrade Unisys MF's three times in a 10 year period.
They do build serious machines. The above grew faster than they had
imagined.
>>That's something I'd not think would be an SCO marketplace - at least
>>based on my view. With requirements
>With UnixWare, this thing SCREAMS on database transactions.
Yup. Keep forgetting the UW7 target. The Compaq's above were
running [as I recall] TruUnix 64 the only 64x64 Unix out there, so
those were probably big Alphas. Part of the DEC heritage.
Along that line Compaq announced that this year they will stop
producing Vax systems. But I guess a lot of people thought those
were gone already.
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>On Mon, 11 Dec 2000 14:02:39 GMT, bi...@wjv.com (Bill Vermillion) wrote:
>>It was a Marconi color-tv standards convertor. First one in the US.
>>Cost about $250,000 when a friend of mine bought it [we were both
>>r'n'r DJs when we first met]. And a quarter-of-a-million dollars
>>was REAL money back in 1970 [about the time of this beast].
>[Bringing topic drift to a new all time low...]
[and I'll conclude it with this post]
>What ever happened to those monitor plus camera in a box kludges that
>Sencor and other bottom of the line equipment vendors were building for
>scan conversion? They worked well enough as long as you weren't into
>extreme quality.
I'll argue the 'extreme' modifier and say they worked OK if you
didn't care about quality. Even standards conversions look pretty
foul to a critical eye.
> I had one on a film chain with interchangeable cameras so I could
>crank out PAL, SECAM or NTSC. I don't recall the cost but it was
>considerably less than $250,000.
It would have to far less. :-) The person who owned the above
has built one or two DVD mastering labs recently - so you can see
he's used to 6 and 7 figures as a minimum. The Marconi above wound
up stored in my partner's warehouse - and his pricey 'toy' was
his Rank Cintel flying spot scanner - that went for $350,000.
>The real headaches were the Ampex
>and Sony VTR's (Video Tape Recorder). I had to sit there with a spray
>can of freon head cleaner and spray the heads while recording or they
>would consistantly clog about half way through the tape. Them was the
>good olde daze.
Sounds like worn-head, bad tape [or storage of tape], and possibly
a combination of both :-)
Back to computers!