|
Grex > Hardware > #83: Question: Do 60 or 66 MHz computers actually get too hot? | |
|
| Author |
Message |
leann
|
|
Question: Do 60 or 66 MHz computers actually get too hot?
| 
Jan 16 17:53 UTC 1995 |
I am currently computer shopping and was given what seemed
strange advice by a friend yesterday. She had been told to
avoid 60 and 66 MHz speed computers because they get very
hot and make rooms where they are being used uncomfortable.
The advice came from my friend's "computer guru" who I
suspect only deals with PCs, not Macs.
I'm shopping for a Power Mac. Does anyone out there
know whether this "hot" rumor has any basis in reality? If
so, is it true for Macs as well as PCs?
|
| 41 responses total. |
scg
|
|
response 1 of 41:
|
Jan 16 22:42 UTC 1995 |
The hot rumor has some basis in reality -- on Pentiums. I'm not sure how
noticable a difference they made as far as room temperature went, but they
apparrently got hot enough for the heat to be bad for the chip, leading to
such jokes as, "Pentium: melts in your motherboard, not in your hand." If
you're buying a PowerPC, don't be deterred by Pentium problems. They're
different chips. Also unlike the Pentium, the PowerPC knows how to divide.
|
n8nxf
|
|
response 2 of 41:
|
Jan 17 13:43 UTC 1995 |
Heat a room with your Pentium? I don't think so. I suspect the heat your
friend was feeling was from tempers building trying to get it to work properly
In all honesty, if the pentium chip disipates more than 10 watts of power,
you could melt solder with it. A bit more heat than a night light would
give off.
Taking it to an unrealistic extreme, your average Pentium PC is going to
come with a 250 or so watt power supply. It can't supply any more than
that. (I'll bet most clone supplies would die if they had to supply that
level of power for any length of time.) The same amount of power given
off by five 50 watt light bulbs. Can one heat the room in question with
five 50 W bulbs?
|
scg
|
|
response 3 of 41:
|
Jan 17 22:38 UTC 1995 |
And if the heat was causing problems for the chips, wouldn't that
indicate that it wasn't disipating the heat very well, and was holding it
in instead?
|
mdw
|
|
response 4 of 41:
|
Jan 18 03:23 UTC 1995 |
There's a very nice video of an intel worker frying an *egg* on a
pentium. Power dissipation and temperature aren't necessarily related -
even a tiny incandescent light bulb internally manages very fiery
temperatures, while dissipatingly only a tiny fraction of the power of a
large wall plug radiant electric heater.
5 50 W lightbulbs is actually equivalent to a 850 btu/hr heater - or
about 1/6th of an ordinary room electric heater. It's certainly capable
noticeably raising the temperature of a small room. Don't forget, too,
to add in extra heat from the monitor (even an old tiny b/w monitor is
good for 60 w - and some of these big modern color displays are a lot
more thirsty). Modems can also be somewhat power hungry, especially
older higher speed units - and the wall plug transformers for some can
also get kind of warm. The thing about all computer equipement is,
whatever else it's good for, it's *always* at least 99% efficient at
converting all of the electricity fed into it, into heat.
People, too, are another source of heat - for people just standing
around, the usual figure is something like 60 w. That's why crowded
parties with 20-30 folks in a room can get kind of warm.
|
jep
|
|
response 5 of 41:
|
Jan 18 04:09 UTC 1995 |
Most high speed PC-type computers have fans on the chip to make sure
the chip doesn't get too hot. I don't know if that's true for Macintoshes
as well, but I expect it is. I don't know of anyone who has a problem
with his high-speed PC generating too much heat for his room, not because
of the chip speed, anyway. People with clunky old MF/M or ESDI hard
drives, and people with large screen monitors (17" and bigger), and people
with computers loaded down with lots of components, are more likely to
have heat problems in their rooms.
I knew someone with 8 small hard drives on his CP/M computer in
college; not only did the computer generate enough heat to warm up his
dorm room in short order, but when he started the thing up, it sounded
like an airplane taking off.
Arbornet ran a computer like Grex for a night in the tiny office in
the NEW Center; after 8 hours, the room felt like it was about 120 degrees
inside (I was told). Arbornet runs 2 486/33 computers in that office
(and 24 modems) without any noticeable problems due to heat.
I don't think you'll have any problem with your Power Mac, but if
you're worried about it, touch the case on one that runs most of the time
where you plan to buy your computer. If the case is noticeably warm,
it'll give you some indication as to how much heat the computer will
generate in your house. For comparison, try touching a light bulb that's
been on for an hour or so. (-:
|
scg
|
|
response 6 of 41:
|
Jan 18 04:55 UTC 1995 |
I have an IBM CGA monitor from 1983 that makes a very nice space heater.
I'm very glad I'm not using it on my computer anymore, for more reasons
than just the graphics.
|
mdw
|
|
response 7 of 41:
|
Jan 18 05:27 UTC 1995 |
I've never heard of tiny fans on MPU's until the pentium came out. I
haven't actually cracked the case of a Power Mac yet, but to the best of
my knowledge, there aren't any of those ridiculous of those ridiculous
tiny fans inside. I know there aren't any of them in a Quadra (68040)
or inside an RS/6K (Power2 chipset, the big brother to the power PC
chip.)
|
leann
|
|
response 8 of 41:
|
Jan 18 16:55 UTC 1995 |
Thanks for all of the input, everyone. I suspected as much...
but it was worth asking before I shell out thousands of $$!
|
rcurl
|
|
response 9 of 41:
|
Jan 18 18:43 UTC 1995 |
I recall that the PowerPC Motorola RISC chip draws a lot less power
than the Pentium, while doing the same "work". Does anyone have the
figures on that? Of course, everything else - monitor, drives,
loudspeakers (;->) - consume about the same power. Oh, for the record:
the PowerBook does not have a fan - except us PowerBook fans.
|
jep
|
|
response 10 of 41:
|
Jan 19 03:01 UTC 1995 |
Marcus, Jemmie Wang of CCS first started talking about CPU fans when he
was pushing his DX2/50 to 66 MHz. When Everex was trying to push an early
486 chip to 80 MHz, they used a water tower to cool it. That was at least
a couple of years ago... Everex went bankrupt about then.
Fans, anyway, are common on high-speed 486 chips and on Pentiums. If
they aren't being used on Motorola chips yet, I think they surely will be
soon as Motorola pushes their chips to keep up with Intel chips.
|
scg
|
|
response 11 of 41:
|
Jan 19 03:13 UTC 1995 |
I think I remember noticing a heat sink on my step mom's PowerPC, but I'm
not sure about that.
|
mdw
|
|
response 12 of 41:
|
Jan 19 05:25 UTC 1995 |
Actually, John - I would be surprised if Motorola did so. Fans are an
extra cost option and something else to break, and chips that take that
much power aren't as useful for many applications. One of the markets
Motorola has been in for a long time is that of automotive electronics -
and in fact one auto company has already comitted to using some flavour
of power PC chip in transmissions. Indeed, fans of any type are bad
enough for automotive applications that they've been fooling around with
all sorts of passive chip coolers long before they found their way into
personal computers. A chip that requires a CPU fan just isn't going to
make it. For portable applications, too, low power is a "must". Quite
aside from the problems of packaging a CPU fan into a portable -
something that eats that much power will do nothing good for battery
lifetime or weight - 2 key parameters in portable design. These are
important markets for Motorola; they are unlikely to favor designs that
would close them out of these markets.
The pentium needs fans because, in a sense, it's pushing the technology
beyond what it's capable of. The PowerPC is a simplier yet more
powerful chip - so it doesn't need to ride the leading edge of
integration quite so hard.
|
n8nxf
|
|
response 13 of 41:
|
Jan 19 13:43 UTC 1995 |
Heat is what kills most electronics, contraction and expansion being the
biggest culprit. Fans can do a lot to help inprove heat disipation. Fans,
however, are mechanical devices and will not have the MTBF of a reasonably
designed chip. I've seen too many clone power supply fans fail to have
much faith in them keeping a Pentium CPU from washing away in its own
puddle of solder.
In the Jan. 12, 1995 issue of MACHINE DESIGN, they show the concept of
using heat piles to carry the heat away from the CPU. Oddy enough, the
drawing shows a Pentium chip in a laptop with the heat pipe on top of it.
The artist didn't elaborate on the power source for this confowlation.
Perhaps a deep-discharge truck battery? ;)
|
jep
|
|
response 14 of 41:
|
Jan 20 02:08 UTC 1995 |
I've also seen heat sinks on some processors. Maybe these are more
common than fans.
If your 80 MHz processor burns out after a year because the fan died,
and you have to replace it, what are you going to replace it with --
another 80 MHz processor, or the newest, latest processor that will fit in
the same slot? I think a lot of people are opting for whatever's being
touted as "the latest". This must break Intel's heart.
|
mdw
|
|
response 15 of 41:
|
Jan 20 06:48 UTC 1995 |
At one point, I saw a power chart for at least 6 risc & cisc
chips, but now I can't find it. I did, however,
find this comparison of the PowerPC and Alpha:
PowerPC 601 Alpha 21064
Technology 0.6-micron CMOS 0.75-micron CMOS
Levels of metal 4 3
Die size 1.09 cm sq 2.33 cm sq
# Transistors 2.8 million 1.68 million
Total cache (I+D) 32 Kbyte 16 Kbyte
Package 304-pin QFP 431-pin PGA
Clock frequency 50 Mhz, init. 150 to 200 Mhz.
Power dissipation 9 watts @ 50 Mhz 30 watts @ 200 Mhz.
This is ca. Jan 1994, so faster 601's may be out today. Particularly
for CMOS, power dissipation is largely proportional to speed - so a 25
Mhz 601 would likely only consume 5 watts. Within a particular batch,
individual chips will often run faster, so it's quite possible that by
retesting 601's at a higher speed and then running them with CPU fans,
one could have a 75 Mhz 601. At one point, a company was doing
something like this with 68040's and Macintoshes.
|
gregc
|
|
response 16 of 41:
|
Jan 21 00:25 UTC 1995 |
To answer the original poster's question:
Yes, avoid the 60 and 66 mhz pentiums. They get too hot. No, they don't have
enough heat to significantly affect the temperature of the room, but the
heat is bad for the chip itself. I have heard reports that the extreme
heat inside a 66 mhz pentium is enough to cause slow metal migration on the
circuit pathways that could lead to premature failure of most 60/66
pentiums after only a few years.
This only applys to 66mhz pentiums. 66mhz 486 DX2's are fine.
|
scg
|
|
response 17 of 41:
|
Jan 21 05:06 UTC 1995 |
(The original poster was asking about PowerPCs, which are also fine)
|
n8nxf
|
|
response 18 of 41:
|
Jan 23 13:07 UTC 1995 |
Yep, 7100 has a massive heat sink. No fan.
|
davel
|
|
response 19 of 41:
|
Apr 9 01:56 UTC 1995 |
If you do get a system with an on-chip fan, be *careful*. I know of
one case where a cable got in the way of the fan, preventing it from
turning. The system produced some really strange symptoms before the
cause was discovered, I think.
|
n8nxf
|
|
response 20 of 41:
|
Apr 10 12:08 UTC 1995 |
Also, many of those fans will die after 2 or 3 years of use. Most of
them use bushings impregnated with low-grade oil and simply dry/gunk
up. Ball bearing fans are much better. Same goes for the fans they
put in most of the clone power supplies.
|
kuran
|
|
response 21 of 41:
|
Aug 31 14:44 UTC 1998 |
Fans also cause a lot of dust built up in PCs, which can cause all sorts
of other problems if it isn't kept in check. BTW, this is my first time
reading this conference, and I it was great. Nowadays you wouldn't
dream of buying a machine without a fan, and most will have a heat sink.
I just picked up a new machine and the sink is bigger than the chip by
several times :P Interesting how attitudes change...
|
n8nxf
|
|
response 22 of 41:
|
Sep 1 09:46 UTC 1998 |
An air compressor fitted with a hose and air gun does a decent job of
blowing the dust out of computers and power supplies. I have also
noticed a trend towards lower speeds and ball bearings in fans.
|
gull
|
|
response 23 of 41:
|
Sep 2 05:03 UTC 1998 |
COMPAQ is one of the few companies that still doesn't use chip fans. Their
preference is for very large heatsinks, sometimes combined with extra case
fans.
|
n8nxf
|
|
response 24 of 41:
|
Sep 2 09:48 UTC 1998 |
the fewer mechanical parts, the better the reliability. A heatsink will
outlast any fan.
|