Reply 40 of 61, by mrau
i dot get that sarcasm here, have you ever seen the heatsink in those apple intel workstations? those are the same cpus that we cool with powerful 50k+ rpm fans in our pcs - all it takes is a well thought sink + case fans;
i dot get that sarcasm here, have you ever seen the heatsink in those apple intel workstations? those are the same cpus that we cool with powerful 50k+ rpm fans in our pcs - all it takes is a well thought sink + case fans;
wrote:Would something like this be enough for anything socket 7 then? […]
Would something like this be enough for anything socket 7 then?
Tried that specific model and t's indeed a very appropiate heatsink for a S7 system (some photos My Pentium MMX machine (finished!)), but you may want to attach a fan to it. Mine is from Fractal Design and it's virtually inaudible. You may also try a common one at lower voltage 😀
wrote:The heat sink has to be the right thickness to maximize heat conductance to the fins while maintaining small enough heat resistance from the base.
But wouldn't a thinner base also mean that less heat will reach the edges of the base, causing some of the spikes near those edges to kinda remain unemployed? I presume they increased the thickness of these heatsinks for a good reason, later ones even got a copper base at some point.
mass is a factor with heatsinks, more moass - more reliable cooling;
wrote:wrote:The heat sink has to be the right thickness to maximize heat conductance to the fins while maintaining small enough heat resistance from the base.
But wouldn't a thinner base also mean that less heat will reach the edges of the base, causing some of the spikes near those edges to kinda remain unemployed? I presume they increased the thickness of these heatsinks for a good reason, later ones even got a copper base at some point.
That's what I'm saying, the thermal solution has to be designed with the right balance. Too thick and you have too much thermal resistance from front to back. Too thin and you don't have effective heat conduction to the outside of the heat sink.
All hail the Great Capacitor Brand Finder
wrote:wrote:wrote:The heat sink has to be the right thickness to maximize heat conductance to the fins while maintaining small enough heat resistance from the base.
But wouldn't a thinner base also mean that less heat will reach the edges of the base, causing some of the spikes near those edges to kinda remain unemployed? I presume they increased the thickness of these heatsinks for a good reason, later ones even got a copper base at some point.
That's what I'm saying, the thermal solution has to be designed with the right balance. Too thick and you have too much thermal resistance from front to back. Too thin and you don't have effective heat conduction to the outside of the heat sink.
Ah, I thought you meant all the fins and weren't specifically talking about the fins around the edges, but we meant the same thing anyway 😀.
If there aren't lots of peripheral components protruding from the mobo close to the socket, you could always use a Socket 370/A heatsink with Socket 7. There are plenty of large and effective models around, that probably could handle the heat even from higher clocked Pentiums without a fan.
If not, even a minimal airflow does wonders. A S370/A heatsink with an 80mm fan at 6-7 volts, or something like that, will be more or less inaudible, and will definitely handle anything a Socket 7 CPU can throw at it.
wrote:Ah, I thought you meant all the fins and weren't specifically talking about the fins around the edges, but we meant the same thing anyway 😀.
Yup, we were pretty much on the same page. CPU cooling is really fascinating. I can think of a litany of considerations for designing (and choosing) a heat sink, and I'm not even a thermal engineer. For something so simple on the surface, it's actually a very complex problem.
All hail the Great Capacitor Brand Finder
wrote:wrote:Ah, I thought you meant all the fins and weren't specifically talking about the fins around the edges, but we meant the same thing anyway 😀.
Yup, we were pretty much on the same page. CPU cooling is really fascinating. I can think of a litany of considerations for designing (and choosing) a heat sink, and I'm not even a thermal engineer. For something so simple on the surface, it's actually a very complex problem.
It's definitely more complex then it first appears to be on the surface.
One thing I thought about, was when I noticed from the patterns of dusty HSFs, that even though the middle of the heat conductive base plate (or whatever these are called) would typically be the hottest part, HSFs that have the fan on top of it, actually cool worse in the middle and better at the edges.
Considering this, a "blow-through"-type HSF makes perfect sense.
But (and this is a thing most of us know by now) the top-down coolers help cool other parts.
Case ventilation is another such thing, I've always preferred tower cases with the PSU in the top, as I figured the bottom-mounted ones would be more susceptible to dust and break, as actually happened to a friend of mine who lost a 600W Antec this way (he forgot to clean the dust filter or something, so it overheated soon afterwards). I think top-mounted PSUs also help with cooling of the caase and the top side of the motherboard (I'm no expert though, I've learned all this by experience and by using common sense..or try to 🤣). That the air cooling the PSU is a bit hotter, I take for granted as this is kinda compensated by the PSU not having to eat so much dust.
wrote:... I noticed from the patterns of dusty HSFs, that even though the middle of the heat conductive base plate (or whatever these are called) would typically be the hottest part, HSFs that have the fan on top of it, actually cool worse in the middle and better at the edges.
That's due to the "dead zone" of the fan where there's no air flow due to the rotor part being in the middle. Has nothing to do with the heatsink itself.
sorry for necro;ing but i think its alot more usefull content wise for me to just add to this thread
id like to see some pics of exactly how large a passive heatsink i would need for a Pentium 120, 133, or 166.
apparently from what i can tell, any socket 370 heatsink+fan or passive cooler will also fit a socket 7 board
is this true??
thats just from observing ebay here:
https://www.ebay.com/b/Socket-7-CPU-Fan-with- … 31486/bn_652123
am i correct??
http://www.oldschooldaw.com | vintage PC/MAC MIDI/DAW | Asus mobo archive | Sound Modules | Vintage MIDI Interfaces
AM386DX40 | Asus VL/I-486SV2GX4 (486DX2-80) | GA586VX (p75) + r7000PCI | ABIT Be6 (pII-233) matroxG400 AGP
wrote:sorry for necro;ing but i think its alot more usefull content wise for me to just add to this thread id like to see some pics of […]
sorry for necro;ing but i think its alot more usefull content wise for me to just add to this thread
id like to see some pics of exactly how large a passive heatsink i would need for a Pentium 120, 133, or 166.apparently from what i can tell, any socket 370 heatsink+fan or passive cooler will also fit a socket 7 board
is this true??
thats just from observing ebay here:
https://www.ebay.com/b/Socket-7-CPU-Fan-with- … 31486/bn_652123
am i correct??
wrote:id like to see some pics of exactly how large a passive heatsink i would need for a Pentium 120, 133, or 166.
I had an IBM PC300GL once with a Pentium 200MMX and it had a passive cooler. It was about the same size as the big blue cooler in the picture Tetrium linked to. The only case fan was in the PSU AFAIR, but I still added a 80mm fan to the front of the case to increase airflow just because it made me feel safer, although that PC worked fine without it for years before I got it.
That big blue heatsink is actually a passive Pentium (non-MMX) heatsink. I salvaged several of those from (mostly or maybe even exclusively) OEM systems that indeed had perhaps only a single fan cooling the system.
The mounting clips of those were very hard to work with (easy to slip with a flatbed screwdriver or no spot for flatbed screwdrivers at all!!), and as I decided to exchange the clips with ones that are easier and less frustrating (and dangerous) to work with, so that's probably why the metal mounting clip wasn't included in the pic.
These passive coolers make a kinda distinct sound when used or when bumped into with (for instance) a screwdriver or another heatsink and came in several pretty colors 😁
They had very thin bases compared to the later s370 Intel stock HSFs also and these heatsinks were usually paired with ceramic Pentium CPUs of roughly 75MHz up to around 133MHz, or at least in the systems I found them in 😀
wrote:My Original Pentium 133 system had no heatsink attached when I got it, ran it for 2 years that way not a single crash
according to the pentium 133 box.. it came with a fan + heatsink??? so that doesnt really match up with what your saying
http://www.oldschooldaw.com | vintage PC/MAC MIDI/DAW | Asus mobo archive | Sound Modules | Vintage MIDI Interfaces
AM386DX40 | Asus VL/I-486SV2GX4 (486DX2-80) | GA586VX (p75) + r7000PCI | ABIT Be6 (pII-233) matroxG400 AGP
Ugh. I'd never try to run any socket 7 CPU without a heatsink. I did, however, run a few without any fan. It seemed that a Tillamook MMX or a Little Foot K6 or older (Chomper) K6-2 can take it without a problem - at least they did stable and not exactly hot. Undervolted and underclocked as much as I could. I chose these CPUs because I expected them to take 2.0V and had a 2x multiplier. They did so, and the system consumed the same or less power than a typical 486.
I think they were perfectly safe with passive cooling (a rather beefy Socket A cooler).
I also recall the stock Pentium 100 systems back in my highschool not having fans, only tall socket 7 heat sinks. They must have been hot.
Shame on us, doomed from the start
May God have mercy on our dirty little hearts
I assumed it was quite common to see CPUs of this era with only a passive heatsink, so long as the CPU is close to the PSU fan. I have a couple of Gateway systems like this, one P233 MMX and one PIII 450, both are super quiet.
VRM still needs cooling, if not for anything else, then longevity. Same thing for video cards.
wrote:wrote:My Original Pentium 133 system had no heatsink attached when I got it, ran it for 2 years that way not a single crash
according to the pentium 133 box.. it came with a fan + heatsink??? so that doesnt really match up with what your saying
Only boxed Pentium CPUs had a glued on heatsink and fan. OEM systems had whatever the manufacturer wanted to. However it was still possible to buy tray CPUs without heatsink and fan directly from computer stores. At least over here that was pretty common. These were usually cheaper than the boxed versions, so were a lot more popular with people that assembled their own PCs.
Remember my friend got his Pentium100 based PC without CPU heatsink at all. And it worked stable in all games we played.
- C&C: Red Alert Archive project
QDI Titanium IIB / P55C-200 / 64Mb RAM / 8Gb IDE CF / ATi 3D Rage Pro 8Mb / Sound Blaster CT2890 / 3D Blaster CT6670