If you plan to put the board inside the case, install an 80x80mm jet fan at the front.

https://www.delta-fan.com/AFB0812SH.html

They didn't need a fan back in the day to run 24/7 and all these years later they still work perfectly fine so I would say they simply dont need a fan for airflow other than what the case provides.

Passively cooled 386/486 CPUs are damn near bullet proof, I cant remember hearing any stories of heat killing them unless the user was doing it on purpose.

zuldan wrote on 2023-12-22, 12:11:

I have a couple of overdrive CPUs that get a little too hot for my liking. For a standard 486 you can simply buy a heatsink and fan that clips onto the CPU (https://www.ebay.com.au/itm/185840766419), however the overdrive already has a heatsink so the heatsink clip won’t work. Is there a way to attach some sort of fan to it?

Hi there, you can try out different things if you like.

a) If the fins of the cooler (heat sink) are long enough, you can screw the fan directly onto the CPU.
Use plastic screws if you're worried of damaging something.

You can also rotate the fan, if needed, so that merely two mounting holes are over the heat sink fins.
That might be necessary, depending on the size of the fan.

Ideally, the fan should suck the heat away from the CPU. But in reality, it doesn't really matter that much, aslong as there's an airflow.

b) You can use a bit (just a bit, a small line or a few dots) of hot glue on the side of the CPU/slot to attach the fan.
If done carefully, it's not a permanent installation and the fan can removed again later on.

Btw, a bigger and slower fan has some advantages over a tiny CPU fan.
a) It draws less current
b) It's more efficient, moves more air
c) It is quiter, usually (at same performance if the tiny one)

The power requirements of small high-speed fans is not to be underrated.
In a small 286 PC (pizza box) with a custom PSU, using such a small CPU fan would technically overburden the PSU.

I'm speaking under correcting here, of course.
That's merely what came to mind. I've mounted fans that way before, but there might be more elegant solutions.

PS: There's also a software-solution to keep the CPU temperature down during idling.

DOSIDLE, an utility, does put the CPU to sleep if there's nothing to do. It uses HLT instruction directly or APM BIOS.

For Windows 3.1, there's WQGHLT (uses HLT).
For Windows 9x, there's AmnHLT.

They're primarily being used to keep VMs from causing a 100% CPU load all the time.
But they're also working fine with real hardware.

DOSIDLE can be fine tuned, even, so that the idle loop detection has the correct intensity (for games etc).

Edit: Here's an example picture of a 486 era cooling fan that's screwed directly to the heat sink..
The screws do go between the fins, while the little plastic clamps hold onto the CPU.

Source: http://www.cjemicros.co.uk/micros/individual/ … de=VAR-CPUHSFAN

You dont put a fan on it. If you want some cooling then you need to get some airflow into your case.

Well yes and no, I think. Generally speaking, it depends.
The heat sink must also allow a given heat transfer.

If the heat sink is too small (under powered), it may won't dissipate heat properly (not fast enough).

A way stronger airflow than normal would be required to compensate for that deficiency.
That's why those angry little CPU fans exist, I suppose. 🤷‍♂️

Anyway, I don't mean to disagree here.
I'm also for big fans and a good cooling in the chassis.

Edit: Again, it's just a general statement. I don't know much about the overdrive processors here.
It may or may not need an extra fan, not sure. 🤷‍♂️

Jo22 wrote on 2023-12-22, 14:39:

Well yes and no, I think. Generally speaking, it depends. The heat sink must also allow a given heat transfer. […]
Show full quote

Well yes and no, I think. Generally speaking, it depends.
The heat sink must also allow a given heat transfer.

If the heat sink is too small (under powered), it may won't dissipate heat properly (not fast enough).

A way stronger airflow than normal would be required to compensate for that deficiency.
That's why those angry little CPU fans exist, I suppose. 🤷‍♂️

Anyway, I don't mean to disagree here.
I'm also for big fans and a good cooling in the chassis.

Edit: Again, it's just a general statement. I don't know much about the overdrive processors here.
It may or may not need an extra fan, not sure. 🤷‍♂️

They come with a heatsink epoxied to them with that awful Thermal Epoxy Intel and others loved using so there isn't anyway to remove it without possibly damaging the ceramic of the chip, that said so long as the case has sufficient airflow the CPU itself doesn't require any other form of cooling than its built in heatsink. It will happily run 24/7 for years with nothing more than some light dusting from time to time and a new case fan which will die well before the CPU ever will.

I can say with 100% certainty that the CPU will happily outlive the very system its put into, even with nothing more than passive cooling, the DX2-66 Overdrive is pretty dang hard to kill.

One of the finest CPUs Intel ever made IMHO.

I have the hexact cpu in my collection and i can say that using it, become hot like a Sun, the only think is that intel do not think that cpu will survive so many years, they think that this dx2 must be replaced and recycled in 3/4 years maximum. Using it for years and years must destroy it, i can't belive it can survive for years at this temperature. If you touch it literally you burn your finger after only 10 minutes.

As @Jo22 mentioned, use DOSIDLE for DOS and Rain for Windows.

https://github.com/DOSAlliance/UTIL-DOSidle/releases
http://www.benchtest.com/rain.html

AlessandroB wrote on 2023-12-22, 15:31:

I have the hexact cpu in my collection and i can say that using it, become hot like a Sun, the only think is that intel do not think that cpu will survive so many years, they think that this dx2 must be replaced and recycled in 3/4 years maximum. Using it for years and years must destroy it, i can't belive it can survive for years at this temperature. If you touch it literally you burn your finger after only 10 minutes.

its a 486, it was designed to run like that with passive heatsink cooling in a case with little to no air flow for years, its not a modern CPU with a 250watt TDP that will die the moment it hits 105c. I grew up with 386 and 486 machines and they are unkillable from heat so long as they have some form of passive cooling or even just air flow from a case fan.

The 486 DX-33 doesn't even need passive cooling, I know it seems foreign to run them that way but its perfectly fine and what conditions they were designed around, adding a case fan near them for a bit of air flow is all that's really needed. IIRC it wasn't till the DX2- 80 and DX4-100 that they needed some form of small fan on them to supplement the passive cooling.

Just for reference 50c will burn your fingers and a cup of hot coffee will also burn your fingers and both are well below 100c, the 486 DX2-66 is fine up to ~85c with nothing more than its heatsink and basic case fan airflow. (It wont ever hit that temp, thats what the heatsink is for)

As for what Intel wanted . .well these CPUs were normally used in PCs at a time when you didn't just go out and buy an upgrade every 2 - 3 years. .CPUs were very expensive at the time so Intel didn't have a reason to do what you suggest and the 486 CPUs were built to last under a variety of stressful conditions . .like being in a dusty case with little to no airflow in some server closet running at high temps.

Trashbytes wrote on 2023-12-22, 15:55:

As for what Intel wanted . .well these CPUs were normally used in PCs at a time when you didn't just go out and buy an upgrade every 2 - 3 years. .CPUs were very expensive at the time so Intel didn't have a reason to do what you suggest and the 486 CPUs were built to last under a variety of stressful conditions . .like being in a dusty case with little to no airflow in some server closet running at high temps.

Intel.. They didn't even get their technical manuals right (afaik). 😁

Anyway, that's the point: back-then.
These days, we nolonger want to do things like it was common back then.

Back then (early-mid 90s), we "abused" our PCs because they weren't meant to last, anyway. They were just tools.
Technology was evolving so fast, that an average PC was obsolete quickly.
No need to keep it lasting for more than ~5 years.

Nowadays, we will rather try to "spare" our PC hardware whenever possible.
That's why we think about heat sinks and cooling fans, I believe.

And it's not too overkill, either, maybe.
Early DRAM and logic ICs can get quite hot.
I've often thought about buying heat sinks for my 2 MB EMS board from the mid-80s (AST RAMPage 286).

Btw, the early ceramic 286 proccesors saw a need for a heat sink, too.
The higher clocked LCC models (8 to 10 MHz) got quite hot, at times.

Source: (link)

A bit like the later released Cyrix/TI 486DLC-40 proccesor, maybe.
It got uncomfortably hot in comparison to an am386-DX which got merely hand warm.

Edit: Whats also endangered by heat is the on-board cache, maybe. Cache RAM is very sensitive to overclocking (overheating).
A good cooling can increase lifetime of a cache.

Edit: It's not necessarily bad if a ceramic package gets hotter than a plastic package.
Ceramic somply can conduct heat better, maybe.
It's generally good if a heat buildup is being prevented. Likewise, a heat sink helps transferring/dissipating the heat.

Edit: I forgot to add. I'm merely thinking out loud here, I don't mean to argue.
I think it's true that back then it was common to have 386/486 systems up to ca. 40 MHz running passively cooled. With and without heat sinks.
I've seen enough old PCs from the inside to confirm that, I believe.

I'm not suggesting to not add the extra case fan or bigger heatsink but rather realise its not 100% required for the dx2-66 to keep on trucking for the next 20 years, heck that CPU will very likely outlive its owner. The extra cooling would however be of a massive benefit for the VRM which was usually placed super close to the CPU in these older boards and the other components would also benefit from more air flow in the case.

I see other CPUs being mentioned here that are not a bulletproof Intel DX2-66 Overdrive and I would suggest that you look at each CPU on its own, Cyrix, AMD, WinChip etc were not always built for the same level of abuse as Intel CPUs of the time.

Heck I guess I'm just saying to not worry too much about the Overdrive getting hot to touch, throw in some extra case cooling and itll be fine, no real need to go overboard and spend cash on fancy cooling for it. (Certainly dont mod the CPU heatsink to fit a fan, I dont agree with damaging stuff to fit cooling)

The technical recommended limit is ~85C, and the max physical temperature before damage occurs is ~110C.

Heatsinks are meant to store heat but they can't actually take it away without strong air flow.

If you can confirm it runs under 85C at constant full load in summer.. then it's fine

dominusprog wrote on 2023-12-22, 12:44:

If you plan to put the board inside the case, install an 80x80mm jet fan at the front.

https://www.delta-fan.com/AFB0812SH.html

The fan mounting holes in all my cases are 70mm by 70mm apart. Any idea if this fan would fit that?

So I measured the temp of my DX2-66 Overdrive and it sits at 57c. I haven’t measured my DX4-100 Overdrive so I’m guessing that will be just over 60c. I guess the temps seem to be ok but I think I’ll get a fan for the case. Would like these girls to last as long as possible.

zuldan wrote on 2023-12-23, 08:27:

So I measured the temp of my DX2-66 Overdrive and it sits at 57c. I haven’t measured my DX4-100 Overdrive so I’m guessing that will be just over 60c. I guess the temps seem to be ok but I think I’ll get a fan for the case. Would like these girls to last as long as possible.

heh, I suspect they will out last all of us, might get to the point there are no working boards for them though, I find the motherboards die first even with overkill cooling.

Which makes me wonder how hard it would be to engineer a cheap replacement motherboard using salvaged chipsets.

zuldan wrote on 2023-12-23, 08:23:

dominusprog wrote on 2023-12-22, 12:44:

If you plan to put the board inside the case, install an 80x80mm jet fan at the front.

https://www.delta-fan.com/AFB0812SH.html

The fan mounting holes in all my cases are 70mm by 70mm apart. Any idea if this fan would fit that?

I tend to use Noctua high CFM fans as replacements for case fans, Noctua makes plenty of fans in smaller sizes which has been a life saver.

That said something like the fan you listed would work well enough, just be sure to check if the motherboard supports 12v or 5v fans ...Ive made that mistake a few times and 12v fans dont run terribly good on 5v. (derp if its a 486 it'll be hooking up to Molex which is 12v..dont mind me, senility is catching up to me 🤣 )

Bit off topic since it's a DX4 overdrive. I pushed two silicon 'Vibration Compensators' in the heatsink, thanks to spacing for the voltage regulator.

PD2JK wrote on 2023-12-23, 08:49:

Bit off topic since it's a DX4 overdrive. I pushed two silicon 'Vibration Compensators' in the heatsink, thanks to spacing for the voltage regulator.

DSC_7073.JPG

Not saying its a bad thing but I would be worried about the silicon insulating the regulator and causing it to overheat, I guess it depends on case airflow and how thick the silicon pads are.. IIRC they could take a bit of heat but its a voltage regulator so I would be cautions for anyone else wanting to try this.

Trashbytes wrote on 2023-12-23, 09:04:

PD2JK wrote on 2023-12-23, 08:49:

Bit off topic since it's a DX4 overdrive. I pushed two silicon 'Vibration Compensators' in the heatsink, thanks to spacing for the voltage regulator.

DSC_7073.JPG

Not saying its a bad thing but I would be worried about the silicon insulating the regulator and causing it to overheat, I guess it depends on case airflow and how thick the silicon pads are.. IIRC they could take a bit of heat but its a voltage regulator so I would be cautions for anyone else wanting to try this.

No worries, it can breathe. 😀
Even gets some airflow from the fan.