• Please review our updated Terms and Rules here

when did cpu's start really needing thermal grease

I own an Intel Overdive with its attached heatsink and its clocked for 33mhz. That's the slowest thing can recall seeing aside form 286's as well with heatsinks.
 
Chuck(G) said:
I have never seen an 8085 with a heatsink mounted on it. Most of those, particularly the HMOS II varieties were PDIP anyway, no? What good would a heatsink have done (epoxy has terrible thermal characteristics)?
Click to expand...


Early Compupro S-100 CPU-8085/8088 Dual Processor Boards had heatsinks glued to the tops of the chips. I used to see some early IBM Clone motherboards with heatsinks glued to the tops of the 8088 processor chips as well.

I think the earliest 8085 and 8088 processor chips were white ceramic. The ones that Compupro used at the beginning were Intel and they were the thickdark colored what I thought were ceramic processors (colored like EPROMS), definitely NOT PLASTIC. Later Intel, NEC, and AMD processors were plastic, but by that time the heatsinks were long gone.
 
The NMOS 8085s were CerDIP; the HMOS ones were PDIP. Still, we used the 8085 since its pre-production steppings and never needed a heatsink. What the heck was Bill Godbout doing to the poor thing to require one? How many boards did he use it on? I can only remember the 85/88 board.
 
Chuck(G) said:
The NMOS 8085s were CerDIP; the HMOS ones were PDIP. Still, we used the 8085 since its pre-production steppings and never needed a heatsink. What the heck was Bill Godbout doing to the poor thing to require one? How many boards did he use it on? I can only remember the 85/88 board.
Click to expand...

I don't know how many processor boards got the glue on heatsink treatment.

I saw them on early dual processor 8085/8088 boards and early 8086/8087 processor boards.
 
Well, I think the point he was making was lots of crap would do the same but duration and time can throw off results of someone that only tests the application one time. Similarly I found that article when I thought I'd search magazines and books for the earliest reference via google. Unfortunately the earliest reference I found was an electronics book that didn't was from 1978 which isn't when it was popular. Popularity in magazines really ends up starting in mid to late 90s though. Here's another earlyish reference from 1996 calling it important on high performance cpus like the i486 or dx2/dx4/overdrive.

Magazine results are sorta interesting though. One reference in 92 but from an end user comment I think, then 96, then the rest are year 2000 when the media caught on that we were apparently doing this often. That of course is only based on whatever magazines google has archived though.
 
Selenium rectifiers, which became common in TV sets in the late 1940s, were made of selenium discs separated by metal plates, which act as a heatsink.

3756304358_e490c14879_z.jpg
 
Before selenium were copper-oxide rectifiers with a similar plate structure. High-power transmitting tubes have had some sort of heatsinking at least since the 1930s. here's a reflex klystron from 1937:

PIC00041.jpg
 
Chuck(G) said:
Before selenium were copper-oxide rectifiers with a similar plate structure. High-power transmitting tubes have had some sort of heatsinking at least since the 1930s. here's a reflex klystron from 1937:



PIC00041.jpg
Click to expand...

The 2C39A/7289 (lighthouse tube) RF finals in the Collins AN/ARC-27 UHF transceiver comes to mind. Guess I just date stamped myself there - eh.
 
Last edited:
Yeah, I remember the lighthouse tubes as well. I went looking for an image of the RCA 827-R, but I couldn't find one. I did find this one of a similar Westinghouse tube (power tetrode). Note the copper finned heatsink--the tube was made for forced-air cooling, circa 1940 (at least it's in my 1940 tube manual):

5736.jpg


Here's a water-cooled triode from the early 1930s:

P9119752.jpg


This was long before the Eimac glass- or ceramic-and-metal power tubes.
 
Chuck(G) said:
Before HP turned into the disaster is is today, they'd use plastic ducting to route the airflow from the power supply fan around the CPU heatsink. It made for a very quiet machine. I can't stand those little fans that eventually fail all too soon, but now it seems that I have little choice.

For years, I've used Permatex anti-seize compound instead of the more esoteric thermal greases. It appears to work just as well and is considerably cheaper. There's quite a bit on the web that demonstrates its excellent thermal characteristics.
Click to expand...

Well you could always get a nice IBM (like an intellistation 9229), they use quality fans and typically favour BTX these days.
My favourite thermal paste is arctic silver, and I also have generic vantec thermal paste for everything else that I got free with some fans I ordered.

Chuck(G) said:
Heatsinks have been used on semiconductor power devices such as transistors, diodes, thyristors and regulators for as long as those devices have been around. But I suspect that the CPU on your C64 is not one of them.
Click to expand...

Yep, the C64 has thermal paste on the CPU, SID chip, and others. I can affirm to this as I took mine apart and cleaned it up. Didn't bother to reapply thermal paste yet until I replace the electrolytic caps.
I'm surprise no one else knows about this?

I think most chips should have heatsinks, the fact that early ones lack them does not mean they're running cool, they get quite warm. Kind of similar to like how most system boards come with undersized aluminum heatsinks on the FSB, when that chip needs much, much more.
 
EverythingIBM said:
Yep, the C64 has thermal paste on the CPU, SID chip, and others. I can affirm to this as I took mine apart and cleaned it up. Didn't bother to reapply thermal paste yet until I replace the electrolytic caps.
I'm surprise no one else knows about this?
Click to expand...

Well, I sure didn't and evidently CBM didn't either:

commodore_64_motherboard_1982_1992.jpg
 
I have indeed seen Commodore computers where the RF shield over the motherboard had metal fingers that were angled down to rest on top of some of the chips to act as a heatsink for them, and upon taking off the RF shield I did see thermal paste smeared on those chips. I've taken apart the VIC-20, C16, C64, C64C, and C128, so I don't remember which one(s) had it, but I've definitely seen it. Obviously anyone who disassembles a Commodore to take a hi-res photo of the motherboard would wipe the grease off to make the chip markings visible, so those photos don't prove anything.
 
I've seen the heat-sink grease and metal tabs on a C128, but I recall (it's been a long time since I had one open) the C64 having a tinfoil-on-cardboard RF shield. It'd make no sense at all to smear heatsink grease on that.
 
Are you sure guys aren't getting confused with the SX-64, which had a glued heatsink for the VIC chip ?
The 64 had a metal tab touching the VIC, but that was it. Now, people over the years did try to heatsink the PLA, SID, VIC, and probably one or two others to help them run cooler, but I don't think C64's ever shipped with a heatsink.
I seem to remember the tinfoil shield RF being a hassle since it interfered with convection cooling.
patscc
 
No. My C64 had thermal paste underneath the shielding which acts a heat sink. In fact, I still have to clean off the smudged residue with some isopropyl alcohol when I put some fresh paste on after recapping. The paste was white and still quite liquified for its age.

A simple google search will show other threads of people talking about removing the paste off of the chips.

If you found no paste underneath the shield, then it was removed by someone else. The paste was added on at the factory (I was the first to open my C64, it sat largely unused in my grandma's basement).
 
You must log in or register to reply here.
Back
Top