Are 8-bit IC’s at risk of overheating

[gs80065xe]

Watched a recent Commodore video on YouTube by a man called Adrian Black. While demonstrating installing an Arduino ROM switcher, he mentioned he added heat sinks to many of the ICs as a precaution. That has me wondering if if the Atari ICs ever get too warm at room temperature?

 

He shows them around 8 minutes

Edited April 30, 2019 by gs80065xe

[flashjazzcat]

http://atariage.com/forums/topic/221132-installing-a-laptop-fan/page-2?do=findComment&comment=2924148

[Rybags]

Most of the legacy Atari chips only ran at 1.79 or 3.6 MHz, and only at 5V.

I believe the C64 by default has some heatsinking for VIC and/or SID. Essentially a bent over piece of metal that sits on top of the chip and is connected to the foil/cardboard RF shielding.

 

Supposedly the PLA on C64 can get hot too (and is a common failure point).

Also note that SID uses higher voltage for part of it's audio generation which could be a contributor.

 

Freddie I believe runs at full 14 MHz though it's just a fairly simple logic array in similar form to the PLA on the C64.

 

Atari at least until the XE and ST came along generally engineered their machines to be very reliable and if a heatsink was needed it would have been included - generally you'll see them on voltage regulators and some other non-IC components.

 

Bottom line for both types though - we've had 30+ years of them being retrocomputers and a pretty decent knowledge base exists of common failing parts and I don't think any of the Atari 8-bit ICs are prone to premature heat-death.

[+DrVenkman]

In my experience with my particular 400/800/XL machines, the only way to overheat a chip would be with a soldering iron. They just don’t get THAT hot.

[flashjazzcat]

Only one thing may require cooling.

 

[+x=usr(1536)]

I have a vague recollection of some (specifically) 800 owners running with the top cover off to allow heat to escape more effectively.

 

Why this was necessary eludes me; mine never seemed to require it.

[+DrVenkman]

I have a vague recollection of some (specifically) 800 owners running with the top cover off to allow heat to escape more effectively.

 

Why this was necessary eludes me; mine never seemed to require it.

The original design OS ROM and RAM boards were encased in plastic housings and could overheat with extended run times, especially in a warmer environment.

[wt808]

With the PDIP-40s right next to each other inside the 5200 Compact and 5200 Handheld, I had to come up with this:

They do not run reliably without when placed in a small enclosure with no vents.

 

[+x=usr(1536)]

The original design OS ROM and RAM boards were encased in plastic housings and could overheat with extended run times, especially in a warmer environment.

 

Thanks; that's ringing bells. Mine had the plastic ROM and RAM modules, but it never seemed to be an issue despite being left running for days on end at times.

 

Then again, I was using it in a room that rarely went above about 68°F. That probably had a lot to do with it.

[_The Doctor__]

You can heat up the cpu and 1 other chip especially when in hot not air conditioned settings.... taking off the shield does little for this, in factory settings we ran a piece of curved spring steel (heat sink grease and riveted to the shielding and adjusted to ever so slight contact the affected chips, heat sink grease applied there as well.. It worked just fine in a cast iron pipe factory in that configuration. I doubt anyone has that sort of environment. It's possible that no air conditioning in apartment houses during the summertime in a good number of southern areas or depending on your orientation, as we approach the equatorial areas, it could get pretty hot as well. It won't hurt to add heat sinks.

 

Early 800's could get pretty toasty in the ram area, You could heat sink the ram, some people de-cased them, folks even drilled holes in the aluminum and added a fan. All this and more has been done over time.

Edited April 30, 2019 by _The Doctor__

[Gunstar]

I watched this video or a similar one on the C64 a while back and caught on to the heat-sink mods then. I've thought of adding heat sinks to my 1200XL's IC's and maybe adding a small fan (it does get toasty with the original PSU circuit and big heatsink in there). Though I have never had any issues with overheating on it or any other of my 8-bit hardware, including my 800 with it's enclosed ram (not that I've left it on for hours at a time anyway). But I thought it looked really cool, certainly couldn't hurt, and maybe even help with the over-all longevity of these 30+ year old IC's.

[flashjazzcat]

http://atariage.com/forums/topic/221132-installing-a-laptop-fan/?do=findComment&comment=2913041

[spinnaker15136]

I regularly operate 8 bit pics at 20MHZ and up and they don't even begin to get anywhere close to warm.

[+Stephen]

http://atariage.com/forums/topic/221132-installing-a-laptop-fan/?do=findComment&comment=2913041

Some people like to "replace all the caps" too

[+bob1200xl]

'Commercial' chips are rated to 70C, 'industrial' chips are rated to 125C. You need to realize that is at the die buried in the chip carrier, not at the surface of the chip. Still, 70C is pretty hot.

 

What you need is a logic analyzer that can trace the system operation. This will help you determine where things go South.

 

Bob

[Kyle22]

In a 'normal' environment, I don't think we have an overheating problem. The only way I see it happening is as The Doc mentioned, in an industrial setting.

Especially if you have taken the covers off of your OS and RAM cards. (Or have an Incognito)

[Sugarland]

Watched a recent Commodore video on YouTube by a man called Adrian Black. While demonstrating installing an Arduino ROM switcher, he mentioned he added heat sinks to many of the ICs as a precaution. That has me wondering if if the Atari ICs ever get too warm at room temperature?

 

He shows them around 8 minutes

 

The short answer is that overheating is not a problem for A8's, for the most part. The 800 ram/os cards were of some concern. There is a video on YouTube that examines the thermal effectiveness of the "shield" (cough) touching the C64 chips and they found it did not dissipate heat.

 

That said, I think it might be a good idea to heatsink the warmer A8 chips like the CPU and GTIA to hypothetically prolong life. However if a metal heatsink gets loose in there it could short something which could have the opposite effect. If a system is running 24/7 it might be a good idea to cool it, imo.

[Kyle22]

It can not hurt, but for most, it is not necessary. If you are worried, go ahead, just be careful not to short anything.

[_The Doctor__]

just make sure to use thermal conductive materials, not the typical double sided sticky tape or standard glues (sadly I've seen this sort of things in all kinds of equipment)... it will make the chips internally hotter still... paste and clips or pressure. or the sticky thermal pads etc....

[ChildOfCv]

To put it in perspective, the IBM 5150 (the first of their line of Intel 16-bit PCs) was engineered like a tank.

 

No heat sinks. Except inside the power supply.

[_The Doctor__]

The Atari's are tough computers. I'd say some of them are indeed tanks.

[Gunstar]

just make sure to use thermal conductive materials, not the typical double sided sticky tape or standard glues (sadly I've seen this sort of things in all kinds of equipment)... it will make the chips internally hotter still... paste and clips or pressure. or the sticky thermal pads etc....

Yikes! I've never come across that before!

 

You want to secure the heat sink, but not permanently to the IC's. Lithium grease should be used in between the heat sink and the IC. the heat sink should be secured to the motherboard somehow, if the grease isn't enough to hold it in place.

Edited May 1, 2019 by Gunstar

[_The Doctor__]

pc's and related electronics are like a box of old fashioned chocolates.... you never know what you're going to find inside.

[ricortes]

Rockwell 6502C made in Mexico are notorious for bad behavior. They will sometimes boot just fine and as they heat up, go south. Can happen anytime from ~1 to only on the hottest days of summer. You can just spray them with chip coolant to see if they recover/hit reset.

 

Rest of the chips I've run up to a 2.12 MHz clock speed and they work fine within the obvious limits of sound pitch, SIO communication, and video. Somewhat over engineered and good quality control considering this was 40 years ago.

 

Would be nice if we had faster replacements for hacking. At some point I see the argument it is emulator vs a real Atari. I'm not sure of the bucket list of things that would have to be done, but things like a POKEY with 1 Mbit SIO or compatible video expansion that you could use for specific applications at speed like an assembler would be good. It would save having to run at multiple clock speeds to keep everything happy.

[+x=usr(1536)]

To put it in perspective, the IBM 5150 (the first of their line of Intel 16-bit PCs) was engineered like a tank.

 

No heat sinks. Except inside the power supply.

 

True, but as an example to the contrary: the Apple ///. Some versions of the story state that ICs would pop out of their sockets due to overheating caused by a lack of forced ventilation; another version states that it was due to poor manufacturing processes.

 

My money's on heat being the issue - if the machine can have all of the affected components reseated, then start experiencing the same problems again after being on for x number of hours with the same fix again mitigating the issue, it's likely not solely a manufacturing problem. I can accept that that may have been a contributing factor, but with heat as the trigger for it.