Why extremely heavy and expensive aluminum castings for RF shielding in the 400 and 800?

[Winston2022]

Why wasn't cheaper (I assume) and lighter sheet metal used which could have been used even for the expansion card slot section of the 800?

[Faicuai]

37 minutes ago, Winston2022 said:

Why wasn't cheaper (I assume) and lighter sheet metal used which could have been used even for the expansion card slot section of the 800?

 

Because the FCC standards of design-time dictated so. On the other hand, my 800's internal shielding won't rot and as crap-looking as my 800-XLs tin-foil shielding...

 

Much better quality and better isolation on the 800. Simply put, built to a higher standard.

Edited November 29, 2022 by Faicuai

[zzip]

Yes, the FCC rules (designed to protect other devices against RF interferance) changed sometime after the 400/800 and enabled the cheaper computers like the C64 and XL line to be produced.

[DrVenkman]

1 hour ago, Winston2022 said:

Why wasn't cheaper (I assume) and lighter sheet metal used which could have been used even for the expansion card slot section of the 800?

The FCC rules for home computing devices were not yet finalized at the time the 400/800 were designed so Warner-era Atari took a very conservative (and expensive) approach at the outset so they would not have to do either a recall or a redesign in case the final rule-making resulted in the stringent requirements originally thought likely. 

[+MrFish]

The 400 & 800 were initially targeted for use in M1 Abrams tanks.

 

[bfollowell]

1 hour ago, MrFish said:

The 400 & 800 were initially targeted for use in M1 Abrams tanks.

 

From the way they're made, I can believe it!

 

[Winston2022]

16 hours ago, Faicuai said:

"Because the FCC standards of design-time dictated so."

I'm aware of that and the later change to the standard, but what I don't understand is how a properly interconnected and grounded ferrous sheet metal faraday cage was inadequate to meet the standard.  Anyone have any Atari documents about this?

[_The Doctor__]

The cage keeps the Atari from radiating outward, but it also keeps the Atari from being radiated inward. A little more than meets the eye here.

[mickster]

FCC required it at the time.  Also during the design  of the atari computer it was not about being the cheapest to market.

[TGB1718]

They were built that way so if the Russians nuked us the EMP would have no effect on the 800/400 so we could still play Missile Command

 

[Faicuai]

1 hour ago, _The Doctor__ said:

The cage keeps the Atari from radiating outward, but it also keeps the Atari from being radiated inward. A little more than meets the eye here.

Ding, ding, ding! We got a winner! 😉💪 

 

And quite a potential benefit with apps like Fujinet, or those many (horrible) shots that I see with 400s and 800s in shows and homes being placed right on top of EMI-rich CRTs !! 🤯🤯

 

 

Edited November 30, 2022 by Faicuai

[mickster]

also, due to the high cost of computers, the design team did not want the computer to feel light.  If the machine was heavy it would be perceived to be worth more.  This was also true on the atari 2600

[Geoff Oltmans]

Would be interesting to see the difference in radiated emissions of the cast cage vs a sheet metal cage. Personally I'm skeptical that they'd be much different.

Edited November 30, 2022 by Geoff Oltmans

[DrVenkman]

1 hour ago, Geoff Oltmans said:

Would be interesting to see the difference in radiated emissions of the cast cage vs a sheet metal cage. Personally I'm skeptical that they'd be much different.

You would be surprised. It's not *just* the thickness of the casting - it's the dimensions of the ground plane of the mainboard; it's the number of fasteners used to ensure a solid connection between each side of the case to the ground plane around the periphery of the case; it's the hundreds of tiny filter capacitors all over the board and at every interface signal line; it's the enormous ferrite ring and the number of windings around it used for the RF output cable, which was also a shielded coaxial cable ... every single aspect of the design contributes to the RF signature produced by a computer of the era, as well as how susceptible that computer is to emissions from other electronic devices nearby. 

 

As I said above, the RF regulations for consumer electronic devices were not final at the time the 400/800 were being designed in 1978 through mid-1979, so Atari was EXTREMELY conservative. There are some indications, depending on which oral history you listen to, that other, larger and better-established electronic companies like Tandy and Texas Instruments, who were aware Atari was entering the home computer market, were lobbying the FCC for extremely conservative (e.g. stringent) requirements. Atari didn't want to be sandbagged and went with a design that would meet any worst-case scenarios. 

[Faicuai]

6 hours ago, mickster said:

also, due to the high cost of computers, the design team did not want the computer to feel light.  If the machine was heavy it would be perceived to be worth more.  This was also true on the atari 2600

Pfft!

 

This is a trivial: drop your iphone on 5G mode inside of the 800 expansion bay, seal it, and try calling... even better start the call before tucking it inside...

 

Do the same with your 800XL in any place it fits underneath the shield.

 

All of the above with the machines powered-down, of course. In the above test, you will see both the results of incoming and radiating RF (high-frequency) emissions.

[Winston2022]

20 hours ago, DrVenkman said:

You would be surprised. It's not *just* the thickness of the casting - it's the dimensions of the ground plane of the mainboard; it's the number of fasteners used to ensure a solid connection between each side of the case to the ground plane around the periphery of the case; it's the hundreds of tiny filter capacitors all over the board and at every interface signal line; it's the enormous ferrite ring and the number of windings around it used for the RF output cable, which was also a shielded coaxial cable ... every single aspect of the design contributes to the RF signature produced by a computer of the era, as well as how susceptible that computer is to emissions from other electronic devices nearby.

 

As I said above, the RF regulations for consumer electronic devices were not final at the time the 400/800 were being designed in 1978 through mid-1979, so Atari was EXTREMELY conservative. There are some indications, depending on which oral history you listen to, that other, larger and better-established electronic companies like Tandy and Texas Instruments, who were aware Atari was entering the home computer market, were lobbying the FCC for extremely conservative (e.g. stringent) requirements. Atari didn't want to be sandbagged and went with a design that would meet any worst-case scenarios. 

I'm aware of all of that and all of that OTHER THAN the use of an expensive non-ferrous casting is standard procedure for RFI suppression and would be expected.  I am wondering HOW the cast aluminum case could be any better than a properly interconnected and grounded ferrous sheet metal faraday cage.  I've never seen any detailed description or discussion of their RFI prototyping trail and error.  On the RFI issue, I've heard in interviews that it was largely because of Sears that they went overboard nuts and that one of the engineers inadvertently fooled TI about a loophole in the FCC RFI standard that fiber optics could be used to exploit which, when the FCC wouldn't buy it, really bit TI in the butt.

[Winston2022]

Atari 2600 6-switch - Rev 1 - Sep 1977 release

 

Aluminum casting shielded:

 

https://the-liberator.net/site-files/retro-games/hardware/Atari-2600/atari-2600-odd-6-switch-woody-rev-1-1978-sunnyvale/Atari-6-Switch-Woody-Sunnyvale-CA-ODD-ONE-REV-1-Boards-008-Motherboard.JPG

 

https://the-liberator.net/site-files/retro-games/hardware/Atari-2600/atari-2600-odd-6-switch-woody-rev-1-1978-sunnyvale/Atari-6-Switch-Woody-Sunnyvale-CA-ODD-ONE-REV-1-Boards-009-Motherboard.JPG

 

Bally Astrocade - Dec 1977 release - sold through 1985

 

CPU: Zilog Z80 clocked at 1.789 MHz
Memory: RAM: 4k-64k (with external modules) RAM, 8k ROM

 

Crude sheet metal shielding:

 

 

--------------

 

Atari 400/600 - Nov 1979 release - cast aluminum shielding to the extreme.

 

Magnavox Odyssey 2 - Feb 1979 release - sold through Mar 1984

 

CPU: Intel 8048 clocked at 1.79 MHz.
Memory:    CPU-internal RAM: 64 bytes
Audio/video RAM: 128 bytes
BIOS ROM: 1024 bytes

 

Stamped sheet metal box shielding:

 

https://1.bp.blogspot.com/-J12ceYkJlvg/Vwp5__eC1QI/AAAAAAAADtM/RG9ml05g8mEA6suQ0O-iWC2G6bzuzvtBA/s1600/Magnavox%2BOdyssey%2B2%2BComposite%2BUpgrade%2BMod%2BKit%2BRetroFixes%2B4.jpg

Edited December 1, 2022 by Winston2022

[DrVenkman]

You need to go dig through the evolution of the nascent FCC regulations - a good law library would come in handy because there’s a decent chance most of the regulatory history (e.g., notices of proposed rule-making, compilations of public comments, etc) haven’t been digitized or have been only partially. 

[ClausB]

3 hours ago, Winston2022 said:

Atari 400/600 - Nov 1979 release - cast aluminum shielding to the extreme.

The case design was done by 1978. Also the metal frame is structurally integral to the case.

[Winston2022]

On 12/1/2022 at 4:30 PM, ClausB said:

The case design was done by 1978. Also the metal frame is structurally integral to the case.

The case was designed around the cast frame.  The case design simply could have been done differently if sheet metal had been used.

 

Note that even the 2600 also used cast aluminum shielding, the only game console from that era I've found that did.  Based upon that and the sheet metal shielding examples I've shown from game systems from the same time period which were approved for release at the same time as the Atari 8-bits with the same CPU clock frequency because of the video sync requirement, game systems which continued to be sold for years afterwards,  I think that maybe the Atari's extreme RFI shielding may have been corporate familiarity/inertia/philosophy more than a need due to uncertainty.

[Winston2022]

On 12/1/2022 at 1:00 PM, DrVenkman said:

You need to go dig through the evolution of the nascent FCC regulations - a good law library would come in handy because there’s a decent chance most of the regulatory history (e.g., notices of proposed rule-making, compilations of public comments, etc) haven’t been digitized or have been only partially. 

Yeah, that's my motive for asking this question here.  I'd hoped someone had already done that.  In the very large amount of reading I've done over the years on the Atari 8 bits and other classic machines, I could swear I've even seen some FCC documents on this and some Atari engineers' notes on it, but I have NO idea where I might find those again if I'm not misremembering.

 

Why is it an even remotely important point?  Because a much lower price point would have helped the Atari 8-bit machines to be even more popular and compete better with the C64.  Too late now, of course.

 

My PC history in sequence- $99 ZX-80 kit, VIC-20, Atari 400 with DIY memory expansion and interface to print from game ports to a Trendcom 100 40-column thermal printer to list programs written using the Assembler/Editor cart, Atari 800XL with DIY memory expansion, Atari 1040STf with color monitor, Commodore PC20 with monitor bought for only $300 on the USAF BX open box clearance table at RAF Lakenheath (it worked perfectly).  My transition to exclusively x86 PCs, all DIY, began with one based on the AM386DX-40 and have been AMD CPU based machines ever since.  My 400 purchase came from my reading of the Byte magazine series about how to use the graphics enabled by the custom chips by the brilliant Jay Miner et al.  I wanted an Amiga, but the 1040 with color monitor was a much better deal and I've always been frugal.

Edited December 3, 2022 by Winston2022

[DrVenkman]

2 minutes ago, Winston2022 said:

Why is it an even remotely important point?  Because a much lower price point would have helped the Atari 8-bit machines to be even more popular and compete better with the C64.

The simplified and cost-reduced A8 was the 1200XL - the 48K 800 has (count ‘em!) 7 boards! Main board, power board, OS board, 3 x RAM boards, and personality board. Yes, the cast metal RF shield and large case were expensive but most of that expense was a sunk cost: design, tooling and setup for the assembly plants and component suppliers. Unit costs for something like that aren’t THAT high, especially as production continued all the way into 1983. The 1200XL takes all that down to a single board, much reduced numbers of discrete components and connectors all over the main board, AND a simple press-cut sheet metal shield. Atari’s (Warner Inc.’s!) mistake was pricing it at a premium instead of passing those savings down to the consumer, which cost them almost a year’s worth of time and lost market share.

 

[Faicuai]

1 hour ago, Winston2022 said:

The case was designed around the cast frame.  The case design simply could have been done differently if sheet metal had been used.

This was addressed before.

 

Those tin-can shields are utter crap. They rot, go black, can't withstand humidity beyond certain point and look like garbage. How do I know? Because I had to find a new set of shields for my ALPS / Rev.C 800XL MoBo, which I had originally purchased 9/1984.

 

You need to consider that, in today's historical context where the technical merits and benefits of this old HW can now be fully appreciated beyond the economical realities and conflicts of their time, there is people here that prefer the beefier, more flexible and (overall) better architecture of the 800, over anything else that came later (even though this, of course, may of no interest to you, whatsoever).

 

And that is just FINE, either way!

 

(Also, consider that the 800 was designed to compete with the king of the hill, the Apple II, which on itself, is also a more flexible and open architecture than the 800, as I can currently see on my Apple II/e Platinum, although shielding there is more of a mystery of its own).

 

 

Edited December 3, 2022 by Faicuai
Cleanup, clarity

[Winston2022]

2 hours ago, Faicuai said:

Those tin-can shields are utter crap. They rot, go black, can't withstand humidity beyond certain point and look like garbage.

I agree, but this isn't about what the machine's internals look like decades after they are made.  It's me wondering if the "unknown future FCC requirements" excuse for the aluminum casting is more of a coverup of less than ideal choices made than an actual RFI problem.  As I said above, I post this here hoping someone can direct me to the Atari FCC documents and engineers' notes on the subject that I think I've seen somewhere before.  Also, I've heard before that the RFI shielding overkill had something to do with a demand from Sears for carrying the product.

[Winston2022]

3 hours ago, DrVenkman said:

The simplified and cost-reduced A8 was the 1200XL - the 48K 800 has (count ‘em!) 7 boards! Main board, power board, OS board, 3 x RAM boards, and personality board. Yes, the cast metal RF shield and large case were expensive but most of that expense was a sunk cost: design, tooling and setup for the assembly plants and component suppliers. Unit costs for something like that aren’t THAT high, especially as production continued all the way into 1983. The 1200XL takes all that down to a single board, much reduced numbers of discrete components and connectors all over the main board, AND a simple press-cut sheet metal shield. Atari’s (Warner Inc.’s!) mistake was pricing it at a premium instead of passing those savings down to the consumer, which cost them almost a year’s worth of time and lost market share.

Yes, I'm aware of all of that, but I'm exclusively and ONLY wondering about the actual necessity of thick aluminum castings in the 400/800 which caused higher retail prices than what would have been possible if the castings had not been used and the stamped sheet metal used by everyone else during the same time period had been used instead, those other machines passing FCC requirements then and for multiple years afterwards.

Edited December 3, 2022 by Winston2022