The recent Blade.org announcement heralds that IBM and Intel are looking to standardize blade hardware designs, allowing blade systems from multiple vendors to work together. It’s no surprise that Intel is interested in this — its own ATX standard has been a massive success. This article looks at the history of ATX and some of the lessons learned about power supply and chassis standardization.
What i always hated about the ATX “Standard” was all the built in stuff. dedicated IRQ resources to useless devices. I ran SCSI hard drives for years so i had absolutly no use for a built in IDE controller.
deactivate the ide controllers in the funky i386 bios?
I’d also remove:
– get rid of 8086,80286 compatibility (who needs them, anyway)
– remove crappy A20
– remove real mode completely
– add protected mode BIOS services instead
– remove PS/2 ports and replace them with USB
– leave serial and parallel ports only as an option
ATX is about case, motherboard and power supply form-factor and electrical standards. And not much else, AFAIK.
Some other specification is to blame for the common PC.
ATX standard applied to tower design is absolute nightmare from thermphysics point of view.
With more reasonable and strict design we could run even x86-based systems, which are rather heat than performance generater by definition, to run without coolers (heatsink only) up to 1 GHz.
Also it is absolutlely ugly from end-user point of view, regarding peripherals plugs. You see it in nowadays PC cases which are permanently inventing wheels with ugly design hacks to bring all that sound, USB and other connectors to front or side of PCs, while Mac-s for example, almost always had convinient placement of those connectors.
I find that after dealing with Old style AT systems I felt at home with ATX, much easier than what I used Before(I deal mostly with desktops so for servers and portable devices, maybe, it’s not the way)
I wish they has continued to use the NLX standard as I feel it had way better thermodynamics than ATX. I stirr use an NLX and I’vd got it powerleap’d to a 1.4 GHz and stirr the thermally controlled fan hardly kicks on. Not bad, up from a 300 MHz.
ATX does not specify the location of USB / Firewire etc headers, as far as I’m aware. It’s motherboard manufacturers who insist on putting them all, for reasons best known to themselves, as far away from the front panel of a standard case as possible.
As for heat, ATX is not the ideal design, but even so, there are fanless ATX cases which can take up to 3GHz CPUs. They cost up to and even more than $1,000, of course, but if you want ’em, they’re out there.
“while Mac-s for example, almost always had convinient placement of those connectors”
Please, please, just stop it!
I’ve always been amazed at the creativity at the user level while following the ATX standard. Some of the mods are incredible.Apparently ATX was a bit more flexibly than most are giving it credit for.
The test of time has proven it expandable, easy to customize, and simple to assemble. Not bad for a spec draw up before some of you were born.
I’d say they did very well, they got most it right.
I build my own systems for years and it work well for me to carefully watch the specs of the case, the motherboard I’ll be using. Now, as cpu’s get more powerful, power supplies and heat management are to be considered carefully. Power supply and chassis standardization will
not only lower the cost of building systems but also will
assure us the the components are compatible.
In the time of 486’s IDE controllers were seperate. Disabling the on board is not good enough. if they were never integrated people would have had a choice of upgrading from a UDMA33 controller to a UDMA133 controller or SATA etc.. I am not a huge fan of integrated Audio or network either but its getting harder to find a mobo that doesn’t have something integrated. The PC design is not the best but intergrating things screwed it up even more. Hopefully someone has the forsight to totally redesign PC’s for future generations.