Examining ATX

The ATX form factor was designed by Intel back in 1995 to improve on the AT form factor. It brought about changes that would help standardize placements with various types of hardware, one being full length expansion cards; in original AT layouts the processor sockets were positioned at the front of the motherboard which kept full length expansion cards from being used. Now that the ATX form moved the CPU sockets to the back right corner of the motherboard full length PCI/ISA cards had plenty of room.

AT also had the problem with drive bay interference since the AT motherboards were longer from front to back. With ATX the motherboard is cut wider, instead, to provide some clearance between drive bays and the board itself. Moving the CPU socket also aided in this change since there was a need for board space at the back right instead of the front left corner of the motherboard.

Some of the other changes from AT to ATX:
  • Power supply connection
  • AT motherboards used dual 6-pin connectors to power the system which was confusing and also potentially dangerous if we did not know what went where. The ATX specification implements a single 20-pin connection from the power supply to motherboard to eliminate confusion and damage to hardware.

  • Power management
  • The older AT specification also had the power button of the case hard wired directly to the power supply. The ATX specification implements the "Soft Power" feature which allows the motherboard to turn the power supply on and off by software. For example, when shutting down the Windows OS in an AT set up the system needed to be manually shut down by pressing the power button. The ATX form factor allowed the system to be automatically powered down after the OS logged off.

  • Integrated I/O, peripheral connections
  • On older AT boards, instead of having the parallel and COM ports we see at the back of the motherboards, extra cables were required to extend those features to the case. We needed to run cables from pin headers on the motherboards and mount them to the cases to be able to connect printers or other devices that used these ports. When the ATX form factor was introduced, these connectors were soldered directly on to the motherboard to both reduce the time to install components and the overall cost of a system, as well as to increase the reliability of the ports.

  • Cooling and air flow
  • Moving the CPU to the back right of the case closer to the power supply and the backside fans had a positive effect on cooling. Warm air created by the CPU would be pulled out through the back of the case as well as up through the power supply if it had a fan facing the CPU.
With the AT form factor, the CPU was at the center of the case and required more fans to cool the entire system. Moving the hottest component off to a side of the case meant more stable temperatures for other components like HDDs and memory.

The ATX had a clear goal in mind; a new design that would reduce clutter, standardize component design and bring more control to the motherboard.  The BTX mandate is a little less clear; reduce the PC footprint while more efficiently cooling the components.  Keep this mandate in mind as it plays heavily on our conclusion.

Index Examining BTX
Comments Locked

77 Comments

View All Comments

  • DerekBaker - Monday, November 15, 2004 - link

    #25

    From here I should think: ">From the picture we see the flow of warm air pulled forward from beyond the VGA card. As it moves towards the front of the case the air flows over the Southbridge and Northbridge passively cooling each chip. We then see from the shades of red the air warms up as it flows through the CPU heatsink then begins to cool as it moves out the front of the case."

    http://www.anandtech.com/casecooling/showdoc.aspx?...


    Derek
  • Oxonium - Monday, November 15, 2004 - link

    Let's see, a new form factor with new case requirements and STILL no standardized connector for the front Power/Reset/LED's/etc.? I'm all for cooler cases, but a standardized connector for those items is something that should have been implemented years ago.
  • HardwareD00d - Monday, November 15, 2004 - link

    #24, who cares since athlons run much cooler and don't need this case form factor.
  • Doormat - Monday, November 15, 2004 - link

    I heard a rumor that it wont work with Athlon 64 chips or any type of chip that has the memory controller integrated with the processor, because the distance between the processor and the RAM banks are too far. Anyone confirm this?
  • Superbike - Monday, November 15, 2004 - link

    Wow AT is back.
  • phisrow - Monday, November 15, 2004 - link

    While I agree that the tweaky enthusiast market doesn't have much to gain at the moment by moving from ATX to BTX, especially as most of them are running A64s or praying-for-death overclocked mobile Athlons, I think there is certainly something to be said for the form-factor. It does seem to be more efficient for cooling than your average ATX(especially for small cases). Even if you are running some PentiumM or a derivative thereof, more efficient is still better. After all, if a given case design can cool a high-end Prescott without deafening its user, it can probably cool a brutally overclocked Pentium M with its fans running at half the speed. For that matter, if the trace length issues aren't actually crippling, I would like to see some A64 motherboards in this format.
  • johnsonx - Monday, November 15, 2004 - link

    Visualize an office full of MicroBTX Desktop computers. Visualize the women in the office cupping their hands around the heat exhaust... visualize the overclocking tips flowing around the female staff as they strive to increase heat output... visualize the men pouring sweat, desparately trying to duct-tape the front vent to block the hot Intel Air.

    I'm sorry, but any spec that seeks to exhaust hot air out the FRONT of any case, whether it be a desktop, tower or SFF, is just plain STUPID.

  • araczynski - Monday, November 15, 2004 - link

    my god, after all these years, they fall back on the packard bell mentality...
  • KristopherKubicki - Monday, November 15, 2004 - link

    GTMan: Temperatures were certainly taken with the cases closed.

    Kristopher
  • GTMan - Monday, November 15, 2004 - link

    I wonder if all those temperature measurements were done with the case open? I would think a case designed for efficient air flow would only perform properly (as designed) when fully assembled!!!

Log in

Don't have an account? Sign up now