|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
|
Prices: CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ |
LOSTCIRCUITS |
||
| ASUS P4S533 Faster | ||
| (Review by MS, June 26, 2002) |
The ATX power connector is at the edge of the PCB, separated from the auxiliary power Smart-Plug by the powersupply fan header. The ATX 2.03-compliant dual 12 V 4-prong header is next to the ATX I/O block and almost at the heart of the CPU power circuitry. The latter shows probably the most important deviation from the P4S333 design in that the established HIP 6301 had to move over to make way for the latest STMicro dual phase controller specifically designed for high power micro processors.

Close-up of the power circuitry with the STMicro main controller chip (arrow). Because the chip markings are laser-etched, they are difficult to read. Each phase consists of two different Power MOS which serves the purpose of "smudging" the phase onsets and, by extension, the switching currents. There are no intermittent driver chips since both drivers are integrated into the STMicro controller.
Again, why is it that 2 phases (180 degree shift) can be better than three phases with a 120 degree phase shift? The answer is that the fastest micro controller is limited by the different components it is controlling and if there are relatively long traces to the driver chips and then additional traces to the actual MOSFETs, even the fastest microcontroller cannot live up to its potential because of trace impedances. As mentioned in several earlier reviews, the trend in the controller industry has been to compact the design, meaning, to integrate the driver chips into the main controller and, this way, shave off latencies as well as enable higher switching frequencies.
The probably best known example are the controllers made by On-Semi used by EPoX and ASUS on their SocketA boards. In the case of the P4S533, ASUS are going with yet another solution, the ST L6917BD 5 bit programmable dual phase controller. This new controller is VRM 9.0 compliant, features integrated dual drivers for high gate currents (up to 2.0A / phase). On paper, the 600 KHz switching frequency appears somewhat low compared to the 1-4 MHz of some of the Intersil or Semtech controllers, however, keep in mind that the latter never had a chance to fully capitalize on these frequencies for the above mentioned reasons.
The L6917BD has a few other nifty features as e.g. soft power-On. Soft Power-On means that instead of immediately blasting the CPU with the full operating voltage, the controller initializes a voltage ramp that reaches the VID (Voltage ID) value of the respective CPU within 2048 cycles. This, of course will result in a noticeable POST delay of roughly 1 µsec but most will be able to live with that. Another neat trick used in the control circuitry is the use of different types of power MOS per channel with different timing characteristics to make the onset of each phase more fuzzy and, thus, reduce switch currents. We have seen a similar approach used by Tyan on the Tiger MPX but somehow it didn't click at the time we did the review. The last point is the integrated droop functionality to reduce the necessary size and ESR value of the output capacitor.
So, does it work? We sure think it does and have some more on that issue later.
next page: => BIOS, Test Configuration =>