Navigate: 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 Search Prices:

What You Get

The Tyan Tiger 2460 ships with the standard components we are used to from any other off-the shelf system. Accessories include the 80-pin UATA cable, a floppy disk cable, the driver CD and the printed manual. The manual is certainly one of the best ever released by Tyan, most of the superfluous information has been deleted and overall the manual gets straight to the point. The driver CD is still a bit clumsy to navigate, requiring manual selection of the model (coincidentally, the S2460 is not listed but the chipset drivers are the same as for the S2642 Thunder) but that is really the only hurdle the user faces. Overall, a big step in the right direction, even though both manual and CD still don't match the standards set by ASUS or Shuttle, they are better than average. The preprinted version of the manual we received still contained a few minor errors that we reported to Tyan and which have been corrected in the subsequent revision

A different angle
Note the position of the floppy connector next to the IDE interfaces at the bottom of the PCB

Quality

We all like Tyan but in the past, there sometimes have been a few small glitches in terms of quality. Some of these issues were related to the fact that Tyan was first to market with some products and glitches only emerged after the product release. To make it clear, with the Tiger, there are no such issues. Every component on the board is rock solid, regardless of the fact that there is still a hand-soldered capacitor on the board which can be excused easily since it is not in any real estate prone to user intervention once the board is mounted. Moreover, a manual rework of each board shows the commitment to steady improvement of the product.

Over the past few months, we have made it a habit to look in detail at the board power supply circuitry as an indicator of overall reliability and quality. In most current designs, we have seen tri-phase power supplies mostly based on Intersil's programmable multiphase HIP6301 core controller. The very same controller is also present on the Tiger, however, for some reasons, Tyan decided to use only 2 phases total for both CPUs while disabling the other 2 phases. This design has advantages and draw backs, in general the number of phases has nothing to do with stability per se or with satisfying the power demands of the CPU. Those matters are simply factors of the switch voltage regulators (MOSFETs) used, however, in theory, the switching speed, that is, how fast the supply current can be adjusted to the load-dependent demands of the CPU sort of correlates with the number of phases used. In practice, however, the impedance between the power regulator IC, the intermediate driver chips and the overall latency of the MOSFETs constitute much graver hindrances towards fast switching currents than a reduction in phases. On the bright side of things, having the power simply split between the two CPUs warrants that both CPUs receive the same amount of power at all times. That, at least is the theory.

Intermission

Usually, glitches are further towards the end of the review, however, in this case, we experienced one of the strangest glitches ever. The reason to mention this here is not to criticize but because there is a possibility that other users may experience the same problem while the workaround can potentially be a matter of a few seconds.

Upon installation of all components, the Tiger was not running very stable. Intermittent crashes, a complete failure to run RC5 dual client were some of the symptoms. The BIOS showed temperature and voltage differentials between both CPUs (37 - 43 centigrades, 1.73 - 1.89 Vre) at the same time. Interestingly, the CPU with the higher voltage showed the lower temperature. At first glance this appeared to be an improperly seated heatsink, alas, repositioning the heatsink did not solve the issues, neither did switching to a completely different set of HSF ameliorate the problems.

The original setup was run off an Antec 300 W power supply with 30A on the +5V line. This power supply could have been the reason for the crashes, nonetheless, upgrading to a 400 W power supply (Antec) did not change the situation, neither did changing the memory or other components increase stability or decrease the indicated voltage and temperature deltas.

The Solution

What completely changed the situation, though, was swapping the CPUs internally, that is CPU0 became CPU1 and vice versa. For whatever reason, the simple internal swap of CPUs equalized the voltages as well as the temperatures and solved all stability problems. There are several possibilities why this might have worked, the easiest is that the two CPUs were from different batches (one of the original two Athlon MPs arrived DOA and was replaced with a second unit from a different production). This could potentially result in a higher power draw on the two phase branches leading to this particular CPU, causing the voltage to drop and subsequently crash the system. A scenario like this is the main drawback of sharing phases between CPUs. A second possibility is a marginal contact within one socket causing inconsistent current supply. Whichever possibility caused the problems, the bottom line is that if there are some mysterious stability problems, it may be worthwhile switching the CPUs, as silly as it may sound.

next page:    => Layout =>