3dfx Voodoo 5 5500 Review
Good and Bad
This content was originally featured on Amdmb.com and has been converted to PC Perspective's website. Some color changes and flaws may appear.
(Click for Larger Image [front])
(Click for Larger Image [back])
The physical characteristics of the card are somewhat troublesome. First, you’ll notice the large length of the PCB. This can make it very difficult for purchasers to find placement in their home system for card of this size. I myself have a Inwin server case, and still had to rearrange the IDE cables to allow the Voodoo 5 to sit correctly. And, though I am not a computer engineer, by looking at the components of the Voodoo 5, I can’t help but notice all of the extra, empty space left on the board. Had the engineers had more time or been more careful, it could have been possible to take another inch or two off the length of the card.
Next, we have the two processors of the VSA-100 chipset(s). I am also disappointed by the size and power of the heatsink/fans on the Voodoo 5 5500. The Aavid brand fans seem ill-suited for an overclocked video card and leave the opponents on the GeForce 2 cards an easy point for overclockability. On this same note, is the RAM on the 3dfx card. The Hyundai RAM is high quality, running at a speed of 350 MHz. Again, there are no heatsinks on the chips. Though they are not necessarily needed, the advantage for overclockers is significant and sways potential customers who look at these features carefully.
(Click for Larger Image [fan])
(Click for Larger Image [back])
Also, unlike the 64mb Hercules GeForce 2 card that we reviewed earlier, the Voodoo 5 5500 does not have an TV/S-Video output options. Again, this kind of option is not necessarily a needed feature but adds to the retail salability. I found the 800x600 output mode on the HDTV’s while playing Unreal Tournament to be quite fun and would highly recommend it!
3dfx does have a strong retail package to give its purchasers. Inside the box you will find the needed drivers, a few game demos, and a apology as to the delay of the VSA-100 chipset. The coolest part is the 3M style mouse pad with the Voodoo 5 face on it. This is not a full size mouse pad, but is decent enough sized to play FPS games on. In this case, 3dfx took the extra step to try and please the customer.
There has also been a lot of talk about the whole FSAA debate (Full-Scene Anti-Aliasing). The technical end of it is that the processor render each frame or scene multiple times, and blur the frames together before they are output to the monitor. The results is less pixelation on screen. In my opinion however, this feature becomes trivial. For the frame hit that you get when this is activated, the visual difference is minimal, especially considering that while playing any game you should be more worried about the opponents than the 'jaggies' that haunt other players. I doubt that any avid Quake III or Unreal Tournament fan would want to sacrifice +30% of their frame rate for a slightly improved image.
Enough with all this trivial information, let’s move on to the true test of a video card, the benchmarking. Here’s the test setup, the exact same as the Hercules 64mb card:
Test System Setup
AMD Athlon 750
1 x 128MB PC133 Mushkin
20.5 GB 7200 RPM Western Digital
Windows 98 SE