Received my PCB, still waiting on parts. Board looks great!

Anyone feeling like doing a card build video? 😀

I might do one. Still have to receive my board though 🙁

I received the 2 boards I ordered. They look great!

This is just such an awesome project, many thanks to Madao for having made this project a reality! 😀

I have received the board I have ordered. Looks amazing!

Thanks Madao!

Can't wait for the rest of the components to finally build mine.

Let us know if you make more PCBs. I am very interested in buying two.

-Brian

I'm also interested to buy one PCB from you.

I'm also interested, especially if it has a feature connector onboard

feartures connector? i make thought about it.

Next Message: Have all people got their PCB "765VL000" ?
Finished ?

regards
matt

Madao wrote on 2022-01-11, 18:38:

feartures connector? i make thought about it. […]
Show full quote

feartures connector? i make thought about it.

Next Message: Have all people got their PCB "765VL000" ?
Finished ?

regards
matt

I have mine. excellent board quality, greatly appreciated. Mind sending me the BOM or your parts list in a DM? 😀

cheers

Hi Madao, do you still have any PCBs left of your 765VL000 board to buy?

Madao wrote on 2022-01-11, 18:38:

Next Message: Have all people got their PCB "765VL000" ?
Finished ?

Waiting on component fulfillment from Mouser, if I get bored this weekend I'll check for alternatives. The board is nice. 😀

I actually have a few hundred of these SOJ-40 sockets on a reel. Could be used with more spacing between the chips on the PCB. Also those S3 Trio/ViRGE chips have weak 135MHz RAMDACs which blur the image at higher resolutions. They cannot even do 1024x768x32bpp without interlacing, so there's no point in producing 4Mb cards unless counting on 3D deceleration features of ViRGE. Things were better with Vision968 which used external RAMDACs of TI or IBM. Especially TI, that 220MHz TVP3026 was very nice, though 175MHz was fairly sufficient. I think adding an output RGBHV driver like MAX9511 can greatly improve display quality.

I finally got around to soldering the chipset on mine. I used TS391AX solder paste and didn't have a single bridge, nor any floating pins.

For anyone trying this who is new to paste, this is about how much paste you want, no more. Too much = bridges and too little can result in floating pins, esp. on desoldered chips.

Cool! By the way, the S3's were also known to work well with cheap 3D shutter glasses..

See

http://www.stereo3d.com/spex.htm
http://www.stereo3d.com/3dmaxlst.htm

The lower end solutions with LCD-BIOS often supported a high refresh rate (say 160Hz) on S3 Trio 64V+ chipsets, which the S3 ViRGE 325 is/was compatible to.

Of course, other VGA cards were compatible per se, too, but maybe didn't reach such a high refesh rate.
Which was necessary to not get a headache due to flicker.

Last but not least, some S3's had a built-in stereo jack connector for shutter glasses.

So it's double nice that an S3 VLB card is being made. 👍

Jo22 wrote on 2022-02-25, 11:01:

The lower end solutions with LCD-BIOS often supported a high refresh rate (say 160Hz) on S3 Trio 64V+ chipsets, which the S3 ViRGE 325 is/was compatible to.

On the other hand, the 3D performance of the S3 ViRGE 325 is slow enough that rendering pictures for one eye is borderline usable at stock frequencies and 640x480 (too bad that lower there are not that many games that allow lower resolutions). Stereo output on a 86c325 will only be useful for viewing still images, but not for gameplay.

mkarcher wrote on 2022-02-25, 19:49:

Jo22 wrote on 2022-02-25, 11:01:

The lower end solutions with LCD-BIOS often supported a high refresh rate (say 160Hz) on S3 Trio 64V+ chipsets, which the S3 ViRGE 325 is/was compatible to.

On the other hand, the 3D performance of the S3 ViRGE 325 is slow enough that rendering pictures for one eye is borderline usable at stock frequencies and 640x480 (too bad that lower there are not that many games that allow lower resolutions). Stereo output on a 86c325 will only be useful for viewing still images, but not for gameplay.

That might be true, but the ViRGE 325 can also act like a good old Trio64v+ 2D card, so little is lost.
If software-rendering is used, both game play and shutter glasses should work just fine. :)

Last time I played Descent II, it was ok, at least.
I didn't "over"clock the 325, also. The 325 was clocked rather conservative by default, you know.
The 3D acceleration feature is nothing to write home about (?),
but comes close to the 10-15 FPS that someone with a classic 486DX33 and an ISA VGA experienced back then, anyway.
So it's not making things worse, just prettier. ;)
The copy of Descent II that I played did turn on all bells and whistles, the bilinear filtering stood out to me.

That's one thing that I think was one of the issues with the Virge's. They had too much features.
And were all (mostly) turned on in-game most of the time.

If S3D games ran accelerated, but with the same quality/resolution as 320x200 VGA,
it might have been an early accelerator, actually.
Or if games ran at 640x480, but without all the filtering and and just 256c, who knows.
Unfortunately, that wasn't appealing to the game developers, of course.
They wanted to show off what they and there games were capable of, the Virge was just a tool.

But for what it was, it wasn't so bad, I think.
I mean, it wasn't expensive. A quality 2D card sold for similar price.
Abd the visuals of the Virge were on par with the Virtua Racing or SuperFX games, at least, which were popular around 1995..

Btw, here are two old videos of me when I did experiment with a ViRGE.
It seems that the VIRGE 325 was a nice extra for owners of early pre-MMX Pentium machines, at least.
Say, a classic Pentium 75 or Pentium 90 in a big tower.
Software-rendering really shines on a Pentium 166 MMX running 98SE, though, I think.
Anyway these are just some experiments.
Real games may behave differently than a benchmark.

Final Reality (software)
https://www.youtube.com/watch?v=qiKq0x4HHaw

Final Reality (Virge 325)
https://www.youtube.com/watch?v=Ujfl3fwb2aE

Edit: I forgot. To be fair, I must mention that I used late drivers.
As such, they weren't available when the ViRGE was young and new.
So the actual experience might have been inferior to users of the time.

Jo22 wrote on 2022-02-25, 21:01:

That might be true, but the ViRGE 325 can also act like a good old Trio64v+ 2D card, so little is lost.

Pardon my reach here, but does anyone have any numbers comparing the Trio64V+ vs. the ViRGE 325 in DOS 2D? I have two of the homebrew Trio64V+ PCBs and was considering soldering a NOS ViRGE chip onto the second, but am hesitant because S3D won't work well in the current VLB environment. If the ViRGE provides no measurable or visual benefit in DOS 2D or Windows 2D, then I will solder the Trio64V+ onto the second PCB. Many thanks!

feipoa wrote on 2022-02-25, 23:53:

Jo22 wrote on 2022-02-25, 21:01:

That might be true, but the ViRGE 325 can also act like a good old Trio64v+ 2D card, so little is lost.

Pardon my reach here, but does anyone have any numbers comparing the Trio64V+ vs. the ViRGE 325 in DOS 2D? I have two of the homebrew Trio64V+ PCBs and was considering soldering a NOS ViRGE chip onto the second, but am hesitant because S3D won't work well in the current VLB environment. If the ViRGE provides no measurable or visual benefit in DOS 2D or Windows 2D, then I will solder the Trio64V+ onto the second PCB. Many thanks!

I ran a few quick tests with the various PCI Virge 325 and Trio64V+ 765 cards I had on hand. It looked like at 320x200, the 325 and 765 were equal; at VESA Modus 100 640x400x8bpp LFB, there may be a slight edge for the 765 card. BIOS seems to matter and my sample set isn't large enough to draw firm conclusions. Here's the results: