What a great photostory and mint condition of parts!

Great Rig 😀
I also love silent computers, but for my vintage builds I love the rustle and crackle of this old drives - makes the build more authentic.
But I also understand if you discard them, after all this wear and tear such old drives are error prone ...

Very nice looking build! I don't see people choosing the Intel 486 chip over the AMD 5x86 much anymore.

Spitz wrote on 2020-12-10, 13:00:

What a great photostory and mint condition of parts!

Thanks 😀

ildonaldo wrote on 2020-12-10, 17:22:

Great Rig 😀
I also love silent computers, but for my vintage builds I love the rustle and crackle of this old drives - makes the build more authentic.
But I also understand if you discard them, after all this wear and tear such old drives are error prone ...

Glad you like it 😀 I don't have a problem with potential errors. I have enough old disks that are still in great condition and even if one of them fails I can quickly replace it.
However, I felt sorry for this PC - the PSU is dead silent, the CPU fan also dead silent... A 25 year old HDD would have killed the silence 😁 I do have others which are noisier and are also paired with old HDDs (if I want to get the full authentic experience) 😁

OzzFan wrote on 2020-12-10, 17:58:

Very nice looking build! I don't see people choosing the Intel 486 chip over the AMD 5x86 much anymore.

Thanks! I have two others with the AMD 5x86 (on the Zida 4DPS & Biostar MB-8433UUD motherboards) 😁
Also have an Intel 486 DX2-66 on another VLB motherboard (the Aopen Vi15G), which I will put in a case at some point in the future (if/when I find another nice case).

I finally gave up on the SB16 IDE connector. No matter what I tried, it didn't work and the CD ROM driver would freeze during loading (I tried two different SB16 sound cards, same issue).
So I just moved the CD-ROM drive on the lower bay and used one IDE cable to connect both the SD-to-IDE adapter (master) and the CD-ROM (slave). Everything works perfectly now. 😀

bloodem wrote on 2020-12-11, 09:30:

I finally gave up on the SB16 IDE connector.

Are you sure that it is an IDE-Interface?
Most old Sound Blaster Pro cards feature a proprietary interface for a Panasonic (Matsushita MKE) drive (e.g. for a "Panasonic CR-652-B" CD drive).
I just looks like IDE but works differently.

ildonaldo wrote on 2020-12-11, 10:46:

Are you sure that it is an IDE-Interface? Most old Sound Blaster Pro cards feature a proprietary interface for a Panasonic (Mats […]
Show full quote

bloodem wrote on 2020-12-11, 09:30:

I finally gave up on the SB16 IDE connector.

Are you sure that it is an IDE-Interface?
Most old Sound Blaster Pro cards feature a proprietary interface for a Panasonic (Matsushita MKE) drive (e.g. for a "Panasonic CR-652-B" CD drive).
I just looks like IDE but works differently.

It says 'IDE INTERFACE' next to the connector.😉 The solder spots for the Creative/Panasonic connector are not populated.

When reading I thought nah I can't do that silent PC's creep me out, but then when I saw you'll still get the FDD boot sound the creepy factor is no longer an issue by my own standards and quite the achievement when you think how noisy HDD's and PSUs were in the day.

Love that case, that display is so classy and even Netware ready, another retro throwback
Overall nice tidy build.

Re CD off the soundcard, Common things are ensuring the the IDE port is actually enabled on the card plus often switches needed adding to config sys to tell the driver to look for the drive on the soundcard rather then IDE controller.
But whatever, its messy, I prefer the simplicity of running as a slave myself it its an option.

Re soundcard, going off your signature you have other Socket 3 PC's so I'd chose the "other" soundcard if they have Creative I'd go with ESS in this build, keeps things different 😀

The IDE port on the Sound Blaster card is usually set to tertiary port so as to not conflict with the primary and secondary ports that may exist on the motherboard. Assuming you don't need to initialize the IDE port on the sound card through a separate utility, you can specify that the CD-ROM device driver look to the tertiary port on the command line. For example, if you're using OAKCDROM.SYS, use:

1device=c:\drivers\oakcdrom.sys /d:OAKCD001 /P:1e8,12

You can experiment with the ports using the common values for parallel ATA:

• /P:1F0,14 (primary)
• /P:170,15 (secondary)
• /P:1E8,12 (tertiary)
• /P:168,10 (quaternary)

1F0 is likely your primary IDE on the VLB card and 170 would be the secondary IDE if it had one. Most DOS CD-ROM device drivers support the /P switch to specify a port. If yours doesn't, you may want to try a different device driver.

Thank you chinny22 and OzzFan for the suggestions. I will try them on another socket 3 build whenever I get the chance!

Yes, I do have other socket 3 builds as I mentioned in a previous post, one of them actually has a Sound Blaster Pro 2 and the other one an Aztech Multimedia Pro 16 IIB-3D. If/when I do another socker 3 build (with a 486 DX2-66), I will use an ESS card for sure 😀

bloodem wrote on 2020-12-10, 11:24:

VGA: ARK Logic ARK1000PV VLB 2 MB (nothing special, but did not notice any compatibility issues during my tests)

Update: Well, it turns out that, apparently, this card is special and I didn't even know it. According to CPU Galaxy, it's one of the fastest VLB cards if not THE fastest (even faster than Tseng ET4000).

Nice case.

I have a similar build.
32mb might be to much for DOS but not for 3.x
Win3x can use that memory no problem.
The problem I had with my VLB controller was supporting large hard-drives. The driver for the VLB controller will only support up 410mb hard-drive. If I used a larger capacity hard-drive the driver would report an error and only run at 16-bit mode and not 32-bit mode.
So slower data transfer rates. But it still runs fast enough for DOS and Win3x.
Win3x VLB driver would report an error to an only run at 16-bit mode.

So the VLB card is useless in 32-bit mode if you use a hard-drive larger than 410mb.
I don’t know how the driver would react to an SD card however.

Did your VLB controller report any errors with the SD card ?
Anyways it should run fast enough even if the SD card is running in 16-bit mode.

Also I was able to down clock my 486dx4-100 to a 286@20mhz.
So you should be able to do the same.
Link: Re: 486 Multimedia dream build ( 1993/94 )

Lovely, clean box you've got going here!
I ought to strip-clean my systems one of these days myself.
32MB is pretty much ideal for DOS on a 486 IMO, you can set up huge caches or play games that need lots of EMS memory.

Intel486dx33 wrote on 2021-01-10, 17:44:

The problem I had with my VLB controller was supporting large hard-drives. The driver for the VLB controller will only support u […]
Show full quote

The problem I had with my VLB controller was supporting large hard-drives. The driver for the VLB controller will only support up 410mb hard-drive. If I used a larger capacity hard-drive the driver would report an error and only run at 16-bit mode and not 32-bit mode.
So slower data transfer rates. But it still runs fast enough for DOS and Win3x.
Win3x VLB driver would report an error to an only run at 16-bit mode.

So the VLB card is useless in 32-bit mode if you use a hard-drive larger than 410mb.
I don’t know how the driver would react to an SD card however.

Well, to tell you the truth, I have no idea if it currently runs in 16 or 32 bit mode. I usually only try and fix if there is an actual performance issue worth fixing. 😀
In this case, the storage performance is so good, that, once I fixed the disk corruption problems, I did not give it a second thought. And I didn't even run any benchmarks, but the thing is so snappy that everything loads instantly 😀
DOS boots in a few seconds (and there are quite a few things that I load in config.sys/autoexec.bat - like the sound/mouse/cd rom drivers). Same with Windows 3.11... I just type "win" and I'm already at the desktop. 😀

Intel486dx33 wrote on 2021-01-10, 17:44:

Did your VLB controller report any errors with the SD card ?
Anyways it should run fast enough even if the SD card is running in 16-bit mode.

No errors at all. It just refuses to work properly in LBA mode with any disk - period correct or not. I could have probably fixed it with a DDO, but I was lazy and went with LARGE mode instead, which works great.

Intel486dx33 wrote on 2021-01-10, 17:44:

Also I was able to down clock my 486dx4-100 to a 286@20mhz.
So you should be able to do the same.
Link: Re: 486 Multimedia dream build ( 1993/94 )

I now upgraded the L2 cache to 512KB, and the speed with a disabled L1 cache has increased significantly - 15.6 FPS in 3DBench 1.0c, up from 10 FPS before, which is now deep into 386 DX-40 territory (not something I particularly like).
However, I just realized that I did not test with L1 cache disabled + turbo off, I imagine it should be slower (hopefully not too slow). Unfortunately the PC is already in storage, but I'll definitely need to test this in the future.

Eep386 wrote on 2021-01-10, 17:55:

Lovely, clean box you've got going here!
I ought to strip-clean my systems one of these days myself.
32MB is pretty much ideal for DOS on a 486 IMO, you can set up huge caches or play games that need lots of EMS memory.

Indeed. Some time ago I bought locally many 8 MB / 16 MB FPM and EDO RAM modules (for $1 a piece). So as you can imagine, I always go with 32 MB for all my 486 PCs (which would have been close to insane for 1993- 1994) 😀
As an example, my first PC which my parents bought me in 1997 (a Pentium MMX 166) had 16 MB EDO and it worked perfectly in Windows 95. In retrospect, 16 MB was probably a bit on the low side for 1997 (I'm guessing that rich people who were buying Pentium 2s would have also gone with 32 or even 64 MB at that point), but in my case, 16 MB was definitely not a bottleneck. 😀

Intel486dx33 wrote on 2021-01-10, 17:44:

The problem I had with my VLB controller was supporting large hard-drives. The driver for the VLB controller will only support u […]
Show full quote

The problem I had with my VLB controller was supporting large hard-drives. The driver for the VLB controller will only support up 410mb hard-drive. If I used a larger capacity hard-drive the driver would report an error and only run at 16-bit mode and not 32-bit mode.
So slower data transfer rates. But it still runs fast enough for DOS and Win3x.
Win3x VLB driver would report an error to an only run at 16-bit mode.

So the VLB card is useless in 32-bit mode if you use a hard-drive larger than 410mb.
I don’t know how the driver would react to an SD card however.

Windows 3.1x 32-bit Disk Access was an early attempt to move away from expensive context transfers using the system BIOS. At the time it was written, Microsoft had the driver working with "BigDOS" compatible drives ( < 504MiB ). The driver didn't support Large or LBA type drives required for larger drives. There are software drivers from both Maxtor and Western Digital to provide 32-bit Disk Access to drives > 504MiB (this is not the same as disk overlay software).

The Maxtor driver works quite well with drives from Connor, Seagate, and Quantum. I generally don't use Western Digital, but they may work with the driver too or you can find their driver instead.

Update:

I eventually replaced the I/O controller with another (newer) one, because I still had all sorts of intermittent issues. Once I did that, all storage related issues disappeared and I can now use LBA mode. Also switched from SD card to CF card with a case bracket adapter.
I also managed to make the Sound Blaster 16 IDE port work with the CD-ROM drive...

But now I have a different problem that I've been trying to debug: I realized that in certain workloads, the SB16 causes VERY loud noises - pops/clicks (with headphones, they can sometimes scare the bejesus out of you).
A few instances where these clicks and pops are always occuring:
- right after MS DOS loads, during the HIMEM extended memory test.
- during 3DBench 1.0C (but much less pronounced compared to the HIMEM test)
- in certain games like Duke Nukem 3D (more rare, but once the pops occur they can be VERY loud).

The pops during the HIMEM testing were a clue, it meant that they are somehow linked to memory. So I tested every possible combination of memory timings / wait states that I could think of, but nothing worked... With the slowest timings and highest wait states the CPU performance drops to DX2-66 levels, but the SB16 noises are still there.
Through trial & error, I eventually I disabled the "memory hidden refresh" option in the BIOS and... the pops completely disappeared. If I enable it back, the noises return. So it's 100% consistently reproducible.
Unfortunately, if I disable hidden refresh, the system is very unstable (it usually freezes within the first 5 - 10 minutes after boot). Tried also toggling slow refresh together with hidden refresh, but the system either remains unstable (with hidden refresh disabled / slow refresh enabled) or with noises (hidden refresh enabled / slow refresh enabled or disabled).

Other things I've tried:
- 3 x SB16 cards (1 x CT2230 / 2 x CT2290), they all exhibit the same issue. Obviously, they work fine on some other motherboards/platforms.
- an ESS AudioDrive ES688 with Yamaha OPL3 (non-PnP) and an ESS AudioDrive ES1688 (non-PnP) - both work perfectly, no clicks and pops.
- an AWE32 CT3980 with Yamaha OPL3 (PnP) - works perfectly, no noises whatsoever (and to my ear it even sounds better than the SB16 cards)
- various 72 pin FPM memory modules (2 x 16 MB, 2 x 4 MB, etc) from multiple manufacturers - no changes with the SB16.
- temporarily switched from VLB I/O & VLB VGA to ISA I/O and ISA VGA, just to make sure that there isn't some weird VLB related quirk at play. Noises are still present, though.

Right now I'm all out of ideas. Seems that this motherboard just hates SB16s but probably works great with any other sound card 😀. Of course, I could use the AWE32 (which would be more period correct for this PC + it's better in every way), but I have other plans for that card...

The noises you're hearing during Himem's memory test with the CT2230 / CT2290 is possibly due to floating op-amps on them picking up noise from the ISA bus. (Creative got really lazy and left several portions of quad amps unconnected.) Changing the surrounding hardware will help very little.

I was also assuming that it's something similar to this. However, the fact that the problem completely disappears when toggling a memory specific option, means that there must be something else going on (besides typical noise on the ISA bus).
I do agree that the floating op-amps might be somehow related to this issue, though.

ildonaldo wrote on 2020-12-11, 10:46:

Are you sure that it is an IDE-Interface? Most old Sound Blaster Pro cards feature a proprietary interface for a Panasonic (Mats […]
Show full quote

bloodem wrote on 2020-12-11, 09:30:

I finally gave up on the SB16 IDE connector.

Are you sure that it is an IDE-Interface?
Most old Sound Blaster Pro cards feature a proprietary interface for a Panasonic (Matsushita MKE) drive (e.g. for a "Panasonic CR-652-B" CD drive).
I just looks like IDE but works differently.

They grew out of that halfway through the SB Pro 2s, by the 16s they had IDE.