TheMobRules wrote on 2023-01-31, 01:53:

leonardo wrote on 2023-01-30, 18:37:

Sorry about the long wait! I got sick and didn't get around to this until today. I've attached a screenshot of the report summary and the BIOS-file. Hope you find some use for it.

Thanks a lot leonardo! By quickly examining the file in a hex editor, it seems your BIOS (A75B - Apr-21 1994) is slightly newer than mine (A75A - Mar-31 1994). Very interesting! I wonder if it has fixes for the bugs I noticed in my version, I will test your BIOS in my board in the next few days and will report back.

I'm on the edge of my seat. 😁

Vic Zarratt wrote on 2023-01-31, 02:40:

That is really a nice build for what's labelled as run of the mill, it's one for that 1994 vibe. The case is fairly correct wit […]
Show full quote

That is really a nice build for what's labelled as run of the mill, it's one for that 1994 vibe.
The case is fairly correct with the tri-digit LCD - the first 100mhz 486's and 90mhz pentiums were from intel were available for evaluation before october 1994 and Dell shipped them in their review machines.
There was a 2-part feature in PCW magazine in the UK 1994 april and may issues (or was it march?) where Frank Leonardt built something like this, and the cirrus 5428 VLB was in high demand at the time.
His goal was to build a multimedia PC that complied with the MPC2 standard for under £1000 I think he had to settle on a 486 sx25 as a compromise. I can't remember if the price includes a monitor but i do have the article cutting somewhere...

It did turn into a fairly nice retro-setup in the end, and it was quite fun seeing it slowly come together. If you find the article you mention it might make for an interesting read.

In other news, sadly my SoundBlaster 32 passed away recently. It was acting up from the day I got it, so I guess it may have been dying from since before I even got it. One day after I booted it, I was greeted with

ERROR: Bad Serial ID Checksum (VendorID 9C008C0E) expected=BC actual=FF

After that no amount of playing with it would bring it back online. I even flashed the EEPROM as suggested in this thread, but to no avail. This does appear to be a somewhat common problem for the model, and as none of the other components have exhibited any warning signs, for the time being I'm still focused on finding a new sound card rather than a PSU.

If i come across the article then i'll do a quick scan and upload it to my webpage at some point - i'll then drop an update back here.
I'm sorry to hear that your soundblaster died on you like that Leonardo, it's gutting when things like that happen.
I do intend to sell some working isa soundcards on my local FB marketplace, including a ct3620. if you want more talk on that than i don't mind you sending me a PM.

leonardo wrote on 2023-01-31, 10:19:

I'm on the edge of my seat. 😁

The BIOS didn't work unfortunately. No sign of life when turning on the system.

Which is strange given that the BIOS string is exactly the same in both, I wonder if there's any difference in the motherboard revisions that prevents this one from running.

TheMobRules wrote on 2023-02-01, 03:15:

leonardo wrote on 2023-01-31, 10:19:

I'm on the edge of my seat. 😁

The BIOS didn't work unfortunately. No sign of life when turning on the system.

Which is strange given that the BIOS string is exactly the same in both, I wonder if there's any difference in the motherboard revisions that prevents this one from running.

I'm a little disappointed. I'll try the dump again (this time paying closer attention to the settings, I didn't check the BIOS shadow-thing for example) and see if the resulting file is any different. I may have a new dump for you soon. That said, it could be a different revision of the board also.

Vic Zarratt wrote on 2023-01-31, 14:50:

If i come across the article then i'll do a quick scan and upload it to my webpage at some point - i'll then drop an update back here.

Thank you!

Vic Zarratt wrote on 2023-01-31, 14:50:

I'm sorry to hear that your soundblaster died on you like that Leonardo, it's gutting when things like that happen.
I do intend to sell some working isa soundcards on my local FB marketplace, including a ct3620...

Thank you for your condolences and kind offer as well, but it turns out a coworker had an old card stashed away and they offered it to me for free:

I had never even heard of this card, but I can say after a day's worth of testing that it seems like a solid sleeper and clone. Windows 95 actually detects and configures it automatically, no extra steps or drivers required! The card is identified as a Media Vision Jazz 16 (SoundBlaster / SoundBlaster Pro-compatible).

The software was straight-forward and easy to set up for DOS too. The FM synth is pleasing and to my ears nearly identical to an actual Sound Blaster (I know some of the more HC crowd here will probably say they sound nothing alike). No advanced features such as IDE- or daughter-board connections, but for basic DOS-gaming use this card seems to be completely fine, even if it doesn't stand above its peers in any way. The volume rocker in the back makes it easier to set the volume levels in DOS than using software, or if old-school passive computer speakers are plugged in.

leonardo wrote on 2023-02-01, 09:13:

Thank you! […]
Show full quote

Vic Zarratt wrote on 2023-01-31, 14:50:

If i come across the article then i'll do a quick scan and upload it to my webpage at some point - i'll then drop an update back here.

Thank you!

Vic Zarratt wrote on 2023-01-31, 14:50:

I'm sorry to hear that your soundblaster died on you like that Leonardo, it's gutting when things like that happen.
I do intend to sell some working isa soundcards on my local FB marketplace, including a ct3620...

Thank you for your condolences and kind offer as well, but it turns out a coworker had an old card stashed away and they offered it to me for free:
mediavision_jazz16.jpg

I had never even heard of this card, but I can say after a day's worth of testing that it seems like a solid sleeper and clone. Windows 95 actually detects and configures it automatically, no extra steps or drivers required! The card is identified as a Media Vision Jazz 16 (SoundBlaster / SoundBlaster Pro-compatible).

The software was straight-forward and easy to set up for DOS too. The FM synth is pleasing and to my ears nearly identical to an actual Sound Blaster (I know some of the more HC crowd here will probably say they sound nothing alike). No advanced features such as IDE- or daughter-board connections, but for basic DOS-gaming use this card seems to be completely fine, even if it doesn't stand above its peers in any way. The volume rocker in the back makes it easier to set the volume levels in DOS than using software, or if old-school passive computer speakers are plugged in.

Phew! All is good for as long as you've got a soundcard!

TheMobRules wrote on 2023-02-01, 03:15:

leonardo wrote on 2023-01-31, 10:19:

I'm on the edge of my seat. 😁

The BIOS didn't work unfortunately. No sign of life when turning on the system.

Which is strange given that the BIOS string is exactly the same in both, I wonder if there's any difference in the motherboard revisions that prevents this one from running.

Third Fourth time's a charm? I created this dump using DEBUG: AMISETUP (Protected Mode):

I also performed some tests with speedsys to see if sis.exe would improve memory output, but I saw absolutely no improvement at all (probably write-back is already enabled in BIOS). I did see the tiniest bump when I changed my cache timing from 3T to 2T though. 😁

All tests are attached.

edit: Swapped the dump created with DEBUG that had a bad checksum for the one created using AMISETUP with a proper one...

Hmm... it looks like my video card has a bug that prevents the use of linear frame buffer for some games/apps when more than 16MB of RAM is installed:

Jolaes76 wrote on 2012-06-03, 17:47:

...The Cirrus Logic 5428 has a hardware bug: if used together with more than 14 MB main memory, it fails to support linear frame buffer.

Does anyone know of a work-around to this when using UniVBE?

leonardo wrote on 2023-02-03, 17:31:

Hmm... it looks like my video card has a bug that prevents the use of linear frame buffer for some games/apps when more than 16MB of RAM is installed:

Jolaes76 wrote on 2012-06-03, 17:47:

...The Cirrus Logic 5428 has a hardware bug: if used together with more than 14 MB main memory, it fails to support linear frame buffer.

Does anyone know of a work-around to this when using UniVBE?

I do have the 1mb ISA version of the 5428, I tested it it once for signs of life but never really played with it.
I guess i could try it in my socket 7 with univbe 2003 and report back by the weekend, though does the bug exist in ISA versions or only the VLB ones?

Vic Zarratt wrote on 2023-02-06, 05:25:

leonardo wrote on 2023-02-03, 17:31:

Hmm... it looks like my video card has a bug that prevents the use of linear frame buffer for some games/apps when more than 16MB of RAM is installed:

Jolaes76 wrote on 2012-06-03, 17:47:

...The Cirrus Logic 5428 has a hardware bug: if used together with more than 14 MB main memory, it fails to support linear frame buffer.

Does anyone know of a work-around to this when using UniVBE?

I do have the 1mb ISA version of the 5428, I tested it it once for signs of life but never really played with it.
I guess i could try it in my socket 7 with univbe 2003 and report back by the weekend, though does the bug exist in ISA versions or only the VLB ones?

I actually didn't know about the bug until I went searching for the error I seem to be getting with certain games/apps that make use of UniVBE. I have not confirmed the bug in ways other than that UniVBE does report linear frame buffer as disabled/unavailable. I have no idea why this is a thing or if it is indeed a thing.

leonardo wrote on 2023-02-09, 12:05:

Vic Zarratt wrote on 2023-02-06, 05:25:

leonardo wrote on 2023-02-03, 17:31:

Hmm... it looks like my video card has a bug that prevents the use of linear frame buffer for some games/apps when more than 16MB of RAM is installed:

Does anyone know of a work-around to this when using UniVBE?

I do have the 1mb ISA version of the 5428, I tested it it once for signs of life but never really played with it.
I guess i could try it in my socket 7 with univbe 2003 and report back by the weekend, though does the bug exist in ISA versions or only the VLB ones?

I actually didn't know about the bug until I went searching for the error I seem to be getting with certain games/apps that make use of UniVBE. I have not confirmed the bug in ways other than that UniVBE does report linear frame buffer as disabled/unavailable. I have no idea why this is a thing or if it is indeed a thing.

Would it be of any use to you if i report back on testing the isa version on a 40mb pentium, to confirm if it's the 5428 chip itself or just poor VLB implementation?

Vic Zarratt wrote on 2023-02-10, 01:19:

leonardo wrote on 2023-02-09, 12:05:

Vic Zarratt wrote on 2023-02-06, 05:25:

I do have the 1mb ISA version of the 5428, I tested it it once for signs of life but never really played with it.
I guess i could try it in my socket 7 with univbe 2003 and report back by the weekend, though does the bug exist in ISA versions or only the VLB ones?

I actually didn't know about the bug until I went searching for the error I seem to be getting with certain games/apps that make use of UniVBE. I have not confirmed the bug in ways other than that UniVBE does report linear frame buffer as disabled/unavailable. I have no idea why this is a thing or if it is indeed a thing.

Would it be of any use to you if i report back on testing the isa version on a 40mb pentium, to confirm if it's the 5428 chip itself or just poor VLB implementation?

I don't know - more info is better I guess. I haven't run into that many applications / games that complain about this or that would potentially see a speed benefit to using VESA mode instead of standard mode.