You could buy a 72p RAM expansion board for AWE64. Then you can test single modules.
No guarantees above 28 Megs. 😁

Refurbished another FX5900 XT 😀

Recapping a socket 939 board.

Silly cap, you're supposed to be 3300uF, why'd you have to burst like that?

I wrapped up my experiments with OS/2 version 2.0 with Service Pack 2 (Fixpak?) on my Pentium 4 overkill retro box. I was surprised to learn that the OS/2 multimedia pack that shipped with OS/2 2.1 will install and function on version 2.o (at least with the Service Pack). I couldn't find much documentation online that it was even possible other than references to an IBM multimedia computer that shipped around the same time. Even better, the improved Sound Blaster 16 driver from Creative Labs works perfectly including the "external MIDI" from the attached Dreamblaster S2.

The only problem I encounter was that I couldn't get sound FX working properly in either DOS or WIN-OS2 (version 3.0a). I could get a Windows CD player to work with analog stereo sound, but I couldn't find a Win 3.0 sound driver that would properly detect the Sound Blaster 16 in Windows (it complained at the DMA detection stage).

Creative Labs "DIAGNOSE.EXE" program properly detected the Sound Blaster 16 in a DOS box and even played the test 8-bit and 16-bit sounds as well as ADLIB music. Unfortunately, every game I tried either crashed the DOS box or crashed the system if I tried to play with sound FX. General MIDI through the Dreamblaster S2 did work in most games I tried in a DOS box.

For good measure, I also installed networking with TCPIP. I could basically only run Ping and FTP, but it still provided some internet capability.

I wish I knew how to screen capture in OS/2 on real hardware. The only screen capture utility I could find for OS/2 version 2.x captures text boxes, not the desktop. I'd love to show off OS/2 in all of its 1280x1024, 256-color glory (using an ATI Mach32 graphics card).

Even late in the life of OS/2 I think programmers would have thought storage performance so low in comparison to data capture required, that it would be pointless, even at 5 frames a second. Any real time compression would be out also, didn't have the Mhz for that either.

stamasd wrote on 2022-02-20, 21:25:

Recapping a socket 939 board.

Those are probably low ESR capacitors. Make sure that you replace them with the right ones.

The ones I'm getting are rated for ESR of 0.015 ohms, should be good enough.

stamasd wrote on 2022-02-20, 21:25:

Recapping a socket 939 board. […]
Show full quote

Recapping a socket 939 board.

Silly cap, you're supposed to be 3300uF, why'd you have to burst like that?

Okay after recapping the motherboard, it works... sort of. I mean it works, POSTs and boots consistently, I can get into the BIOS without it freezing etc.

But it takes a very long time to POST. Like, 5 minutes long between printing the CPU ID and the moment when it starts counting memory. After it starts counting the memory, all is fine and proceeds at the normal speed.
And it does this delayed POST very consistently, every single time. I've never seen a motherboard do that.

stamasd wrote on 2022-02-22, 00:15:

Okay after recapping the motherboard, it works... sort of. I mean it works, POSTs and boots consistently, I can get into the BIO […]
Show full quote

stamasd wrote on 2022-02-20, 21:25:

Recapping a socket 939 board. […]
Show full quote

Recapping a socket 939 board.

Silly cap, you're supposed to be 3300uF, why'd you have to burst like that?

Okay after recapping the motherboard, it works... sort of. I mean it works, POSTs and boots consistently, I can get into the BIOS without it freezing etc.

But it takes a very long time to POST. Like, 5 minutes long between printing the CPU ID and the moment when it starts counting memory. After it starts counting the memory, all is fine and proceeds at the normal speed.
And it does this delayed POST very consistently, every single time. I've never seen a motherboard do that.

What happens if you press reset?
Cold boot, hot boot; resets? I guess you tried all already.

Reset gives cold boot. But even a warm boot (ctrl-alt-del after boot) results in the same long delay during POST. Just without memory count at the end of the delay.

I have a motherboard that I recapped which doesn't shutdown or reboot properly from Windows 98/ME. Shutdown will usually work but reboot will always hang the system on a black screen with a blinking cursor, requiring a hard reset. Either option will always trigger the startup scandisk which detects an incorrect shutdown (even though normal shutdown turns the computer off completely).

I only noticed this problem after I replaced the stock capacitors. I used genuine Nichicon capacitors ordered from Digi-Key. I don't know if the issue existed prior to replacing the capacitors since I hadn't done this exact test.

I have suspected that the capacitors I used as replacements are not low ESR, which is why I offered the warning. The ones I used are Nichcon HV(M) series which are marketed as low impedance. I'm not even sure if this has anything to do with the problem I'm experiencing, as I only have the one board.

Well I've cleared the CMOS a few times, didn't help. Flashed the latest BIOS version 1.D from the original 1.A and it also didn't help.
I noticed that during the time it waits during POST, if I press the NumLock key, all 3 LEDs on the keyboard flash, then go blank. Do this 5-6 times in a row, and eventually it continues to POST. So it shortens the duration of the wait during POST. Weird.

You're in the same boat as me now. Without a working motherboard with the stock capacitors, we have no way of confirming whether or not this is the original behavior.

I guess I'll go ahead and install XP and see if it's stable. 😀

Kahenraz wrote on 2022-02-22, 01:10:

I have a motherboard that I recapped which doesn't shutdown or reboot properly from Windows 98/ME. Shutdown will usually work but reboot will always hang the system on a black screen with a blinking cursor, requiring a hard reset. Either option will always trigger the startup scandisk which detects an incorrect shutdown (even though normal shutdown turns the computer off completely).

I only noticed this problem after I replaced the stock capacitors. I used genuine Nichicon capacitors ordered from Digi-Key. I don't know if the issue existed prior to replacing the capacitors since I hadn't done this exact test.

I have suspected that the capacitors I used as replacements are not low ESR, which is why I offered the warning. The ones I used are Nichcon HV(M) series which are marketed as low impedance. I'm not even sure if this has anything to do with the problem I'm experiencing, as I only have the one board.

The last time I put my Soyo P4 board through it's paces was after the VRM recap, since then Iv done the rest of the primaries and big-medium secondary caps and the power switch wouldn't power it off, but the reset switch worked perfectly. However if I tried to put it to sleep it would soft crash XP

Anybody familiar with a DEC Dazzle Multimedia Rev 3 PCI card? I got this thing in a lot and I can find precisely fuck all about it. Apparently it was made in 1997 with a Digital Multimedia 21230 at its core from 1995 (https://techmonitor.ai/techonology/dec_chip_h … desktop_editing) with what appears to be 1MB of Alliance memory on the back end. Just a fancy MPEG1 card I'm guessing? Any practical usage?

Also acquired a Rage 128 AiW, a Matrox G2+ with memory expansion module, and a SB Live 5.1 in the same lot. Not bad for something I received for free.

I guess it may help if you plan on using your retro machine to make and playback VCDs. 😀

stamasd wrote on 2022-02-22, 01:38:

Well I've cleared the CMOS a few times, didn't help. Flashed the latest BIOS version 1.D from the original 1.A and it also didn't help.
I noticed that during the time it waits during POST, if I press the NumLock key, all 3 LEDs on the keyboard flash, then go blank. Do this 5-6 times in a row, and eventually it continues to POST. So it shortens the duration of the wait during POST. Weird.

What happens without a keyboard? Same 5 minutes wait?

I didn't try that. I may do it tonight when I get home. I did set the CMOS to "halt all errors except keyboard" so that should work.
FWIW I set the above because, at first POST after recapping, it gave me a "keyboard error, press F1 to continue" - despite the keyboard working fine, pressing F1 got it past that etc. This is also a keyboard that works flawlessly on all other retro machines I have (a dozen or so). Come to think of it, after I flashed the newer BIOS it did not give me that keyboard error anymore, but I still set that parameter in the CMOS anyway.

Working on a project on a 1982 TRS-80 model 4 with 128K (also runs on a TRS-80 Model 3 with 48K), to replace the need for old Floppy Disk Controller board as well as floppy drives, by connecting to the FDC port of the mainboard and running 4 floppy drives from an SD-CARD, here is a peek of the work at hand; it handles so far LDos 531, LDos 631, CPM v2.2, TRSDos 1.3, and RapidDos 1.3 (more under testing) will be completed with a menu loader to mount images on the fly from DMK or HFE files format. so far has been tested with FREHD as the booting disk and FreHD hard drive images for LDOS 531, 631, CPMv2.2 and NewDos 80v2, carrying up to 5 x 7MB HDD Partitions (35MB Total) with 2 Floppy Drives (180k, 360k or 720k formats supported) UNit has been tested to work without issues from 2, 4, and 6 Mhz Z80 CPUs systems, further testing to be done on 8MHZ System later.

yes I also use an Amber Monitor on my Model 4