Ran Speedsys on my mid-late 90s desktop trio, a 486DX4-100 build, a Compaq Presario 2200, and an IBM Aptiva 2170-175. The 486 and Compaq are two computers from my childhood, the IBM is just an old system someone gave me a couple years ago that I use as a "media intermediary" since it has both a 5.25in floppy drive and Zip drive.

Trying to run Speedsys on the Compaq sure was an adventure, because due to some sort of incompatibility the screen goes completely blank once it loads, and remains blank until a reboot. Not only that, but pressing the wrong keys causes it to hard-lock as well! Only by memorizing the key sequence was I able to successfully run it! 😵

On a side note...I have yet to find a single benchmark program that correctly measures the processor on the Compaq at its correct speed of 180MHz. The various utilities in Phil's benchmark pack all give wildly different (and incorrect) results, ranging anywhere from 50MHz to ~350MHz. 🤣

Interesting, I too have a Compaq Presario 2232 media GX, never managed to get speedsys to run. I didn't think to check wether it was in fact working under a blank screen. Yet to find a utility that correctly ID's the CPU either.

BSA Starfire wrote:

Interesting, I too have a Compaq Presario 2232 media GX, never managed to get speedsys to run. I didn't think to check wether it was in fact working under a blank screen. Yet to find a utility that correctly ID's the CPU either.

Start Speedsys, let it run for 5 mins or so for the memory test to finish. Check Caps Lock or Num Lock, if that responds you're good to continue. Then, hit spacebar once followed by Enter to run the hard drive test. After all drive activity stops, hit R then Enter to save the text and PCX results. At this point you'll need to turn the power off then on using the switch on the back, because CTRL+ALT+DEL will hard-lock it even if you manage to blindly get it to exit back to DOS.

Stuffed 1024MB (4x256MB) of SDRAM today in my Abit LX6. It looks like the C3 "Samuel 2" handles it just fine, the younger C3 "Nehemiah" didn't behave well when I passed 512MB in earlier attempts. CrystalCPUID doesn't detect correct original speed of the CPU, it's actually just 650MHz.

Finally got around to uploading a video to YouTube of me demonstrating a homemade dial-up server for my vintage computers that lack a network card. Can be used to connect them to the internet just for fun, or more productively to connect them to the LAN for (admittedly agonizingly slow) file sharing purposes.

https://www.youtube.com/watch?v=jpey-EQJe54

Wanted to use a properly stored and tested Cirrus Logic GD5424 in my in-progress 386 build, only to find out that sitting for 4-5 years in a box made it angry and was shaking/trembling the image/letters.
Go go gadget soldering iron, a few (well, quite a lot actually) capacitors later the image is rock solid again. It feels good to resurrect stuff!

torindkflt wrote:

Finally got around to uploading a video to YouTube of me demonstrating a homemade dial-up server for my vintage computers that lack a network card. Can be used to connect them to the internet just for fun, or more productively to connect them to the LAN for (admittedly agonizingly slow) file sharing purposes.

https://www.youtube.com/watch?v=jpey-EQJe54

Cool. I always thought it'd be a neat idea to do this to get some use out of my old modems (and to hook up computers without Ethernet).
Did not know about line voltage and that you have to generate a ring signal, thanks for the info!

stoof wrote:

torindkflt wrote:

Finally got around to uploading a video to YouTube of me demonstrating a homemade dial-up server for my vintage computers that lack a network card. Can be used to connect them to the internet just for fun, or more productively to connect them to the LAN for (admittedly agonizingly slow) file sharing purposes.

https://www.youtube.com/watch?v=jpey-EQJe54

Cool. I always thought it'd be a neat idea to do this to get some use out of my old modems (and to hook up computers without Ethernet).
Did not know about line voltage and that you have to generate a ring signal, thanks for the info!

Yeah, my vision is a matched XPS and PowerEdge going all the way back. Don't know if anything i'm interested in is ethernet incapable though.

I'm messing with a first revsion Asus A7V333 (REV. 1.01) VIA KT333 Socket A motherboard and a Thoroughbred B Athlon XP 2400+, released during the spring and summer 2002 respectively.

My plan is to see what kind of performance one could get from a pre nForce2 AMD Socket A setup and also to see if the Asus A7V333 is as bad as the forum threads from back then seem to indicate. The early revsions of the Asus A7V333 are supposed to have issues with 166 MHz FSB, general performance issues, general memory issues and alot of AGP issues. Even the later 2.xx revsions are supposed to only work half decently with Bios 1015 as BIOS 1014 and earlier had AGP issues that was not fixed until BIOS 1015 but then again broken in BIOS 1016, 1017 and 1018.

Although my board is the release version REV. 1.01 I have not yet found a single issue and 166 MHz FSB runs fine as long as I don't use the "turbo" chipset timings. I have not tested any other VIA KT333 motherboards so I don't know how the performance compares but the stability has been perfectly fine so far.

I was going to cheat a little by using a Thoroughbred B Athlon XP with a very late manufacturing date so it could mimic the "T.bred" 2600+, 2700+ and 2800+ but I found that the late ones I had were "super locked". When it comes to a super locked Athlon XP CPU it's impossible to adjust the multiplier unless one modify the CPU to present it self as an Athlon XP-M CPU and then only with CPUMSR or similar programs. I do not have an conductive silver ink pen although I should get one as graphite dosn't really work well on Athlon XP CPUs, for now I'm holding back on the modding.

It turned out that I diddn't really need a late revision Thoroughbred B to mimic the top Thoroughbred B speed grades (understatement). I actully think this early T.Bred Athlon XP 2400+ I'm using is better than the cherry picked T.Bred Athlon 2600+ Toms Hardare tested in this review from August 2002 http://www.tomshardware.com/reviews/At-The-La … P-2600,507.html. They used chilled water cooling and still needed 1.9V for 2400 MHz. I'm using this tiny and quiet Akasa cooler which dosn't have a larger footprint than the socket and the CPU runs great at 2400 MHz with 1.75V. 😀

I will probably post some results in the K7 thread tomorrow or early next week. 😀

Continued inventorying my motherboards (amongst other stuff).

It's slow progress partially due to me wanting to write down more details than just the motherboard number.

I did find something odd, the silkscreened jumper settings on the back of a 486 motherboard 😜

Part of the print is in mirror 🤣

Got myself a fully functional Dot Matrix Printer for my Packard Bell Pack-Mate 28 Plus, got some QBasic files, and installed Windows for Workgroups 3.11 onto the hard drive (CF to IDE).

I made already couple of videos about the printer aND the QBasic program and text editor that I ran (BAS file):

Citizen GSX-230 Dot Matrix Printer test

Microsoft QBasic QB Write

Started the replacement and recovery of the 486's failed/failing 8GB HDD.

Last night put a 15GB Quantum Fireball into the sled the Maxtor was living in, most trouble free install ever, plus read/write speed has improved greatly (it's the first time I've seen a 486 DX4-100 totally skip the Windows For Workgroups 3.11 boot logo, it reads that fast, though it did come back as I added back my drivers and other stuff). Granted, it's only using 8GB of that drive, but that's fine with me as I may try "dual boot" with something else eventually.

Maxblast DDO seems to work with darn near anything - thus far I've used it on my Maxtor 8GB drive, this 15GB Quantum, and a 20GB Western Digital. Uses the EZ-BIOS DDO, which is the same one my 286 uses as well on it's Seagate drive (though probably through an OEM Seagate DDO diskette and not a Maxtor one like I'm using). Then installed MS-DOS 6.22 and Windows For Workgroups 3.11.....new notes....

- Just leave the NIC NE2000 compatible
- I really need to make a floppy diskette with TCP/IP 32B on it
- I've been doing this waaaay too long, I've got every setup step memorized, rebuilding this system is easy and fast

Process was something like this
- install new drive
- install DDO, set 4 FAT-16 partitions (1 2.1GB for Boot, and 3 others as Logical DOS drives in the Extended DOS Partition)
- Install MS-DOS 6.22
- use Windows 98 SE Boot Disk to copy OAKCDROM.SYS from diskette, then edit that into my Autoexec.BAT and Config.sys
- reboot, install Windows for Workgroups 3.11
- When setting up network I really should have just left my adapter with the generic NE2000 driver, it works best with that
- swap out drives, go to Windows 95 drive, copy TCPIP32B from my Win10 machine to the 486, make floppy with TCPIP32B on it
- swap out drives again, go to DOS, make C:\TEMP dir, copy TCPIP32B.EXe to the directory and extract it
- run network setup in WFWG again (Win /N) change to NE2000 Compatible with TCP/IP 32B as the only protocol
- reboot
- install Opera 3.62....dangit, I don't have my key..... (yes, I own this version, I "bought" it a few years ago, someone from Opera just gave me a key with my name for free because it's an old version)
- install drivers for my SoundBlaster AWE64
- install AfterDark 3.2
- setup shares in File Manager
- start copying all the crap from the old drive over while I sleep (had a backup of MOST of it).....but.....

This morning finished backing up all the other crap I need to a 16GB Flash Drive.....well, I can't use Win10, so I wound up using another USB with WinPE for Windows 7 because as it turns out, Windows 10 cannot handle FAT-16 partitions properly. Hmmm....XCOPY was my friend here getting roughly about 9 or so GB worth of programs, data, and my key for Opera 3.62 over. Plugged the drive into a USB to SATA/PATA cconverter and power source and went about copying - hit a few more disk errors in the process, that old Maxtor was really on it's way out. Glad I decided to tackle this when I did. Time to upload some of the more irreplaceable stuff to the cloud so I still have it.

Already, only roughly an hour or so of actual work and I'm almost back to where I was (save for waiting for file copy operations - which ARE going to take awhile when you have almost 2.1GB of games on one of the partitions).

I have been testing/benching vlb vga cards on my Abit 486 board.
It is running with a dx2-66 write through, 128kb 20ns cache.
I know this is just a quick and dirty benchmarking, yet it tells me what to expect
from a final build in the future. I am quite happy with this setup with the S3.

Are these results the standard results that are to expected?

CL-5428-VLB 1mb:
3D-Bench 1.0 = 45.3
3D-Bench 1.0c = 44.3

S3-805-VLB 1mb:
3D-Bench 1.0 = 47.6
3D-Bench 1.0c = 47.4

Here's a fun story.. I'm putting together a Pentium PC from scrap parts I've collected from here and there. The board fives me the following error at every boot:

CMOS checksum error - Defaults loaded
CMOS vattery failed

Press F1 to continue, DEL to enter BIOS Setup

The board is almost 20 years old so I assumed the battery was dead so I didn't think much of it. However, yesterday I actually replaced the CR2032 battery and.. it did not resolve the issue.
When the PC is plugged to the wall, the BIOS settings hold.
When the PC in unplugged, I get the same errors.
There are 2 3-pin jumpers right below the BIOS chip that apparently controls two things: 12V/5v for BIOS Flash Chip, and CMOS Reset. These pins are undocumented online (PCChips A101 v1.04 board, manual is unavailable online) so I had to resort to some experimenting.
JP2 (12v/5v) was shipped in 2-3 position and I did not touch that lest I damage something.. The BIOS chip is a Holtek chip, googling the text on the chip yielded absolutely nothing so I don't know if it's 12V or 5V but I do know that the current BIOS is Small Block and 03/03/97-i430VX-02071997C-00
Jp3 (CMOS) was shipped at 2-3, which I believe is almost always the Clear CMOS position on any board I've ever seen, so I moved that to 1-2 instead
I then smiled to myself smugly, thinking I had fixed it.. But no, the same behaviour continued.
So I moved the CMOS Reset jumper back to 1-2, and the board started losing BIOS settings at even soft reboots.
This was when I decided something had to be wrong with the battery holder, so I removed the PCI cards, and the battery to inspect it. There is absolutely no corrosion and the connectors touch solidly. The battery is solid 3V on the voltmeter. There is a slight crack in the plastic that makes the battery slightly wobbly but I can't be sure if that's the problem, so I start putting everything back together..
And when doing this, I realized something very strange - the VGA port of the PCI VGA I am using, a Matrox Mystique 220, actually presses onto the battery when folly inserted and screwed. It presses onto the battery so much that the bottom connector flattens and apparently it has trouble contacting the upper connector.
I really can't believe this could be the real reason, at this point I'm still worried that I may have a bad CMOS chip instead, but I move the cards around in the PC so that there is nothing on the PCI slot above the battery..
Reboot with JP3 in 1-2 position.. And everything works? I unplut for 30 seconds, plug in, and it still works?

Has anyone ever experienced something like this? I'm really hoping I fixed this issue and my first retro build can actually hold BIOS settings when I get home today.

I also found a patched Award BIOS for my board that allows up to 64GB Hard drives, so I'll try flashing that at the earliest opportunity 😀 A question, by the by, is there any reason to not use the latest awdflash.exe to flash an award bios?

appiah4 wrote:

Here's a fun story.. I'm putting together a Pentium PC from scrap parts I've collected from here and there. The board fives me […]
Show full quote

Here's a fun story.. I'm putting together a Pentium PC from scrap parts I've collected from here and there. The board fives me the following error at every boot:

CMOS checksum error - Defaults loaded
CMOS vattery failed

Press F1 to continue, DEL to enter BIOS Setup

The board is almost 20 years old so I assumed the battery was dead so I didn't think much of it. However, yesterday I actually replaced the CR2032 battery and.. it did not resolve the issue.
When the PC is plugged to the wall, the BIOS settings hold.
When the PC in unplugged, I get the same errors.
There are 2 3-pin jumpers right below the BIOS chip that apparently controls two things: 12V/5v for BIOS Flash Chip, and CMOS Reset. These pins are undocumented online (PCChips A101 v1.04 board, manual is unavailable online) so I had to resort to some experimenting.
JP2 (12v/5v) was shipped in 2-3 position and I did not touch that lest I damage something.. The BIOS chip is a Holtek chip, googling the text on the chip yielded absolutely nothing so I don't know if it's 12V or 5V but I do know that the current BIOS is Small Block and 03/03/97-i430VX-02071997C-00
Jp3 (CMOS) was shipped at 2-3, which I believe is almost always the Clear CMOS position on any board I've ever seen, so I moved that to 1-2 instead
I then smiled to myself smugly, thinking I had fixed it.. But no, the same behaviour continued.
So I moved the CMOS Reset jumper back to 1-2, and the board started losing BIOS settings at even soft reboots.
This was when I decided something had to be wrong with the battery holder, so I removed the PCI cards, and the battery to inspect it. There is absolutely no corrosion and the connectors touch solidly. The battery is solid 3V on the voltmeter. There is a slight crack in the plastic that makes the battery slightly wobbly but I can't be sure if that's the problem, so I start putting everything back together..
And when doing this, I realized something very strange - the VGA port of the PCI VGA I am using, a Matrox Mystique 220, actually presses onto the battery when folly inserted and screwed. It presses onto the battery so much that the bottom connector flattens and apparently it has trouble contacting the upper connector.
I really can't believe this could be the real reason, at this point I'm still worried that I may have a bad CMOS chip instead, but I move the cards around in the PC so that there is nothing on the PCI slot above the battery..
Reboot with JP3 in 1-2 position.. And everything works? I unplut for 30 seconds, plug in, and it still works?

Has anyone ever experienced something like this? I'm really hoping I fixed this issue and my first retro build can actually hold BIOS settings when I get home today.

I also found a patched Award BIOS for my board that allows up to 64GB Hard drives, so I'll try flashing that at the earliest opportunity 😀 A question, by the by, is there any reason to not use the latest awdflash.exe to flash an award bios?

That was actually a fun read, well spotted! 😁

These are actually the kinds of problems that will ever happen twice when using different hardware, but it helps to have an eye for details here and spot anything that may be out of the ordinary. Spotting the battery issue along with noticing the Matrox card was pressing down on it (after having checked that the battery holder itself was in good-looking condition) is a way of solving problems that are unexpected and happen so rarely that one will probably have never heard of it (not even experienced people).

This happens a bit more frequently when it comes to the rig working perfectly fine, but problems appearing when the side panels are put back and pressing or warping something that makes a problem that was juuust about not to happen to a problem that juuust did start to happen 🤣.

Well done 😜

I've been reading the MIDI specs, MPU-401 and IBM Music Feature Card Technical Reference manuals, in preparation of adding MIDI support to my 8088/DOS VGM player.

Picked up a trio of as-is old laptops from a semi-official "Ewaste drop off that also allows you to buy parts and pieces from the skids of parts people dropped off" I found near work. Spent 20 bucks CAD.

Got a Toshiba Satellite T1910CS, a Toshiba Satellite 110CS, and a Compaq Presario 1200.

The two Toshiba's worked right away - the T1910CS is a 33mhz 486 with 4 MB of RAM and the 110CS is an early Pentium with 40 MB of RAM. Hard drives and 3.5" drives in both work great.

The Compaq... well the Power Cord attached light comes on, so at least that circuitry works... but no other signs of life just yet. I'll pull it apart and see what I can find - this one'll be nice if I can get it working, nice 1998-1999 era laptop I can use for mid 90's portable gaming.

Finished setting up the 486's new DOS 6.22/WFWG311 drive last night.

- installed win32s - got a newer version than the outdated version I've been using for years
- used cutemouse instead of AMOUSE.EXE which I've had since I bought a Radio Shack serial mouse in 2001
- installed all of the SIM games
- got all my games copied over to the E partition
- got all my DOS applications copied over to the D partition
- setup a Driver Repository that's better organized on the G drive (this is for when the O/S gets sick and needs reformatted or whatever, makes installing drivers and programs easier)
- Setup multi-boot configuration and honed it down this time to 4 options instead of seven, I stuck with the same options I use most of the time (one with EMM386, one without, one clean, one for Windows)
- ran MEMMAKER so the maxxed out configs have 590K of Free conventional memory so that things like Destruction Derby and Freddy Pharkas run
- copied my game dev software over

I now need to copy some stuff over from some of the game Ideas I'm working on graphics for....I've got something like 5-6 ideas, 3 in serious development, the only one I can mention is a DOS based Five Nights at Freddy's fangame of sorts using the AGS game engine....why I'm doing this? Mostly just for fun.

My idea on how this game works is this....
- player stays disabled as it's in first person, though I've been tempted to put a graphical adventure element in there where the character CAN leave the office, sort of like the underground catacomb scene from MOnkey Island 2: Lechuck's revenge where LeChuck chases him around....hehehehe.

- the animatronics are turned on and off like objects in the room depending on if they are present, so if say, Chica walks into the east hall corridor, instead of showing a totally different room, I can just turn on Chica's graphics for that room and even animate them if I want to. I may also be able to do this by teleporting the characters to a set of coordinates based on a timer, or even just turning an object on or off if they are in the room.

- The above saves me time in the artwork department because instead I can overlay graphics over pre-existing rooms instead of separate screens/rooms for each camera in each state. Plus I can have more than one animatronic per room if I want.

- Also, I can create the jumpscares, I've seen how fast this thing throws full screen animations in AGS up at full speed, I'm pretty sure I can make a terrifying animatronic screaming/biting your head off in 320X200 8bit color.

Scali wrote:

I've been reading the MIDI specs, MPU-401 and IBM Music Feature Card Technical Reference manuals, in preparation of adding MIDI support to my 8088/DOS VGM player.

Today I wrote a rough first draft of the MIDI player... It can now play CANYON.MID through an MPU-401 UART compatible interface.
Perhaps tomorrow I can try adding some support for the IBM Music Feature Card.

@Scali: cool stuff! Looking forward to your next improvements. 😎

Tonight I finally dug into the PowerMac 7200 I've had for TWO YEARS with the intent of getting it going. Pulled it apart, everything was tidy & clean inside; no leaky battery or bulging caps, so that's nice! But it's TOTALLY stripped. Found System 7.6 ready to install on original floppies, when did I even get those?? Slid in a SCSI HDD & CD-ROM from a junked 6100, didn't even have to change the caddy. No ROM SIMM. No problem, internet says the ROM from the 6100 will work. No VRAM. No problem, internet says there is 1MB onboard.

No RAM. Oh look, it takes SDRAM. That should be easy right?

Nope ... it's not SDRAM, it's 168-pin *5V* RAM in the same form factor as SDRAM. Oh apple...

Wait, I have a stick of that somewhere! I rifle through every box & storage bin looking for all the places I might stuff RAM. Find my precious DIMM. 64MB, 5V, 60ns. YES, perfect! Wait, it's EDO. Internet says this is one of THREE specific models Apple made that will NOT take EDO AT ALL and will infact become damaged if you try to stick it in. WTF.

I need 5V *FAST PAGE* memory ONLY, on a 168-pin DIMM.

Where do I even???

Looks like I'm putting everything back in its cardboard box and banishing it to the back of the cupboard, hopefully for another decade.

BLAGH.