VOGONS


Reply 240 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

I've started to prepare the next episode and I must say it's going to be a MONSTER.

I'm going to take TL;DR to the NEXT LEVEL!!! 😁 90 pictures ready! 😁

EXTRA: some good ol' fun with my first PC 😁 45 pics prepared.

Scandal - Just Let Me Dance (Maxxi Soundsystem Remix)

https://www.youtube.com/watch?v=Ir-mZaTAxRU

Attachments

  • VLI486.jpg
    Filename
    VLI486.jpg
    File size
    212.33 KiB
    Views
    1658 views
    File license
    Fair use/fair dealing exception
  • RIVA.jpg
    Filename
    RIVA.jpg
    File size
    113.59 KiB
    Views
    1658 views
    File license
    Fair use/fair dealing exception

Reply 242 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

HUH 😁 I wish I had a better video card and a better CPU but I think that can be arranged soon 😁 for now I'm taking it easy 😁

NEXT WEEK I'M WAITING FOR A VERY SPECIAL DELIVERY!!! 😁 I'M HYPED!!!!

Attachments

  • VLI486-026.jpg
    Filename
    VLI486-026.jpg
    File size
    1.26 MiB
    Views
    1644 views
    File license
    Fair use/fair dealing exception

Reply 243 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

Almost, VLB Heaven...

For more than two months I didnt buy anything retro-HW related. While I was in REHAB 😁, I thought I got over the worst part of this addiction. Unfortunately for me, I was wrong 😁 as this week I'm waiting for a package which contains MORE RETRO HW 😁. YEP! I did it again...need I say more?.... It's not a prototype, a monstrous graphic card, an IBM 350 (RAMAC) or something extremely rare but the contents make me remember the times in highschool in 1994...my memory is still fresh. The first time I sat on the chair that I used for the next four years, which had carved on the back the words: IBM PC and between my legs: Don't worry / Be Happy . Go figure 😀

Let's return to the present day and the story I'm about to tell you.

Almost, VLB Heaven.

Why Almost and Heaven in combination with VLB? What made me use these words together? Well, I wish I had a more powerfull mix of components for this endeavour as it would've been even more enjoyable. After this experience, I've aquired the taste for even older parts. While the newer parts, usually have a higher casualty rate, the older parts have yet to disappoint me and they still work "knee deep in the dead", after taking on the Apocalypse from DOOM, being left to the mercy of elements, have escaped the shredder or got away from the vandals who see them only as a GOLD Recovery source.

Up until this moment, when I heard about VLB stuff, I was thinking about something complicated which was also shrouded in a mysterious fog. Why should I bother with something like that? (Little did I know that in the end I was going to like them A LOT!)

In September 2017 through a stroke of luck, I bought from the flea market, an ASUS VL/I-486SV2GX4 REV. 2.0 motherboard and a mighty - FDD/HDD/SERIAL/PARALLEL/GAME PORT Winbond W83787F, W83758F / Promise PDC20630- EIDE / Super-I/O VLB Controller. I paid a very low price for the entire package. The Asus smilled at me the moment I saw her even if the first time I didnt bought it and I had to get back from my car and recover it from the pile in which it sat. Looking back, I say it was a good decision as I'm pretty sure, it would've haunted me beyond the grave if I wouldn't've done so 😁

This is how I was bitten by the VLB bug. GOD DAMN IT!!! It's crawling under my skin and I CAN'T ESCAPE!!! Something has changed inside me and I can't turn back!

First I cleaned the VLB controller. Without this piece, the ASUS motherboard is just an inert mass. NO HDD, NO mouse, NO CD-ROM, zilch, nada, ZERO. A BIG FAT ZERO! I must say again.

The VLB controller, was in a so-so state. Dusty and dirty with a bent bracket which also lost its shine. The good part was that it was complete.

VLBCTRL-001.jpgVLBCTRL-002.jpgVLBCTRL-003.jpg

I took off all the jumpers, about 30+ of them and I cleaned them with an old toothbrush and isopropyl alcohol 99%.

VLBCTRL-004.jpg

Next came the cleaning of the board with plenty of isopropyl alcohol 99%. Before this stage, I took the decision to remove the surface dust using a soft animal hair brush as I was pretty sure it would've made my life a whole lot worse when I was to use the alcohol. IT WAS THE RIGHT MOVE! as everything went smoothly. I paid great attention to detail. It was my first VLB Controller, after all.

VLBCTRL-005.jpgVLBCTRL-006.jpgVLBCTRL-007.jpgVLBCTRL-008.jpgVLBCTRL-009.jpgVLBCTRL-010.jpgVLBCTRL-011.jpg

The bracket got back its shine and I put back all the jumpers. It took me a while, but the pictures I made earlier were of big help. I didnt have a manual for the controller and if I would've had missing jumpers, this operation would've been a PITA even with the instructions printed on the back.

VLBCTRL-012.jpg

Final results. FLAWLESS VICTORY!!!

VLBCTRL-013.jpgVLBCTRL-014.jpgVLBCTRL-015.jpgVLBCTRL-016.jpgVLBCTRL-017.jpgVLBCTRL-018.jpgVLBCTRL-019.jpg

gallery: https://postimg.cc/gallery/15c7bv6yk/

As I said before, I found the ASUS motherboard and the VLB controller at the flea market. I took a picture after I bought them.

VLI486-001.jpg

ASUS VL/I-486SV2GX4 REV. 2.0

This motherboar was my ticket in a new stage from the evolution of the PC. As soon as I saw the printed name on the motherboard, I knew I was staring at an ASUS board and my mind was racing towards the dual socket Pentium PRO motherboards. The font is specific to the boards belonging to this era. An ASUS is still an ASUS no matter what.

As expected, the motherboard was dirty and had that specific smell which I cant stand to this day even after more than a hundred cleaned components. Come to POPPA to make you GREAT AGAIN! 😁

VLI486-002.jpgVLI486-003.jpgVLI486-004.jpgVLI486-005.jpgVLI486-006.jpg

I took off all the CACHE chips, the BIOS chip and the jumpers. The jumpers were brittle and I knew I had to replace them. ISOPROPYL ALCOHOL 99%. Need I say more? I DON'T THINK SO!

VLI486-007.jpgVLI486-008.jpgVLI486-009.jpgVLI486-010.jpgVLI486-011.jpgVLI486-012.jpgVLI486-013.jpgVLI486-014.jpgVLI486-015.jpgVLI486-016.jpgVLI486-017.jpgVLI486-018.jpgVLI486-019.jpgVLI486-020.jpgVLI486-021.jpgVLI486-022.jpgVLI486-023.jpgVLI486-024.jpgVLI486-025.jpgVLI486-026.jpg

After I cleaned the motherboard I started to stare at it and the thought of populating all the CACHE sockets came into my mind. There were only 128KB and four chips. The motherboard had 8 sockets! Let's DO IT! 256KB!.......Little did I know this would mean more work and less FUN!

I checked the manual and I found that for 128KB I needed 4x32KB CACHE chips and a 8KB TAG chip. CHECK! I didnt have 64KB chips so I had to use 8x32KB instead of 4x64KB. For the 8x32KB I needed a 32KB TAG chip. At this moment, I didnt know if the TAG chip had to be made by the same manufacturer of the CACHE chips or if it would've had to be of a different construction or of a different brand. I searched vogons.org and the Internet and I found out that these facts dont matter too much as long as the CACHE chips and the TAG have the same specs as stated in the manual. So I searched in my stash of organ donors and parts that will not be restored, and I found on a motherboard 9 identical 32KB CACHE chips. SUPER. Now the ASUS had 256KB CACHE and it looked wicked.

https://www.philscomputerlab.com/asus-vli-486sv2gx4.html

I didnt have a tool for the extraction of the CACHE chips so this operation was nerve racking. I didnt know that I will repeat this process at least three times...😁

VLI486-027.jpgVLI486-028.jpgVLI486-029.jpgVLI486-030.jpg

OK. All was clean and shiny, and I wanted to pick a CPU to test the motherboard. I found from the Internet that if I was to use an AMD 5x86 de 133MHz the motherboard would need a specific BIOS version to recognize it and I was pretty sure that it didnt have the correct one. Besides, I couldnt FLASH the BIOS chip because the motherboard and the chip didnt support this. DEAD END! I took from my stash the next BIG thing, an AMD 486 DX4-100MHz 100NV8T CPU.

I set up all the required jumpers and I took some pictures with the wonder. I also replace the mix of black jumpers with new white ones. I had to use a pair of pliers with smooth jaws to straighten some pins. Now the motherboard look GREAT! WICKED!!!

VLI486-031.jpgVLI486-032.jpgVLI486-033.jpgVLI486-034.jpgVLI486-035.jpgVLI486-036.jpgVLI486-037.jpgVLI486-038.jpgVLI486-039.jpgVLI486-040.jpgVLI486-041.jpgVLI486-042.jpgVLI486-043.jpgVLI486-044.jpgVLI486-045.jpgVLI486-046.jpg

The following day, when I tought I had everything ready for the real test, something was bugging me. The Revision 2.0 of the motherboard doesnt have settings for 3V, and my AMD DX4-100 MHz IS 3V! Even if I had jumpers settings for this particular CPU, I decided not to pump 3.45V in a 3V rated part, even if I was told I could do so. The 5x86 133MHz is rated at 3.45V but the motherboard wont recognize it...

My board only has the JP32 jumper and in combination with the JP16 I can get: 3.45V, 3.6V, 4.0V and 5.0 V.

If JP16 is shorted on 1-2, then JP32 sets the voltage. In which case,
JP32 1-2 = 3.45 V
JP32 2-3 = 3.6 V
JP32 OPEN = 4.0 V

If JP16 is OPEN, then the voltage is 5.0 V

Asus VL/I-486SV2GX4 rev. 2.0. - Voltage selection

https://groups.google.com/forum/#!topic/alt.c … sus/svoQg7GSRbQ
http://www.elhvb.com/mobokive/Archive/Asus/48 … 4/sv2g-cpu.html
http://www.dosforum.de/viewtopic.php?f=1&t=8175

I started to look in my boxes and what do I find? The Intel 486SX de 25MHz which I saved from demise in 2015 and it was featured in the episode: Little bent PIN -
The story of the BENT CPU PIN(ssssssssss) or should I say lots of CPU Pins 😁

Re: Hello, World! :D - Robert B's PC builds - oogle away freely :)

5V? YEAH BABY! This slow 4x86 will help me test this high-end motherboard. GREAT!!! I have to set again all the jumpers...DON'T forget! 25MHz BUS instead of 33MHz!

VLI486-047.jpg

I got everything ready for the test. I powered up the Minebea PSU from the previous episode and I waited to hear the HAPPY BOOT BEEP.

NOTHING! NO BEEP, NO IMAGE on the screen! Just a cricket chirping....😁

What THE....?!?! I used another stick of RAM. STILL NOTHING. I took out all of the RAM sticks .....NOTHING! Well I'll be F.....

I changed the PC SPEAKER as it might be damaged. NOTHING!!! Hmmmm

I tried to connect an FDD and a HDD. I changed a few jumpers on the VLB controller.

STIL NOTHING!

THOSE CACHE CHIPS might be the culprit! There simply is no other explication. I took out all of them and I powered the system. BEEP BEEP!!! I'm alive! I still had no image on the screen... the card Oak Technology OTI077 - 1077082003 REV. G - 512KB was dead...RIP

VLI486-070.jpg

I took out another video card: a Trident TVGA9000C - 7133 Rev. B1 - 512KB and BAM! I got a clear image. All was OK as it should.

VLI486-053.jpg

I tried my third ISA card: a Trident TVGA9000i-2 - 7210 Rev. H1 and I received more good news. It was working!

VLI486-054.jpg

At this moment I decided to put back the original CACHE chips and combine them with other 32 pin 32KB CACHE chips plus another 32KB TAG. This worked and I got the 256KB I so much desired.

VLI486-048.jpgVLI486-049.jpgVLI486-050.jpgVLI486-051.jpgVLI486-052.jpg

The motherboard is finicky about RAM in regards to capacity, manufacturer, type: SS or DS and the way you fill the slots. The MANUAL was of great help. I managed to get a maximum of 16MB as I only had 4MB and 8MB sticks. The two 16MB sticks I have are in the 5x86 build and they arent going anywhere no matter what.

VLI486-055.jpgVLI486-056.jpgVLI486-057.jpgVLI486-058.jpgVLI486-059.jpgVLI486-060.jpg

I ran a few tests but I didnt like what I saw. Most likely, THE CPU IS WAY TOO SLOW for this motherboard. The CACHE CHECK program complained about supposedly fake cache chips. It is not the case, as the CPU, the MIX of RAM, the MIX of CACHE chips might be blamed for this behaviour...I didnt set anything in BIOS and only the defaults were loaded...All these "problems" can be easily fixed in the future.

VLI486-061.jpgVLI486-062.jpgVLI486-063.jpgVLI486-064.jpgVLI486-065.jpgVLI486-066.jpgVLI486-067.jpg
VLI486-068.jpgVLI486-069.jpg

During testing I decided to use a cooler for the 4x86 SX 25MHz as it was quite warm to the touch. It can be used without a heatsink but I wouldnt do this over long periods of time.

VLI486-071.jpg

In the end I reverted to the 128KB CACHE and the original chips. A bigger CACHE wont help me if I dont have a RAM capacity to match.

IF IT WORKS DON'T FIX IT!!! 😁

VLI486-072.jpg

Here we are at the end of the : "Almost, VLB Heaven..." story. I'm pretty sure you understand the title by now. It's not about the difficulties I encountered when I wanted to test the system it is about the fact that a more powerful CPU, more RAM, an ET 4000 VLB video card, etc... would've made all of this more enjoyable.

At the very least I'm glad that the ASUS and the VLB controller are in good working condition.

I might revisit this VLB moment sometime in the future...we'll see.

gallery: https://postimg.cc/gallery/1wjod5nuk/

Last edited by Robert B on 2018-04-20, 05:49. Edited 4 times in total.

Reply 244 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

EXTRA: Is it possible to use a RIVA 128 4MB PCI in a SiS 496/497 socket 3 motherboard? (I'm sure this question keeps you awake at night! 😁 )

I was asked by a fellow vogons member: amadeus777999 if I could test a RIVA 128 in my socket 3, 5x86 133MHz, SiS 496/497 build. In his case all his efforts were fruitless as the system hanged at the stage he tried to install the drivers in Windows. In the end he found a solution as he had to set a jumper on his motherboard and he also got a RIVA 128 PCI working on his SiS setup: "the LuckyStar(SiS) board does indeed work with the Riva128 - there was a misplaced jumper that caused cache corruption and hence the failed assumption of incompatibility."

I wanted to test a V1 in my first PC but in the end I got to test a mighty NV3 aka RIVA 128 😁 : Diamond nVIDIA Riva 128 - Rev. D - VIPER V330 PCI 4MB

Test configuration.

1. CPU: AMD 5X86 133MHz - AMD-X5-133ADW/Am5X86-P75 3.45V
2. Motherboard: Tomato board / ZIDA 4DPS v2.11- chipset SiS 496/497 - 256KB cache/3 PCI/3 ISA/2 RAM SLOTS
3. RAM: 2x16=32 MB RAM SIMM FPM
4. VGA: Cirrus Logic 5446 - 8260B/V6, 2MB, PCI
5. AUDIO: ESS AudioDrive 1868F ISA
6. HDD: QUANTUM TRAILBLAZER TRB850A -850MB- 4500RPM
7. FDD: NEC
8. CD-ROM: Goldstar/LG CRD-8160B
9. PSU: Minebea Electronics 200W AT PSU
10. Cooling: CPU-Noctua NF-4x10FLX 40 mm x 40 mm x 10 mm 4500 rpm/SSO2 SYSTEM: Scythe Mini Kaze 60 mm x 60 mm x 20 mm 2500 rpm/sleeve.
11. CASE: Generic AT Case manufactured 1998.

RIVA-_TST-001.jpgRIVA-_TST-002.jpgRIVA-_TST-003.jpgRIVA-_TST-004.jpgRIVA-_TST-005.jpg

I prepared a driver package version 1.30 I found on the internet and two CDs taken from the vogons driver library called DIAMOND SUPER CD 98 1,2.

http://www.helpjet.net/Fs-41580795-4007 ... 35073.html
http://www.vogonsdrivers.com/getfile.php?fileid=548
http://www.vogonsdrivers.com/getfile.php?fileid=156

RIVA-_TST-006.jpgRIVA-_TST-007.jpgRIVA-_TST-008.jpgRIVA-_TST-009.jpgRIVA-_TST-010.jpgRIVA-_TST-011.jpgRIVA-_TST-012.jpgRIVA-_TST-013.jpgRIVA-_TST-014.jpgRIVA-_TST-015.jpg

The Diamond SUPER CD is awesome as it contains lots of drivers for Diamond graphic cards. First I tried the V 1.30 package and it installed right away so I didnt use the DIAMOND Super CDs. I had to burn a CD with only the drivers as my floppy disks were a pain to write with my two USB FDD units...errors after errors...this brings back sweet memories....😁

I encountered no image problems or any kind of issues after I used the DIAMOND VIPER V330. All was stable. The colours and image clarity were great.

I was expecting a greater speed in SpeedSys 4.78 for the Viper v330. The CL 5446 2MB PCI got a VESA memory speed of 19376KB/s and the Viper V330 22289KB/s. I think the CPU doesnt let the RIVA 128 spread its wings.

b-3.jpgRIVA-_TST-016.jpgRIVA-_TST-017.jpgRIVA-_TST-018.jpg

I tried a few games: Shaddow Warrior, DOOM, DOOM II, DUKE NUKEM, DESCENT, HERETIC and they ran great. When the RIVA was in my system the Settlers 2 and Heroes 2 wouldnt start saying that the mouse was not present on the COM 1 port or mouse not detected even if I had a PS/2 mouse. However Heroes 2 ran great with the CL 5446 2MB PCI and the PS/2 mouse. I didnt want to try a serial mouse but I'm sure if I would've tried the games would work. The Lost Vikings Game which is a pure WIN95 game wouldnt start with the RIVA 128 saying that it needed a file/gave a Windows error but the game worked from the get go with the CL 5446 2MB PCI.

These are small issues which can be solved but I didnt have the time to do it.

RIVA-_TST-019.jpgRIVA-_TST-020.jpgRIVA-_TST-021.jpgRIVA-_TST-022.jpgRIVA-_TST-023.jpgRIVA-_TST-024.jpgRIVA-_TST-025.jpgRIVA-_TST-026.jpgRIVA-_TST-027.jpgRIVA-_TST-028.jpgRIVA-_TST-029.jpgRIVA-_TST-030.jpgRIVA-_TST-031.jpgRIVA-_TST-032.jpgRIVA-_TST-043.jpgRIVA-_TST-044.jpg

I took Phils test suite and I ran a few tests.

https://www.philscomputerlab.com/dos-be ... -pack.html

The RIVA 128 ran Quake at 640x480(see the RIVA-TST-034.jpg) and 320x200. I ran the tests from WIN95 and I got a 4.6 score for 640x480 and 10.9 score for 320x200. After this I restared in MS DOS mode from Win95 and I got a score of 4.8 for 640x480 and 11.3 for 320x200.

I also tried to run the Quake 640x480, 320x200 tests on the CL 5446 2MB PCI and only the 320x200 test worked. I got a 11.3 score.

So I'm sure my CPU is not feeding the RIVA 128 all the data it needs to show its strength.

RIVA-_TST-033.jpgRIVA-_TST-034.jpgRIVA-_TST-035.jpgRIVA-_TST-036.jpgRIVA-_TST-037.jpgRIVA-_TST-038.jpgRIVA-_TST-045.jpg

With the RIVA 128 I was able to play DN at 800x600 and it worked even if the frame rate was low.

I also tested the Viper and CL cards with 3D Bench 1.0 and 3D Bench 1.0c. I got the same results for the Viper and CL cards. If I had another test system with a Pentium CPU I might've seen other scores. Who knows...

The RIVA 128 ran the PC Player Benchmark in 640x480 and 320x200. The CL 5446 ran only the 320x200. For both the RIVA 128 and CL 5446 the PC Player Benchmark 320x200 test returned the same score 14.6.

RIVA-_TST-039.jpgRIVA-_TST-040.jpgRIVA-_TST-041.jpgRIVA-_TST-042.jpg

All the tests were made after I restarted in MS DOS from WIN95. I dont have a pure DOS machine yet. Also I didnt tweak the settings in BIOS. The defaults were loaded.

I dont know if there are DOS drivers for the RIVA 128. I think the card shows its strength in Windows environment.

So there you have it: 45 pics 😁 No witchcraft or something else. THE RIVA 128 WORKS ON SiS 496/497 😁 kind of...😁

gallery: https://postimg.cc/gallery/2k7j1d2pi/

Last edited by Robert B on 2018-04-19, 19:50. Edited 1 time in total.

Reply 245 of 844, by TheMobRules

User metadata
Rank Oldbie
Rank
Oldbie
Robert B wrote:

The following day, when I tought I had everything ready for the real test, something was bugging me. The Revision 2.0 of the motherboard doesnt have settings for 3V, and my AMD DX4-100 MHz IS 3V! Even if I had jumpers settings for this particular CPU, I decided not to pump 3.45V in a 3V rated part, even if I was told I could do so. The 5x86 133MHz is rated at 3.45V but the motherboard wont recognize it...

The Am486DX4-100 actually requires 3.3V with a tolerance of +/- 0.3V (see here and here). 3V is the minimum, and it will run just fine with the 3.45V provided by your board. That CPU should give you a nice boost compared to the 486-25 😎

Reply 246 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

Light speed compared to the SX 😁 - I know about the +-0.3V but because it is my only DX4 and I didnt have all the BEST parts for this test session I decided to play it really safe 😀 I'll need more RAM, better CACHE chips, a better video card and perhaps other misc items 😁 You can never have too much HW! 😁

I'm sure that at some point in time I'll post a sequel to this episode 😁 This was kind of a semiglorious victory 😁

The next episode will feature something really nice 😀 - I think on Friday I'll receive the latest components. OH BOY!!!

Working title for the next episode: How good is your DIVISION? from a mathematical point of view...

Reply 248 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

Next week...I must wait a little longer...may the package arrive safely 😁

Attachments

  • P5-001.jpg
    Filename
    P5-001.jpg
    File size
    20.65 KiB
    Views
    1465 views
    File license
    Fair use/fair dealing exception
  • P5.jpg
    Filename
    P5.jpg
    File size
    100.52 KiB
    Views
    1499 views
    File license
    Fair use/fair dealing exception

Reply 249 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

BOOM!!! 😁

The parts I've been waiting a whole week have arrived. At a first glance, the CMOS battery is not the right one as it should be a DS1287 and the BIOS chip doesnt have the AWARD sticker. I hope it is the original BIOS chip, otherwise I'll have to buy a BIOS programmer and a new BIOS chip (I've beed planning this for some time). Also I'll have to find a Dallas DS1287 battery and do a CR2032 mod.

Except the facts mentioned above, the Pentium I socket 4 platform is complete:

ASUS PCI/I-P5MP3 Rev. 2.4
CPU Intel Pentium 66MHz - SX837 - A80501-66 - L4102613 - 94025376AA MALAY

By the looks of it, this CPU has FDIV written all over its face 😁

I took off the black stickers from the Intel chips of the motherboard.

The retaining clips of the heatsink are a little bit strange and I wonder if this heatsink is the original one. On the back of the heatsink, I found two pieces of a small round sticker and on the CPU, I could see a faint round mark, sign that maybe the CPU and the heatsink worked together sometime in the the past. In the future I'll have to make a retaining clip from a piece of steel wire or modify a retaining clip from another heatsink.

I also received two ISA modems, an ISA controller and a S3 Savage 3D AGP card. The state of these cards is rough and ATM I dont know which is going to be restored.

The Pentium I kit was bought AS IS. I dont know if it works or not. The seller told me that the last time it worked it was somewhere in 1999...

PCI-_P5_MP3-001.jpg PCI-_P5_MP3-002.jpg PCI-_P5_MP3-003.jpg PCI-_P5_MP3-004.jpg PCI-_P5_MP3-005.jpg PCI-_P5_MP3-006.jpg PCI-_P5_MP3-007.jpg PCI-_P5_MP3-008.jpg Z-_MISC.jpg

gallery: https://postimg.cc/gallery/nasleyb6/

More later.

Last edited by Robert B on 2018-04-19, 19:51. Edited 1 time in total.

Reply 251 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

I'm working on it as we speak.

Tomorrow I'll finish the battery mod. The board should've come with a DALLAS DS1287 but it has a BENCHMARK BQ3287MT. I found on the internet that they are interchangeble so I keep my fingers crossed 😁 The mod is easy to do and as usual I'll post the TL;DR version. I separated both the + and - leads for the battery. I never leave my work half done 😁

More later.

Reply 252 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

TONIGHT!

A proper RTC battery MOD for Benchmark BQ3287MT / Dallas DS1287 / ODIN OEC12C887A 😁

Attachments

  • BQ3287MT.jpg
    Filename
    BQ3287MT.jpg
    File size
    744.62 KiB
    Views
    1334 views
    File license
    Fair use/fair dealing exception

Reply 253 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

Hey Grandpa, do you still lose track of time?!?!?

Have no fear, I have the solution right here, said "the greenhorn" 😁 yep, true story 😁

As you know well, recently, I have aquired a socket 4 motherboard and CPU, an illusive creature which only comes out at night 😁

The Pentium 66 MHz is a real treat, aka THE ORIGINAL PENTIUM!!! To put things in perspective, we are talking about THE PENTIUM 66 MHz, top dog back in its heyday and not a Pentium 60 MHz aka a Pentium 66 MHz which didnt make the cut. We all know its story, FDIV and stuff...

Before I received the socket 4 bundle I knew I will have to do an RTC: Real-Time Clock module mod. After decades, it was obvious that the lithium cell/battery inside the RTC module went the way of the DoDo bird. Do you still believe in fairy tales?... I know I still do.. kind off...but I wasnt expecting that the battery would still be good...

When I received the package with the Pentium 66, I saw that I didnt get the right RTC module. Instead of a Dallas DS1287 RTC module I received a Benchmarq BQ3287MT module. I used the Benchmarq RTC module but I wasnt able to make the socket 4 system to post or boot. Regardless, this story is not about me not being able to solve the P66 puzzle, which is just a temporary setback, as I'm going to present to you an RTC module mod for vintage/old systems.

The RTC module is kind of special in the sense that it's more than just a CMOS battery: "The CMOS BQ3287/BQ3287A is a low-power microprocessor peripheral providing a time-of-day clock and 100-year calendar with alarm features and battery operation. Other features include three maskable interrupt sources, squarewave output, and 114 bytes of general nonvolatile storage."

http://pdf.datasheetcatalog.com/datasheets/90/122373_DS.pdf

Without this humble black toffee, an older PC is dead in the water not beeing able to boot, get past the boot screen or load an OS. Quirks of ancient technology.

You can find on the internet plenty of RTC CR2032/RTC chip/module, mod/hack/rework, Dallas DS1287 / Benchmarq BQ3287 / ODIN OEC12C887 tutorials. The gist of these mods/hacks is represented by the attachment of a regular CR2032 socket+battery, to a couple of specific terminals of the RTC module after they are disconnected from the dead lithium cell/battery inside.

http://www.mcamafia.de/mcapage0/dsrework.htm

After several years, these lithium cells die and can't hold the required charge to keep the BIOS settings which are lost after you disconnect the PC from the mains, power off the system or you just restart the PC.In some instances this fact can render an older PC useless.

Said and done. I read again the instructions and I was ready to ROCK and ROLL!

To make things more challenging, fate did so, that a pin from the Benchmarq BQ3287 RTC module broke off when I pulled it out of the socket... call it a bad hair day 😁 BEHOLD it is the PIN 13!!! FTW!!! 😁 Wicked!!! IN YO' FACE!!!

PIN_13.jpgBQMOD-001.jpg

I took my trusty callipers and I took a measurment to determine the exact spot where I was about to use a hacksaw to expose the cell battery contacts. PIN 16: (-) and PIN 20 (+). I hacked away until I felt the metal.

BQMOD-002.jpgBQMOD-003.jpgBQMOD-004.jpgBQMOD-005.jpg

I took a multimeter and I measured the voltage of the cell battery inside the RTC module: 2.96V what the f..k??? At first, I thought I was about to open up a good RTC module but quickly I figured out that the charge was from the test of the socket 4 system I did earlier and I was sure that the charge wouldn't've been kept for long, minutes, hours, a few days.... My hunch was confirmed after I disconnected the battery when I made a new measurement and I got a BIG FAT ZERO VOLTS.

BQMOD-006.jpg

After I made the initial cuts with the hacksaw, I used a few fine files of various shapes to fully expose the contacts between the RTC module and the battery. As you can see in the pictures, the pin 16 and pin 20, have been bent the other way to make contact with the internal lithium cell battery.

BQMOD-007.jpgBQMOD-008.jpgBQMOD-009.jpg

I used a fine screwdriver and with a firm rocking motion, left-right, I severed the contact of the pin 16 and pin 20 with the lithium cell battery. I took care to remove all the metal bits that resulted from this operation.

BQMOD-010.jpgBQMOD-011.jpgBQMOD-012.jpg

I prepared the CR2032 socket that I wanted to attach to the BQ3287MT RTC module.

BQMOD-013.jpg

To make things bulletproof, I decided to also cut the tips of the pins I exposed. Better safe than sorry.

BQMOD-014.jpg

I took a razor and I cut four little bits of plastic from the CR2032 socket as they would've hindered me when I was to attach it with super glue to the RTC module.

BQMOD-015.jpg

I prepared the wires that I was about to use.

BQMOD-016.jpg

I cleaned everthing with and old toothbrush and 99% isopropyl alcohol.

BQMOD-017.jpg

Looking good!

BQMOD-018.jpgBQMOD-019.jpg

I used black electrical tape to mask the exposed contacts of the dead cell battery. Looks are VERY IMPORTANT!

BQMOD-020.jpg

After all of this, I tackled the problem of the missing pin. I tried to solder the sheared tip of the pin 13 but I wasnt successful, so I took the hacksaw and the fine files and I did a third groove in the RTC module. A CACHE CHIP donated a pin and the problem was solved. I used a french key to keep the things in place while I did these delicate procedures.

BQMOD-021.jpgBQMOD-022.jpgBQMOD-023.jpgBQMOD-024.jpgBQMOD-025.jpgBQMOD-026.jpg

I soldered the wires to the CR2032 socket/battery holder.

BQMOD-027.jpg

I made a set of CUSTOM heat shrink tubes for the CR2032 socket, just for looks.

BQMOD-028.jpg

Almost done.

BQMOD-029.jpg

I determined the ideal length of the wires.

BQMOD-030.jpg

I used super glue to attach the CR2032 socket on the Benchmarq BQ3287MT RTC module. To obtain a strong bond I used again the french key. A pen was used as a spacer to protect the pins of the BQ3287MT RTC module.

BQMOD-031.jpg

I soldered the wires from the CR2032 socket to the BQ3287MT RTC module.

BQMOD-032.jpgBQMOD-033.jpgBQMOD-034.jpg

Looking SEXY!!! I used TRANSPARENT POXIPOL to cover the soldered joints of the RTC module.

BQMOD-035.jpgBQMOD-036.jpgBQMOD-037.jpgBQMOD-038.jpgBQMOD-038-01.jpg

HOT! HOT! HOT!

BQMOD-039.jpg

gallery: https://postimg.cc/gallery/241ax5y0i/

The entire operation took more than three hours.I cant establish the time with precision as it went by in a flash. Now, I can make this mod in less than an hour and all this experience was worth it even if I wasnt able to wake up the ASUS socket 4 motherboard. This will have to wait a little more...

I have to get a Dallas DS1287 RTC module, hack it, and then try to power up again the Pentium 66...

DS1287.jpg

More later.

FTW!!! FYW!!! FTW!!!

Moonwalk - Girl For You

https://www.youtube.com/watch?v=xFc1nrwXUKg

Last edited by Robert B on 2018-04-19, 19:51. Edited 1 time in total.

Reply 254 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

The JPN CORPORATION CA8302E-1 controller for the socket 4 P66 system is ready to rumble in the jungle.

I had to change a 10uf 50V capacitor and to straighten all the pins on the back.

I dont know if I will be able to find a DS1287 RTC until the end of the year to establish the state of the P66...we'll see...

CTRLR.jpg

More later.

Last edited by Robert B on 2018-04-19, 19:52. Edited 1 time in total.

Reply 255 of 844, by DeafPK

User metadata
Rank Newbie
Rank
Newbie

Quite the thrill reading through this page! Both the VLB and socket 4 stuff takes me back to fiddleing with computers my primary school threw out in 2000-2001. I started my computer hobby trying to make the most out of those outdated PCs. Nice to see the 501 @ 66MHz , never had one of those.

"an occasional fart in their general direction would provide more than enough cooling" —PCBONEZ

Reply 257 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

I did power up the P66 and I saw that the CPU gets warm and the motherboard chips also get warm sign that the motherboard is alive. Also the chips of the graphic cards I tried got warm to the touch sign that they receive power.I checked the BIOS chip and it seems legit.

With the BQ3287MT, with and without the battery mod, I didnt get any sound from the PC Speaker. I tried ISA and PCI graphic cards. I removed the RAM. I fiddled with the jumpers. I removed the CACHE chips.I disabled the CACHE from jumpers. I set a jumper on TURBO.I changed the PSU. I checked the clock crystals specs... NOTHING HAPPENED!!! Not even a beep on the PC Speaker.

At this point my best bet is the RTC so I'll get a DS1287, try it, with and without the CR2032 mod to see if the P66 works.

If this doesnt work, I'll really really really have to buy a BIOS programmer to see what's on the BIOS chip. I find it curious that it doesnt have the AWARD sticker on it.Maybe this motherboard was a weird revision. It came with black stickers on the motherboard chips and I've removed them. Maybe because it wasnt powered in more than 20 years it lost it's data...I've never fiddled with old parts as this one 😁 It's something new to me 😀

The BIOS chip is a CSI CAT28F010P-15

"The CAT28F010 is a high speed 128K x 8-bit electrically erasable and reprogrammable Flash memory ideally suited for applications requiring in-system or after-sale code updates."

"The CAT28F010 is manufactured using Catalyst’s advanced CMOS floating gate technology. It is designed to endure 100,000 program/erase cycles and has a data retention of 10 years."

http://www.datasheetcatalog.com/datasheets_pd … CAT28F010.shtml

I had to use isopropyl alcohol 99% with a cotton stick and a strong light to see the specifications. This was tricky as I had to take pictures, keep the alcohol wet, close to a bright hot source of light that made the alcohol evaporate quickly. I read on vogons that someone with a 486 board had such a chip and this is why I think/hope is legit.

BIOS repair

BIOS_CHIP.jpg

The DS1287 will clear this up, one way or another...I dont know if I'll be able to find one by the end of the year, but I have a source for one.

More later.

Last edited by Robert B on 2018-04-20, 05:49. Edited 2 times in total.

Reply 259 of 844, by Robert B

User metadata
Rank Oldbie
Rank
Oldbie

Thanks amadeus777999 - I keep my fingers crossed! 😁

Sneak peek! 😁 Socket 4 preview 😁

I still have work to do but I'm confident I'll be able to wake up THE BEAST!!! I've tried poking it with a stick but I got no reaction 😁 I guess for now it is content to play dead and tries to ignore me! 😁

Sock4-_SNKPK-001.jpg Sock4-_SNKPK-002.jpg Sock4-_SNKPK-003.jpg Sock4-_SNKPK-004.jpg Sock4-_SNKPK-005.jpg Sock4-_SNKPK-006.jpg

gallery: https://postimg.cc/gallery/i4ifmjhy/

Stromae - Je cours

https://www.youtube.com/watch?v=wIrI_MOfxG0

More later.

Last edited by Robert B on 2018-04-19, 19:55. Edited 1 time in total.