I "fixed" my ThinkPad A31p and I'm very happy about it. Had the problem that it was drawing too much power und tripping the power supply (which was running very hot) after a few minutes.

So I got a thermal camera to debug this, disassembled the whole thing, found nothing, reassembled it...and the problem went away.

What the...

The only explanation I have for this is that I might have accidentally produced a short when I put it together the last time and did it properly this time.

Completed another build. I present to you the "average" '98-'99 PC.

Scope of this build is to recreate the very first PC I used which was an Olidata preassembled. Sadly it got dismembered and most components thrown away over the years.
I can't find the case online so I had to buy a cheap new one in the meantime.

MB: Olidata OLI-BX3 (a rebranded Asus P3B-F)
CPU: PII@350
RAM: 256 MB PC100 (but I will probably lower it to 128)
GPU: Matrox G200
SOUND: Creative AudioPCI
STORAGE: LG DVD-ROM and 60GB SSD with ide-sata adapter.

Both GPU and sound card are original from my first PC, still with the Olidata sticker.

Well, I got a bit further with my PT-432B. Today my shiny new XGecu T48 arrived and I was able to look into the BIOS chip. Turns out it was completely blank, as in FF all around. Wrote the only available BIOS from RetroWeb on it and it verified just fine. So I have a board that tries to post now:

Sadly it gets stuck and cycles between code 62 and 63. Re-seating Cache and swapping memory didn't change anything. It's hard to find a reference for those codes, but some references suggest that the test of DMA controller #2 fails, an exception is raised and it's retested in a loop.
One possible explanation is that this was the original problem of the board that drove the seller to try and update the BIOS, fudge it and then leave it alone. Or the one BIOS file I found is incompatible. So, if anyone owns that board and has a BIOS dump I'd appreciate it very much!
Either way, I'll try and find some more BIOS files to try and then try and narrow down the problem on a board level. If one part of the UMC chipset is dead I'll try and source one of those.

Did some more work on my intended six soundcard 486 build-in-progress. Swapped a PSU from another case for one with more stable voltages.

Also got a Pentium OverDrive to try out in this system. Unfortunately, it yielded the same outcome as when I tried the DX4-100 OverDrive processor: parity error on boot-up.

The parity error was triggering as a result of HIMEM.SYS. I then tried several different sets of RAM, before finally finding a set that worked. But I was still faced with the issue of the L2 cache not working.

Ended up replacing the BIOS with a MR BIOS v1.65 for this particular chipset (per this post) and both the original RAM and L2 cache finally worked.

This motherboard is clearly not "OverDrive Ready". 😠

KungfuPancake wrote on 2023-09-07, 11:59:

Either way, I'll try and find some more BIOS files to try

Well that came to a speedy conclusion - tried some BIOS files from OEM variants (SR-M401-A) of this board and found one that POSTs. Now it's time to test if everything works 😀

KungfuPancake wrote on 2023-09-07, 12:29:

KungfuPancake wrote on 2023-09-07, 11:59:

Either way, I'll try and find some more BIOS files to try

Well that came to a speedy conclusion - tried some BIOS files from OEM variants (SR-M401-A) of this board and found one that POSTs. Now it's time to test if everything works 😀

IMG_0435.jpg

Post screen looks amazing in such a big screen xD

I did replace Novell Netware for mars_nwe for my retro Windows install
And reverted to etherdfs for DOS/W3.1

I needed a shared drive for all my retro PC / OS to be able to access and the ability to upload files on the server from my recent hardware.

Samba was a no go (don't support smb1 anymore afaik)
ftp is annoying to use
NFS have issue when using client and server from different epoch all in the same time
etherdfs work great, for DOS client

So I did initially revert to installing a Netware 4.11 VM, with did work great for DOS, W3.11 and W95 clients, but out of the box I can't connect from recent Windows and Linux. And I was running it on a KVM VM. The CPU is always running at 100% in that kind of configuration, so it generate noise. So I decided to install the TCP and IP support then install FTP server on that netware server so I can dump file from my recent PC to the network share. Reliability go down like a rock when I was using FTP...

Ended up trying 6.5 SP8, mostly because it come with some kind of unlimited license, native support for TCP/IP and it also come with a FTP server by default. But then while installing and using it I noticed it ate 12 poutines a day between the 4.11 and 6.5 version. So bloated and slow... Take > 5 minutes to boot easy on a 4 cores VM (i5 8400t)... For comparison, the 4.11 take like 5 seconds to boot... And the 100% CPU usage and the Instability when using FTP are still present.. only, instead of just getting big lag for the many clients, it freeze the server (and the clients 🤣).

So I managed to compile IPX linux module (using DKMS so it recompile it automatically on kernel upgrade), And mars_nwe. Seam to work great so far, So I can continue to use etherdfs on DOS and W3.1, but windows 95 can connect to the share with the same exact netware client I was using for the real thing. And I can just dump file on the server, from recent machines, using samba.

Well - this is what I'm going to be playing with tonight - a NETGEAR WNCE2001 Universal WiFi Adapter for my vintage Desktop machines. I've got the new scheme figured out by the 4 seasons using 3 machines: 1985 Tandy 1000A (currently setup), 1991 Compaq DeskPro 386s/20, and my 1995 CreepingNet 486 DX4-100 box

SPRING = 486/COMPAQ
SUMMER = COMPAQ/TANDY
FALL = TANDY/486
WINTER = COMPAQ/486

Basically, I'm getting away from having to have all the bloody wires everywhere. Then, once I live somewhere where I can have a dedicated computer-den with all three permanently setup, I can put the WiFi Adapter on a switch and use it to bridge to my WiFi network. Not like it's a big bandwidth hog since all three are 10mbps.

KungfuPancake wrote on 2023-09-07, 11:59:

Well, I got a bit further with my PT-432B. Today my shiny new XGecu T48 arrived and I was able to look into the BIOS chip. Turns out it was completely blank, as in FF all around. Wrote the only available BIOS from RetroWeb on it and it verified just fine. So I have a board that tries to post now:

I see two bioses available for 432B https://theretroweb.com/motherboards/s/pine-t … -432b#downloads

KungfuPancake wrote on 2023-09-07, 11:59:

Sadly it gets stuck and cycles between code 62 and 63. Re-seating Cache and swapping memory didn't change anything. It's hard to find a reference for those codes, but some references suggest that the test of DMA controller #2 fails, an exception is raised and it's retested in a loop.

DMA controller is located in UM8886 chip, pinout and pin descriptions: Re: UMC8881/8886 Datasheet

KungfuPancake wrote on 2023-09-07, 12:29:

Well that came to a speedy conclusion - tried some BIOS files from OEM variants (SR-M401-A) of this board and found one that POSTs.

https://theretroweb.com/motherboards/s/sowah- … 401-a#downloads
PT-432B 203 is identical to SR-M401-A 203 suggesting those boards should use same bios
did 201 work for you?

gmaverick2k wrote on 2023-09-01, 21:27:

Installed Windows 7 on iMac 2011 27". Tried to get to windows 10. No cigar... Windows 10 is way better than windows 7 and protected... It wanted a product key. This iMac has a gorgeous lg 2k display with Inky Black's and nice internal speakers, really bright too. Beats my benq ex3210r in picture quality and it's great for media consumption. Best £100 spent on tech. Tried YouTube in win 7, don't know what magic Apple do but the sound is less full and the colours aren't saturated with iMac profile.

Found dvd copy of win 10 pro with OEM sticker from pc I had bought a while back. Tried it and it activated on iMac 2011. Bootcamp working fine, donated it to family. So impressed with Apple quality, bought 500 nit iMac 2017 5k 27". Gorgeous display. Has Radeon pro 580, 7600k. If only it allowed for alt codes in Excel, I'd jump ship fully...

rasz_pl wrote on 2023-09-07, 18:09:

I see two bioses available for 432B https://theretroweb.com/motherboards/s/pine-t … -432b#downloads […]
Show full quote

I see two bioses available for 432B https://theretroweb.com/motherboards/s/pine-t … -432b#downloads

https://theretroweb.com/motherboards/s/sowah- … 401-a#downloads
PT-432B 203 is identical to SR-M401-A 203 suggesting those boards should use same bios
did 201 work for you?

So there's actually only one BIOS file that can be downloaded for the 432B, the other entry is for information purposes I guess. I tried both for the SR-M401-A though, and the 201 version is the one that actually worked. Maybe there was a cutoff point for this particular board, possibly linked to a board revision.
I do have some weird behavior though which I feel might be connected to the RTC: The time can't be set and doesn't advance, and all forms of resetting the Board (CTRL-ALT-DEL, Reset Button, Software Reset, BIOS save/discard reset) end in a hung system. It's almost as some of the registers of the RTC do not work and/or a connected system watchdog is malfunctioning. The RTC is a rebadged Dallas chip which can be sourced very cheap so I'm going to get one just for safety.
I guess I'll have to break out the scope and properly debug this though..

Edit: just started probing and saw that the 32khz crystal is not doing its thing. Figures. I'm gonna rummage through my crystal collection tomorrow and see what I can find 😄

Well today I think I've broken an Abit VH6T while swapping the cpu... Because the large cooler mauled two nearby caps and I noticed them only after failed post. Now even with original cpu it does nothing.
Tomorrow will arrive the new caps. Fingers crossed.

I continued with my Socket A build after I got home from work. Last activity for the evening was to install my old water block to Geforce FX 5900XT:

It also crossed my mind, that it is pretty close 21 years when I've installed this block last time to a GPU. Back then it was first in my GF3ti200, but it was roughly around this time 21 years ago I installed it to Radeon 9700 I got. Next GPU I bought was X800XT PE for my Athlon64 system. This block wasn't installed to that particular card at all, I removed it from the 9700, installed original cooler back and sold the card. Since then this block has been in my parts bin. Until now, that is.

A few days ago I received a K7 PRO and an Asus V6800DDR, both of which were in a questionable condition and presented no signs of life.

Both were bathed as well as scrubbed well and today the K7 partially recapped, leaving only the original Sanyos and smaller caps in place.

Against all expectations the board and 600Mhz Athlon + GORB booted straight away, as did the GF 256. However 256 gives a garbled screen after 1-x reboots, while 1st boot is just fine - what is this usually a sign of? Memory issue or bad caps?
E: and now that I plugged the thing into my actual modern ATX power supply instead of a test power supply, looks like the card works without issues. Weird, maybe I immediately need to throw the test unit off a balcony?

Got in the WiFi adapter last night, and spent all evening on the Tandy 1000A surfing Bulletin Boards and messing with IRC, and upgraded mTCP to the latest version. It works great. So I think that's my next incarnation of retro-computing networking setup is this whole wifi bridge setup.

I have a decent Nokia 445ZA monitor here that I rescued from a recycling container years ago. It didn't work and needed some effort to get back to operation but it got done and since then it has seen 15k hours of use (with 17k on it before I got it).
There has always been the problem that convergence is a bit shit and in one place and no matter what I couldn't get it right... until today, when I did the nuclear option of taking the yoke off, cleaning everything (it had accumulated plenty of soot from the fireplace in the old place I used to live) and starting over, and now I have near perfect result. I still have to put some magnet strips under the yoke in startegic places but I have a result that factory conf couldn't get with its 6 magnets 🤣. I did some improvements in video path too and it is actually really decent now, far better than the stock state... now if the kinescope itself could have its working hours turned back 🤣

I have to say, that is the most impressive CRT repair I've seen. You did a wonderful job and I sure need to see how many hours my Orion has along with my PVM-9L1. I have a 15" KFC that has very low hours on it and man, does she look sharp. The shell needs a retrobrite, but it adds character.

I also repaired my DGi DGi-P-90C laptop. All it needed was a new clock battery and it powers on without issues. The main batteries need a rebuild, the display inverter boards (on both FMA7600 laptops) need a recap due to a coil whine if its brightness is set to 100%, the displays need to be replaced later on (I have OCD when it comes to displays, but they're TFT, which is a must, yet dead pixels are on those displays), the one needs its hinge mounts fixed in since they broke out, but are in 1 piece with either epoxy or putty, the other needs a drive caddy and drive (I have one that was in my Austin laptop, but I might do solid state so I can copy files to the SD card easily), and it also needs a charger along with a clock battery replacement (doing it externally like the semi-working unit).

This is my crappy Cherry membrane PS2 test keyboard wich i dont care for if it gets dirty.
Everything has yellowed except of one row of keys?! Who knows why...
This weekend we had pretty hot and sunny days so i decided to try my luck in sun brighting.
I just put it out of the window for 1 and a half days in full sun. And when i was done i cleaned it.

And i really think it looks better. Maybe 1 or 2 more days in the sun and its even better.
You can still see the sides of the keys are a little bit more yellow because i didnt rotate it during the day, but overall it looks better.... i think...

DerBaum wrote on 2023-09-10, 18:02:

This weekend we had pretty hot and sunny days so i decided to try my luck in sun brighting....

I was amazed that it really works (at least on the keyboard) ... so i decided to test a little bit more...
I took the yellowest stuff i can find and put it in the sun... Lets see how it will turn out...
This time i even have a "control piece". I use 2 identical yellowed drive bay covers... one goes into the sun, the other one not.

Hmm these Nichicon cap(s) gone bad, my Tulip TCX17 power supply was completely dead. Better heat up the iron then and hope for the best.