kingcake wrote on 2024-02-27, 01:02:

I just got one of those tiny grinder pens from ali express after seeing Northride Fix use one. What a game changer compared to scraping with tweezers. And now I don't keep ruining the tips of my tweezers 🤣

I just found this product as well and immediately ordered one from AliExpress. This is *exactly* what I need. I have a Voodoo 2 in queue for repair and have been carefully considering how to dig out the broken legs from inside the chip. I have done this previously with an X-Acto blade with middling results. I was considering getting a dremel, but wasn't sure if it would be fine enough for the job. This grinder pen is perfect-- and cheap!

I'd probably get the drill stand set up with the ball shaped grinder and take the first 80% of that off though since it's a relative lot of hard material to go through. On mine though I could rig it so it hits the stop at maximum depth so couldn't go too far in no matter how hamfisted I am.

Kahenraz wrote on 2024-02-27, 09:27:

kingcake wrote on 2024-02-27, 01:02:

I just got one of those tiny grinder pens from ali express after seeing Northride Fix use one. What a game changer compared to scraping with tweezers. And now I don't keep ruining the tips of my tweezers 🤣

I just found this product as well and immediately ordered one from AliExpress. This is *exactly* what I need. I have a Voodoo 2 in queue for repair and have been carefully considering how to dig out the broken legs from inside the chip. I have done this previously with an X-Acto blade with middling results. I was considering getting a dremel, but wasn't sure if it would be fine enough for the job. This grinder pen is perfect-- and cheap!

IMG_0558.jpeg

Hey that looks familiar! 🤓

Anyway, the PSU in the Compaq Deskpro XE 450 I recently acquired is kaput. ☹️ I’ve cracked it open and poked around with my multimeter. Nothing appears shorted or blown up. There’s a Rifa in there, but it looks in one piece, no cracks. I can see 160V going in to the big step down transformer and then about 0.5V coming out. I put a load (duff HD) on the 12V rail and that didn’t help.

Out of my depth on this one - gonna pop the cover back on and set it aside, for my own safety. At least it’s a standard form factor AT PSU, so I can easily replace it. (Either that or dump the guts and put a pico ATX in the box, but that’s def option B)

ubiq wrote on 2024-02-27, 20:17:

Anyway, the PSU in the Compaq Deskpro XE 450 I recently acquired is kaput. ☹️ I’ve cracked it open and poked around with my multimeter. Nothing appears shorted or blown up. There’s a Rifa in there, but it looks in one piece, no cracks. I can see 160V going in to the big step down transformer and then about 0.5V coming out. I put a load (duff HD) on the 12V rail and that didn’t help.

Start a new thread and post some pictures of the insides of the PSU there. If you have a multimeter and are willing to poke with it a little, that PSU shouldn't be too hard to revive. Most AT PSUs are either half-bridge or STF or flyback topology and fairly trivial. The Rifa needs to go regardless of its condition. If it's after the fuse, check the fuse and the NTC thermistor. Since you measure 160V, though, those are probably fine. Also, I imagine this measurement implies there is only one large primary 400V capacitor and that's what you measured the 160V across? If so, you probably have a flyback or STF topology PSU.

In any case, before starting any troubleshooting, do the hair dryer trick first: open the PSU (unplugged from the wall) and heat it up to about 40-50C (components should feel nice and very warm to almost hot to the touch.) Then plug it in and see if it starts. Make sure you put a good load on the 5V rail and not necessarily the 12V rail. AT PSUs expect most of their load on the 5V rail, and typically at least 1 Amp.

momaka wrote on 2024-02-27, 20:48:

ubiq wrote on 2024-02-27, 20:17:

Anyway, the PSU in the Compaq Deskpro XE 450 I recently acquired is kaput. ☹️ I’ve cracked it open and poked around with my multimeter. Nothing appears shorted or blown up. There’s a Rifa in there, but it looks in one piece, no cracks. I can see 160V going in to the big step down transformer and then about 0.5V coming out. I put a load (duff HD) on the 12V rail and that didn’t help.

Start a new thread and post some pictures of the insides of the PSU there.

Good idea!

I'll also try the hair dryer trick, but I don't have an actual hair dryer just a hot air rework station so I'll have to be just a little careful...

Had some fun with a board I hadn't managed to work on quite a good chunk of time since its recap.

Quick specs of the build:

Seasonic SS-350FS PSU (disguised as ANS LC-B400ATX 🤣)
DTK PRM-27i E0
P3 Coppermine 800EB (133FSB)
384MB RAM
ASUS V9180SE / GF4 MX440-8x 64MB
NEC USB 2.0 card
Infineon AN983B LAN card
SB Live 5.1 SB0100 PCI soundcard
IBM Deskstar DPTA-371360 HDD
Optiarc DVDRW
98SE

Tested my new arrival, a S.462 ECS KT600-A.
Hiren's MiniXP ok. Ubuntu 4 and 7 ok. On 7.04 I tested sound with Rhytmbox Music Player, Radio. 2 out of the 4 work... From 2007 😀 HBR1.com

Edit: Sata drives won't work unless capped @150, wouldn't even get detected; not in the manual, but seen elsewhere.
Nice little board for 10€

I finished a new build a P4 3.0 - 2gb 667 and Gf 6600GT, it's the clone of a pc i had back in 2004, i wanted to test a game i played a lot back in the day, vampire bloodlines.
Gotta be honest i was expecting better performances, it works well mind you but in large areas framerate drops to 25, it doesn't matter if i drop the resolution or all the details to low, the framerate stays the same, which makes me believe that the pixel\vertex pipelines are just not enough for this hot mess of unoptimization, i will try an X800 GTO on this same system to see if a card from the same generation but with more pipes can deal with the game any better.

RandomStranger wrote on 2024-02-25, 07:24:

Made a clean install of Windows 95 on the Olivetti M4-100 and noticed something that was right before my eyes all time. It has no L2 cache. So that's why it felt so uncharacteristically sluggish. Even more surprised it stayed in use as an office PC until the mid-2000s.

Edit: And another reinstall after it started blue screening when loading up the desktop. Sometimes I forget how finicky W95 can get.

After ironing out all the issues with the Olivetti, sold it on a local trading site. The buyer said this model was his old PC back in the 90s. Packed as well as I could. Hope the post won't find a way to still destroy it.

iraito wrote on 2024-02-28, 00:06:

I finished a new build a P4 3.0 - 2gb 667 and Gf 6600GT, it's the clone of a pc i had back in 2004, i wanted to test a game i played a lot back in the day, vampire bloodlines.
Gotta be honest i was expecting better performances, it works well mind you but in large areas framerate drops to 25, it doesn't matter if i drop the resolution or all the details to low, the framerate stays the same, which makes me believe that the pixel\vertex pipelines are just not enough for this hot mess of unoptimization, i will try an X800 GTO on this same system to see if a card from the same generation but with more pipes can deal with the game any better.

Is your 6600gt one of the unlockable cards? I remember I had a 6200 that I could unlock the pipelines and it ran stable as basically a 6600.

Sadly the 6600 isn't unlockable, the frame rate isn't bad like 80% of the times, this game is simply unoptimized and runs at like 25 in some scenes.

Recapped my first motherboard (successfully)!

The board is a EPOX EP-BX3.

I also tried one of the voodoo 2s i have in the build and it works, albeit some image distortion/flickering. I will clean the contacts and see if it solves it. I'm happy though as I thought the voodoo was not working at all. Now I can get on with the repairing the rest of the voodoos. 😀

I got a new FM1 motherboard and tested it. I also wanted to test a cpu, a A6-3500, 3 cores.

The board is fine, no issues. The cpu gets to 50°C in a matter of seconds. It was shorting Vcore. One pin is broken in the middle of the cpu (ground, no issues really).
I decided to open it. The smds are my damage.

The pin I'll use to test my replacement skills while desoldering/soldering new.

Prepping a TNT2 M64 to be sold. I'm also reminding myself of the old almost-period-correct drives that I use for testing 🤣, this time it's a Fujitsu MPB3021AT-E that's chuggin' near my ears, and the test build is a 500MHz Katmai on a Luckytech P6BX2 mobo, which has had a SMD MOSFET mod done to it.

Nexxen wrote on 2024-03-01, 20:17:

I got a new FM1 motherboard and tested it. I also wanted to test a cpu, a A6-3500, 3 cores. […]
Show full quote

I got a new FM1 motherboard and tested it. I also wanted to test a cpu, a A6-3500, 3 cores.

The board is fine, no issues. The cpu gets to 50°C in a matter of seconds. It was shorting Vcore. One pin is broken in the middle of the cpu (ground, no issues really).
I decided to open it. The caps are my damage.

The pin I'll use to test my replacement skills while desoldering/soldering new.

Just to let you know, it's very hard to solder SMD components to a CPU, due to how easily it wicks away heat. A pre-heater or hot air station to warm it up first is highly recommended.

CPU pin repair of an Intel 486DX-50

Repairing damaged capacitors on an Intel Xeon E-2278G

Kahenraz wrote on 2024-03-01, 21:35:

Just to let you know, it's very hard to solder SMD components to a CPU, due to how easily it wicks away heat. A pre-heater or ho […]
Show full quote

Nexxen wrote on 2024-03-01, 20:17:

I got a new FM1 motherboard and tested it. I also wanted to test a cpu, a A6-3500, 3 cores. […]
Show full quote

I got a new FM1 motherboard and tested it. I also wanted to test a cpu, a A6-3500, 3 cores.

The board is fine, no issues. The cpu gets to 50°C in a matter of seconds. It was shorting Vcore. One pin is broken in the middle of the cpu (ground, no issues really).
I decided to open it. The caps are my damage.

The pin I'll use to test my replacement skills while desoldering/soldering new.

Just to let you know, it's very hard to solder SMD components to a CPU, due to how easily it wicks away heat. A pre-heater or hot air station to warm it up first is highly recommended.

CPU pin repair of an Intel 486DX-50

Repairing damaged capacitors on an Intel Xeon E-2278G

Second method is a life saver, I recently tried to put back in place a cap and it was hell.
Next time I'll use solder paste.
Not mentioning that they get scalding hot 😀

I'll try to preheat the cpu, remove the foot that is still on the cpu and then maybe use solder paste to help a bit.
Not sure, I need to see what happens when it is removed.

Thanks for the suggestions.

And a quick photo (sorry that I couldn't do a proper printscreen.) of 3DMark 99 Max for the TNT2 M64. Looks good to me.

Making a batch of my slightly enhanced voltage blasters.

Differences from Phil's version:
Surface mount regulator allows heatsinking to copper planes of PCB
Protection diodes as recommended in the 79xx regulator datasheet
LEDs for the 4 main power rails on the ISA bus
Disable/Enable jumper

I made an ATX power test jig to verify operation. It also lets me switch in load resistors. And of course has test points to monitor output voltage with a meter.

It has a PCI slot because I was experimenting with some PCI power supply stuff.

Kahenraz wrote on 2024-02-27, 09:27:

I just found this product as well and immediately ordered one from AliExpress. This is *exactly* what I need. I have a Voodoo 2 in queue for repair and have been carefully considering how to dig out the broken legs from inside the chip. I have done this previously with an X-Acto blade with middling results. I was considering getting a dremel, but wasn't sure if it would be fine enough for the job. This grinder pen is perfect-- and cheap!

IMG_0558.jpeg

How has that repair been going? Are you going to do a thread on it?

I also wrote a python script to combine option roms for burning into a single eprom. It also automatically fixes boundary alignment, the length byte, and checksums.