Yes, they're heatsinks. I've seen them before.

You might be stuck with that power supply because Gateway had a proprietary form factor (still ATX power connectors, but smaller than normal). I've also heard a lot of complaints about them failing, even though they came from some of the better manufacturers (I've seen Astec, FSP, and Newton Power).

I didn't realize that Gateway had a 120mm fan ducted to the CPU heatsink in those things. It's a pretty high-speed fan by modern standards, too.

lti wrote on 2025-06-30, 04:57:

Yes, they're heatsinks. I've seen them before.

You might be stuck with that power supply because Gateway had a proprietary form factor (still ATX power connectors, but smaller than normal). I've also heard a lot of complaints about them failing, even though they came from some of the better manufacturers (I've seen Astec, FSP, and Newton Power).

I didn't realize that Gateway had a 120mm fan ducted to the CPU heatsink in those things. It's a pretty high-speed fan by modern standards, too.

*groans* aww... When I read that, I thought "WHAT?!?" and went and picked it up and it was immediately obvious it was a weird form-factor... I only removed it and plopped into the "pile" for cleaning. And OH yeah, that fan moves some air, BUT it was going the wrong direction...

I got most of the Gateway cleaned, so I'm going to finish work on that right now. Windows98 is installed on the drive that's going in (Seagate ST310212A). I've only run some DOS games on it so far and it works pretty good. I'll put it in the case and then install the voodoo. I have not tested anything but the board/mem/CPU.

The floppy was PACKED with dirt, 🤣 Makes sense, though, because the CPU fan was ALSO an exhaust fan. There were NO HOLES for incoming air and the "nothing" panel on the other side had a sheet dust bunny under it. CD drives shared the same fate; lens was completely obscured. It'll be nice if only the hard drive was bad.

I can also report that this was a smokers' computer, back when it was used, as the yellowing that I thought was just "normal" yellowing, washed away... The "nothing" side and top are also plastic and came back to white. Very nice!

Got the Gateway fully cleaned and powered on again. Tested the power supply and had good voltage with and without load. It's got a bit of a squeal, though. This PSU has another oddity: The fan plug that goes to "PWR SUPPLY FAN" on the motherboard; it goes into the PSU, and I didn't look if it just goes straight to the fan connector, but I suspect this is what is happening. The fan did not run during testing.

And ALSO, the case DOES have screw holes for a regular PSU*!

All good, so I'll put the video card back and the sound card and see if those work.

I think Gateway intended for those tiny holes on the front to be the air intake so the hard drive (and maybe graphics card) got some airflow. Otherwise, you would basically have that side fan blowing directly into the PSU fan. It obviously wasn't enough ventilation, though.

The proprietary PSU situation is weird. Some of the cases had the full-size ATX PSU cutout and an adapter plate to mount the proprietary one. Surprisingly, those Astec PSUs (I almost typed "Astec POS," but that might come from seeing modern Artesyn stuff at work) seemed to be the most failure-prone. I didn't see too many of the FSP (branded Power Tronic) models, but the Newton Power ones seemed to be better. Gateway claimed that these had a fan controller, but if the fan doesn't spin at all, you should make sure it still spins with a different power source. I don't think anyone would have made a semi-fanless power supply in 1999.

The power supply fan DOES seem to be working normally now. I unplugged the PSU fan header from the board and it kept going. Maybe I missed something when it was apart. I did test the fan separately after I had it apart for a lube-job.

I attempted installing drivers last night, but my untested CDROM drive was a bust... Of course when I don't test it... SO, I just took it apart and found the obvious culprit (should've seen it earlier, I've been tired as F). The lift guide for the head assembly was off the track.

I reset it, then opened and closed it a few times and BAM, it came to life. The activity light wasn't even lighting up before as it thought it was open still. It reads great now. The Voodoo3 driver disk read right away and I got those installed.

Only one thing left:

And success! I was surprised how smooth it was compared to other 8M/16M video cards I've tested. It looked a bit different as well. Now to install Quake(s) and Half-life!

Ope. What in the heck did this psycho just cook up? It all started when I thought "Hey, I've gotten better at soldiering, I should shorten my test mouse cord". I indeed made it shorter...

Awww, 🤣 Duude, it clicks good, though!

I have wanted to do this since I was a kid!

Counter strike.

I've been playing Quake 2 on the Gateway for the last couple of days. I put it away for a bit to work on a few new entries (and finish a few).

First off, these guys:

Gemlight GMB-P56IPS-V0 and a Micronics Twister AT

The Micronics said it was broken, and noted a bad regulator. The Gemlight looks good, only a dead Dallas. Easy on that one, so I started with the Micronics yesterday.

I started with a cleaning and replace the battery.

I set it up with my test AT power supply (no memory, VGA) and it started beeping, very angry CPU beeps.

I checked the voltage at the regulator(s) and it was good. 3.3 I/0 3.2VC. I removed the CPU and noted the voltages were still readable without a CPU (no angry beeping with CPU out). I set VCore to 2.8V and put a known-good MMX in. Then it was memory beeps. I put in two of the memory modules it came with and a video card and it booted up. The VCore regulator does get HOT for sure. It has been replaced before, it seems, but I haven't checked the datasheets yet. I may need to replace it.

The Gemlight had a soldered Dallas, so that was the first place to start:

I replaced it with a socket and then went to town on the Dallas. I found a pretty fast way to deal with it this time.

The first test was good, so I added the battery and it keeps time/CMOS. I only booted to DOS from floppy so far on this board.

I'm working on the boring PowerSpec right now. Installing Windows XPx64 and drivers only, since it's so boring. I will be back to the Gateway to finish Quake 2 before I go over the AT boards again.

I finally finished Quake 2 (Normal difficulty) . So many flashbacks to when I played it the first time. Very good still.

After that, I cleaned my desk and brought out my computer from work. I was using it as my personal alignment machine. It has been having problems keeping it's CMOS when the power went out. The battery is new.

I have NOT resolved the battery issue, but I fixed a few other issues I haven't noticed before, since it went straight to work. It was NOT able to detect the CPU properly. It showed up in windowsXP as a single core and Everest was also a bit confused.

Checking the BIOS, it was older than the CPU, so that was an easy fix. I updated to the 2006 version, since this CPU was from 2005 (Athlon64 X2 3800+). It was then detected by the BIOS and then Windows (after a restart). Strange that it didn't show up right away. There was a noticeable increase in speed and responsiveness.

I then replaced the single 1GB DDR333 with two 512M of DDR400. With 1:1 Bclk/Dram ratio, things REALLY sped up! It boots like lightning now.

I also added a sound card, (since the onboard wasn't working properly) an Audigy LS (SB0310). It doesn't NEED sound, but it's nice to hear the ding when I need to roll the car forward and back, etc.

I'm going to revisit it in the future and get some help with figuring out why it can't keep time/CMOS. The board will have to come out for that, and it's a real pain. I need it at work, so I'll let it slide for now.

Tonight, I'll be looking at the Micron Twister AT board. I'll be investigating the datasheets on the regulators and possibly remove the vcore one, to replace it, if it's not correct. I'll have to look around, but I think I might have a few. Possibly one I can steal from a bad board, too.

I would like to test the AMD CPU it came with in another board as well; I have a TX98-XV that looks like it will do the job.

Catch-up time again...

I looked up those datasheets, and found, the specs were similar enough to be suitable.

Getting HOT does seem to be the norm, as closer inspection had revealed a bit of a brown stain around that EZ1082CT. I found a suitable replacement on Digikey, but hot DAMN, they want nearly $20 USD each! Noooope. Not until I REALLY need one am I doing that. 89 in stock and the datasheet says it's for a Pentium even.

So with that, I'll call that board "Good". It hasn't crashed or anything weird and has been working pretty solid the last week. So, I then grabbed that TX98-XV from the drawer. I powered it on and it booted, had the time, etc. Good still. I pulled the CPU and installed the K6 that came in the Micronics (K6-233ANR). I changed nothing. (2.8Vcore, CPU was the same speed)

Of course. Works just fine, even at 2.8. I set Vcore and the only difference, was that it got hot. Now, looking at my tray of K6, they are all 2.2Vcore, and luckily the board goes down to 2.1. I 've never tested ANY of these CPUs and was assuming (for whatever reason) that they all work. Well, test after test, they DO work!! The only one I didn't test was a K6-2/500AFX (Clock was far exceeding what the board says it can do). I Ran the 333 at 300 (max of the board) and it worked fine.

Just from the data I collected here, I can see that whatever I build with the K6-2 is going to be pretty speedy.
I have not done an OS test other than DOS yet.

The only issue I was seeing, was with the BIOS. As soon as you move the cursor, the whole screen glitches out. Still usable, doesn't crash, but it's not right. All the K6 I tested did this.

Here's the next thing I worked on:

It was a sexy, chonky, full height external SCSI enclosure with a full-size AT power supply and a EXB-8200 tape deck. Damn, it was so cool-looking and the enclosure I could do something with. Grabbed that boy-o for a single dollar. It was VERY dirty. My brother started taking it apart right away, so I started preparing some new devices to put into the enclosure. My Multispin 6X and a SyQuest555.

The power supply tested good, so we just reassembled it and both drives were functioning. Pretty awesome little enclosure.

The Exabyte drive is absolutely beautiful. I just seen a picture of it clean now, and it's good. I can't wait to test this drive! It's such a chonker! The manual states it need a power supply, all for itself (to avoid SCSI reset). It also has a cooked-looking 5V pin on the molex that went to it! This is one of the most power-thirsty drives I've ever seen.

It takes D8-112 (which I have plenty), so I hope it works!

I think this is the last thing. This Epson AN-4C50. It's a 50MHz 486 with 4MB ram and 512k video. It is probably going to get a lot of work done to it.

The main problem, is that the HDD was broken. My brother replaced the drive a while back, but it's been traveling through time seemingly, very rapidly, so now the screen doesn't work, so that's another issue I would like to work on. What it was here for, is a RTC modification.

I did the deed (knowing it might not fit in the limited space) and it turned out great. BUT, it indeed does NOT fit, but I would gladly take it and put in into my P55TVP4 (it has a fancy, smaller, modern, mod).

I pulled out the 'ol boombox powered it on (I usually check before modifying existing systems) and was completely disappointed. A double whammy, more on that in a bit. The P55TVP4 was NOT able to keep time with the RTC that was in there. Ugh. It does NOT keep time with the one I just modified. (I tested it in the laptop, it kept time in there). Sigh. The TVP4 has an issue other than the RTC. Damn.

And if you thought that was a double whammy, here's the whammer: The RTC from the TVP4 does NOT keep time in the Epson either. Switched to the one I modded, it keeps just fine.

So, the "aftermarket" RTC is sh*t AND my TVP4 has a fault. I'll want to take care of this asap; I like this computer.

So, I was waiting to find another computer to steal a Dallas out of and found a NEC PowerMate VP75 that said "Needs RTC, POSTs - $5". Yup take that. If it needed a dallas, it wouldn't POST, so there has to be one in there still, so I indeed picked it up without seeing the insides at all.

Got it home and started taking it apart. Yes Dallas. But also, there is a ESS 1688 with CDROM interface (unused) and another working floppy drive.

The star of the show finally showed itself after I had the board out. I of course, assuming by the model number, was expecting 75Mhz pentium in there. Added bonus for 5 bucks, right?

Aww jeez. I was so excited when I got the heatsink off. I finally have one in my collection! The GOLD Pentium! I thought it would never happen. That was the best part of my weekend. My brother has the Dallas, he wants to try doing the mod. I have the non-working unit here I will use to test this computer. I'm almost ready, all the parts are clean, just a bit of assembly left.

The PowerMate seems to work just fine. It was locking up after the memory check, while waiting for (F2) and it was simple problem. POST card showed 7E-96 when it halted. This is usless information, 🤣 After checking the memory and jumpers, I realized the "Sleep" header on the board was populated. It was like that when I got it, so I put it back (one of the LEDs is 2-way) assuming it was for the sleep LED. Nope. A sleep SWITCH. The LED was jumping the switch and causing my issue. (and probably the last guy)

Other than that, it works good! I was playing Keys of Kryodacia late into the night and FINALLY finished it. It was one of the hardest "Keens" I've played. I never used "save game" so much before. I would NOT have been able to finish some levels if I wasn't saving frequently. I played Doom2 for a bit, to listen to the MIDI, and it sounds really good. Possibly makes it into my top 3, for sure top 5. Some distinct voicing differences vs. the ES1869 I've been using a lot recently. Duke3d also sounded really good. The "what it's supposed to sound like" sound/instruments.

One of my favorite features about this tower, is the service panel on the bottom. One screw removes the bottom and you can access the cards and CPU and a few other things without taking the casing off (AT style, wraps around the whole thing).

Since I'm stealing the Dallas, it is not very convenient to boot, so I'll try to pop a working Dallas in it someday, as it's kinda grown on me. It's somewhat industrial-looking and has a solid footprint. Damn heavy, too. I'm planning to "retrobright" the face and I know I'm absolutely going to love this thing once the original color is back. Just needs some bay-filler

For the last couple of weeks I've been trying to repair 4 5.25 floppy drives I've been hoarding. I'll want to stick one in here if I can get them all working. I'll go over those, since I'm going to be working on them some more.

1. Epson SD-600. This drive took a few hits and had corrosion all over the board. Did nothing on power-up. I removed most of the stuff on the board in the area and cleaned it all up (replaced the headers) and re-soldered it back together. It now pulls the head back when powered on, but nothing else. Fails floppy check.

2. HP JU-475-4E53. Drive did not read/pass floppy check. Moved jumper to DS2 and it passed floppy check, but considers itself an empty drive. Fully disassembled/cleaned the drive, then reassembled. No change.

3. HP JU-475-2EAF. Drive did nothing. Inspection revealed someone attempted to re-cap it (with used parts) and destroyed the VIAs. One leg on one cap wasn't attached to the board and had a glob of solder on the end. This sucked. I used some leg to bodge all the traces. Kinda jank, but the contact is good now. I indeed replaced the capacitors, too. After all that, progress!! I spend so much time on this drive, I was glad when the light blinked with the head-pull maneuver when powered on. It actually does the floppy check and passes it, too. Disk read error. Almost....

4. Teac FD55F-03U. Drive failed floppy check. Moved jumper to DS1, drive checks and passes. Does not see a disk in the drive. Drive was disassembled and cleaned. Re-test showed no change. Passes check, does not see disk.

All of these drives were tested on Gemlight GMB-P56IPS with a double cable and the drive on the "A" side. (3.25" on B) I tested them on "B" side later and booted from the 3.25. Image Disk revealed the Teac has all "?" for data. The 2EAF has misaligned heads (Top head reads "1" at track 0). The 4E53 is off one track as well. Reads "1" at position "0". I didn't test the Epson again, yet.

They should all be repairable. Not sure about the Teac. Possible bad heads? Maybe it needs caps, too? All of these drives have some knobs for adjusting 'things'. I haven't moved them. Need to read more on floppy drive repair to get some ideas on what to check. Hopefully aligning the heads fixes the HPs.