TechieDude wrote on 2025-11-05, 15:05:

The 2.5V1F capacitors you showed in your other post are the clock capacitors I was referring to. They help the console keep time […]
Show full quote

The 2.5V1F capacitors you showed in your other post are the clock capacitors I was referring to. They help the console keep time for a little while in case it losed power. They're black on every revision except the last. The gold ones are typically more reliable and less likely to leak, but still can't be fully trusted after 20 years.
These sites should provide useful info for you to work on them:

https://xboxdevwiki.net/Hardware_Revisions

https://consolemods.org/wiki/Xbox:Versions

These go into detail on the different revisions of the XBOX and the clock capacitor failures on the earlier revisions.

This led me down the right rabbit hole, thanks! 😁 I see the board in xbox#2 is 1.4, so it should work without the cap. I was already sticking a LED in the holes, just because. I'll leave it and read more on different failures/symptoms, I feel like it's close.

I made it to "Old New York" in the Futurama game and decided to open xbox#4. Wow. Just wow. I hadn't opened a fresh one before, so I wasn't sure how it was held together. Many more screws involved compared to these modded ones, which have NO screws... at all. The casing fits together better, too, slides right together. Actually, the whole thing is like that! It is dirty as they come, but it came apart so nicely compared to the "old wh*res", 🤣

I plugged it in with the board exposed and hit the power button and could hear a light "fizzle" sound for about 2 seconds. It would do it every time. I removed the board (another 1.6). and replaced the obvious plague crew (5x6.3V3300). I slid the board back in and plugged it in. Hitting the power button, the system now boots!

This one only has "Lego: Star Wars" on it, so I went for it. Works great and looks good, getting a bit hot, but it needs to be closed for proper cooling.

I ran Call of Duty on xbox#1 and it looked like absolute crap. Probably because I've always ran it on PC and even with the worst settings on the worst computer, it looks FAR better. Not feeling like I missed anything back then, excepting a few games.

I'm surprised the Xbox can even "run" Doom3. Doom3 was harder to get running on my PC back then... PC looks better, but I couldn't get any kind of "playable" state out of it. 10FPS with dips and freezing is what it was back then. I only had the demo for years, and when I got the full version, my PC was more appropriate (PentiumD 3.75/8800GT) and felt like it was whoopin' that game!

I'm going to properly clean Xbox#4 and replace the clock cap with a regular cap. I think, when I'm fully assembled, I'll get those caps on the A8N-SLI sorted-out...Thing has been laughing at me all week 🙁

Three of four xbox are in perfect working order and two of them actually look pretty good, too!

I then worked on the A8N-SLI for a while, but got tired of looking at it and put it aside again. I replaced most of the capacitors on a QDI Advance 6T that has just been sitting on the floor, because I ran out of places to stack them, 🤣

Testing reveals that it works fine. Speedsys completed successfully, so I wrapped it up and put it away. Not very impressive board, with no AGP. One thing it DOES do, is Tualatin CPUs. I tested it with a Tualaron and it POSTed first try. I'll play with it sometime, but I need to finish all the board laying around.

After that, I immediately started on another board, a ASUS P5LD2. I only replaced the 5x6.3V1500µF that looked bad. Got it cleaned and set aside. I was on fire yesterday, 🤣

I grabbed the Abit IS7 next and removed the two bulgers and powered it on, got POST, so continued with removing ALL of the capacitors. I wanted to test it first, because it looks like it had trace damage on the back. Wetting it and zooming in, looks questionable still, but with how hard it was to repair a tiny trace like that, I'm not prepared for that yet. I figured I'd just go for it and see if it's any good first.

It has all the caps off still right now and I tested the old ones. Almost EVERY capacitor was failed or out-of-spec. Most fails on a single board so far! I'm surprised it POSTed at all.

I was working on it pretty late last night and I ran out of flux. I got some more now, so I can finish it and then, I think I have four boards I can test in a row, before going back to the cursed, A8N-SLI.

I made some changes to the capacitors on the IS7. When it was operational, I removed several more capacitors, so I could probe them from the top while it was running. 25v100µF seemed like a garbage choice to me, so I measured and most of them had 3.3 or 5V going through them. Some were connected to transistors with 12V. I know the caps for the audio did NOT need 25v100µF. I replaced almost everything down to 16v100µF, since I didn't probe the whole board, I'm not sure exactly what was all 12V, so I didn't want to risk going to 6.3V capacitors. I know the caps that looked like they were for USB could get 10v1000µF and the audio is now 16V470µF. I'll do some double checking before sealing-the-deal.

I've read before, that on computers, having tolerances closer to what the voltage actually is, results in better filtering. Others argue, higher voltages can lead to longer life, but lots of noise and work like crap. Which is possibly why, I've seen "budget boards" that work like crap, even when they were new, because they had "cheap" capacitors on them. Cheap being the "high voltage" caps of 25V and 50V as they are significantly cheaper at almost any value. "oops, all one type" is also a way manufacturers cut costs; only having 2-3 electrolytic capacitor types was/is cheap and easy. Strategically replacing them with more appropriate values can add stability to a previously unstable board.

Done with the IS7 and +1 more for testing. I replaced the caps on a P4SD-LA. I tested this board before, so it should work even better now. Several of the 7x6.5V1800µF caps were leaky and open.

I'm testing the IS7 first. It POSTed right away and is keeping time fine. It passed a Speedsys, so it's installing windowsXP.

This install should be good for testing the P4SD and the P5LD2.

The IS7 seems to work great! I started off with a GF2 (since it was universal) and it was able to pass a 3dMark 2001 run.

I don't really have much around to beef that up, but that score (1575) was dissatisfying, so I had to do something.

This red card, a 9700XT, which has been dead in three other boards... works fine in this one. This is the first time I've seen it output anything.

It then continued to have the same problem as usual; install ATI drivers, everything works like crap. I went back to Microsoft drivers and then uninstalled the ATI ones. It works again and does everything just fine. Pretty much every ATI card I have has this problem.

I wanted to test a few untested CPUs in this configuration. I upped the PSU to a 650W and changed the memory to 2x512MB DDR333 and CPU to P4 (2.4/1M/533). Added thermal paste this time, P4s can actually get hot.

I next changed the RAM to 2x512MB DDR400 and the CPU to P4HT (3.2/512/800).

Not a big change between the Pentium 4s. The best thing about this board so far is the POST logo. The splash logo changes with CPU and has at LEAST the three different ones, Celeron, Pentium4, and Pentium4HT.

I'm loading more stuff onto the drive and will stress the board a bit before swapping the board with the P4SD-LA and running the same tests. I'm expecting lower scores and maybe AGP functionality issues. I've never seen this 9700XT working before, so it might not work over yonder.

What are the clocks on that card as ATI never made a 9700xt?

Grem Five wrote on 2025-11-10, 04:37:

What are the clocks on that card as ATI never made a 9700xt?

Wow, I did not know that! I should've looked it up, but you know how it goes... Someone just penned on it with sharpie "9700XT" so I was just rolling with it.

Somehow, I hadn't looked at the info on it yet. It's never worked before, so I couldn't check. Still working today. The specs:

I can see now, that someone mixed their X and T, 🤣 It claims to be a 9700 TX. I messed with drivers again, but still... the card only works with Microsoft drivers and works GREAT when it does. As soon as drivers are installed, absolute slugfest/crash-party.

I'm going to run this set-up for a bit; I'm really liking the board. It is very stable and as a bonus, I have the I/O panel for it!

I figured it was a TX as I think its the only version of 9700s I have seen with infineon memory.

I thought TX is some Dell version.

Anyway, doublecheck the cooling of these cards. R300/R350/R360 cards die a lot when not taken care of properly. Carefully remove the shim around the die if you can.

PD2JK wrote on 2025-11-18, 11:07:

I thought TX is some Dell version.

Anyway, doublecheck the cooling of these cards. R300/R350/R360 cards die a lot when not taken care of properly. Carefully remove the shim around the die if you can.

Just leave the shim where it is unless you can prove that the DIE is not making good contact to the cooler. Not all R300s have this issue. But I agree on cooling, those cards love a good airflow and proper cooling.
3.3ns RAM usually works pretty well on the non 9700Pro cards. Rarely seen it die.

Regarding the drivers, that's a bit odd. I've tested a lot of 9700s and 9800s and the latest ATI drivers never gave me any windows hickups.

tehsiggi wrote on 2025-11-18, 14:20:

Just leave the shim where it is unless you can prove that the DIE is not making good contact to the cooler. Not all R300s have t […]
Show full quote

PD2JK wrote on 2025-11-18, 11:07:

I thought TX is some Dell version.

Anyway, doublecheck the cooling of these cards. R300/R350/R360 cards die a lot when not taken care of properly. Carefully remove the shim around the die if you can.

Just leave the shim where it is unless you can prove that the DIE is not making good contact to the cooler. Not all R300s have this issue. But I agree on cooling, those cards love a good airflow and proper cooling.
3.3ns RAM usually works pretty well on the non 9700Pro cards. Rarely seen it die.

Regarding the drivers, that's a bit odd. I've tested a lot of 9700s and 9800s and the latest ATI drivers never gave me any windows hickups.

To tell the truth, I still haven't removed the heatsink, so next time I use it, I'll give it a proper clean and re-paste. Still no idea as to why the drivers are doing me so dirty.

I've been stuck on a board for the last week. Stuck as in "damn, that's a nice board" and have been playing games on it and making sure it's good before returning it.

I did end up doing caps and testing on the P5LD2 and P4SD-LA. This whole process has been taking a while, since I enjoy sitting around fiddling with my hardware while my software is working hard. I used a WD Expert 310000 as a hard drive when I went to test the P5LD2 and wow... That made it feel like a slow piece of sh*t!! I find this horrifying mismatch entertaining, but not practical, so it's not common practice. I ran out of SATA drives is the horrible truth. I have a Samsung, that works with data, but can NOT boot a windows XP, it crashes. Tried with three different installs on three different computers (775, AM2/AM3) same result. I also have a 2.5" WD Scorpio, which also works, but is invisible to the P5LD2. Works fine everywhere else. Just those two hard drive issues alone cost me hours. I had them prepared on another computer the first time, too.

I started with the P4SD-LA, since the HDD I had set-up for the ABIT IS7 would boot it without hassle. It did. Board feels snappier than it did before the capacitors, but the 3Dmark result is the same. Called it good, moved onto the P5LD2

This same hard drive would NOT boot a 775 (gave it a try anyway) so I had to install XP again. Next subject was the P5LD2. This is where my drive woes started and kept me up waiting for failed installs to fail. I ended up with a Maxtor DiamondMax +9 working. Board passed all the tests and endured some gaming before I called it good. Pretty boring board, but I'm glad it works. It will make a nice replacement for my P5B-E computer, which board failed in the line-of-duty one day. I haven't diagnosed it, but built another computer and stole the HDD/GPU, so we could continue the mayhem. I knew the board was on it's last leg and had that backup pretty much all ready.

This Gigabyte board I'm about to re-test after replacing caps was a total surprise. It had one visually bulging cap. I used it pretty heavily and it puffed up even more. I think it reached max capacity, so I went and replaced 5x3300µF6.3V and 4x1800µF16V. Testing the old ones is where the surprise came in. All of the 3300s tested 8000µF or higher. The 16V ones were in-spec still, but I replaced them anyway. This is a very pretty board. Works very well, too. I couldn't find the exact driver disc for it, but the drivers were easy enough to obtain. Most of them were already stockpiled.

My first test with it, loaded a 80GB with about everything I had to try and stress it, to a certain extent. I didn't play Crysis, so I have no idea how dragging that really is. I have it and have played it in the modern-era, but my PC that has it on, absolutely dominates it. When it came out, I was still rocking 478 Pentium 4 and after, had Pentium 4 551 for YEARS, even made it out of the Dell it was in and into a performance board. Finally kicked the bucket years ago, but this is the system(s) I would have tried it on (with a 6600GT on AGP and 8400GS on PCI-E) and they would've probably been the slugs of legend. Worst part of that story, I DID in fact have Crysis DVD, but I did NOT have a CDKEY!! I do now, 🤣

The board I'm now talking about and testing is the Gigabyte DUAL POWER, K8NXP-SLI.(not on TRW, surprisingly) I'll say it again; beautiful board. First board, I believe, Gigabyte used the "DUAL BIOS" system as well. That DUAL POWER is just so damn cool. It is a riser card with MORE CPU VRM, and according to a marking on the PCB, it can function as a backup OR in parallel (not this model, I don't see any jumpers). It also has a flip-chip, that needs to be re-inserted for running in SLI mode.

I've been testing with a EVGA 8600GT (256MB, boo) and will continue to do so. The memory is Mushkin DDR400, which is pretty cool-looking (heh) as they have heatspreaders, not common in my RAM piles. What surprised me most, like I mentioned, is how this even functioned as well as it did with the junk capacitors. The CPU is a Athlon 64-X2 4800+, so not a cookout, actually, I knew it wasn't going to be, so I didn't waste TIM (thermal paste) on it. The hot-headed P4s definitely NEED it for any test lasting longer than 60 seconds.

The board is still working fine, so I'll play just a little more before moving on. I actually have a quick job to do. It's on a Tyan-based PC with a AT-style ATX case. Pretty cool-looking PC-prop type aesthetic. Such a basic looking unit is appealing to me, since there are no funky edges or weird plastics. Anyway, we'll see it soon. All I need to do is make an image of the drive (C: is Itchy and D: is Scratchy) then reformat and reinstall Windows 98 and copy a bunch of stuff that will run on it. I'll be looking into the board more, too. It should be able to do coppermine, but it doesn't. It might just need a BIOS update. I'll find out soon enough.

Damn. I want to mention where my weekend went real quick: BMX!

I found a sweet ride for $5 at the thrift store and best part: It was one of the ones from the magazines I collected.

Back in the late 00's, I wanted a "real" BMX so badly, and even though Diamondback was pretty much "entry level", I wanted one. What I had: Huffy, magna, and other junk. I got injured so badly breaking those damn things, it made me scared to try as hard when I finally DID get a real BMX (GT Mach 1). I never became an expert or anything, but I can still jump around and do barspins. Anyway, I spent all weekend restoring a 2009 DiamondBack Mr.Lucky to its former glory. The only non-original part was the seat (tires). I had tons of parts still, but it only needed a LOT of de-rusting and balls. Lots of ball bearings were crunchy and pitted, so replacement was necessary. The pedals, crank, hubs/cassette, and headset all needed some replaced. The headset was caged, I uncaged it. It's a pretty fun toy now. I still need brakes, but I just use it at the shop, so it might not need them.

That same thrift moment yielded one of my greatest gaming finds ever: Quake III Arena! I've been copying the same copy for the last 15+ years!

You'd be absolutely INSANE to not grab that for $2.99!! Such an awesome thrift run

After I was done with the Gigabyte, I setup the M2N-SLI Deluxe and tested it with the same drive. Booted right into windows and I already had the drivers, so it just need a reboot before it was usable. Works pretty good. Don't know why, but it feels snappier and quick compared to just running the Gigabyte. I played COD:WaW on both computers and this one WAS indeed a bit slower, since I used Athlon X2-3600+. I don't have many of these CPUs sitting around, so I couldn't do anything faster, but it worked real good (I should note I was running everything in 1024x768, the 8600GT only has 256M of memory, so most games crash if I go higher)

While I was playing with that, I was talking with my brother and he found a Shuttle X50v2+ and were joking about how it was supposed to run windows 7, which, Shuttle was bragging about. The thing is pretty slow. Dual-core Atom 1.8Ghz and 2GB SODIMMs (DDR2). Windows XP is supported, so he's going with that, but that had me wondering about running 7 on slow, older hardware. Googling it led me back here, but what set me off, was Google AI chiming in and trying to tell me to NOT do it. Pffft. Fat chance. It's happening now!

And here she is:

I used that Advance6T with the 1k Tualaron and 2x256M PC100. O.M.G. It took 2.5 Hrs to install and it can't do ANYTHING with it's onboard video. Windows doesn't go back far enough for me to install a PCI (no AGP) FX5200. The earliest support from Microsoft, is 6 series.

Conclusion: It was doable, but not practical on this particular board (or any board, I'm sure). The GUI is very slow. Everything you do takes time to think about it. I haven't done nothing but run the GUI and the computer is sweating bullets. It uses 165MB of memory at desktop (I already modified the services, so I was probably far higher. It WAS much slower initially).

Overall it was pretty fun and felt illegal, 🤣

Got another sweet pile of junks yesterday! The items of great import are: Athlon 64-X2 5600+ and a GeForce 8600GT. The GeForce is important only because it is the EXACT same one I've just been using to test my PCI-E boards!! What is that!? Do I hear SLI calling?!

The other item(s) were bracket-less coolers. After that, I found ANOTHER sweet vintage toy:

This. This is going to be fun to play with! I'm not going to get to it for a while, but I'm excited for it.

So now, with a beefy X2 and dual 8600GTs, I decided to put together a build. I gutted an ugly one to start with.

It had a very ugly board in it; still haven't tested it, 🤣 I just can't be interested in it; at all. I DO like the case though - a lot. This system will be a near duplicate of my work PC, but with SLI. I have a matching set of Nanya 2G DDR2-800 to put in it, too. So random that a set I had matched a pair I got at another time. I need to clean the GPUs still, but it's coming together. I did a bunch of testing with the "new" 8600GT already, and it work just like the other one, just a bit hotter. After service, I'd expect them to be the same.

I'll be finishing this up and then cleaning some more boards, testing, and putting them away. I almost have my surfaces back.