I got curious about some comments on this TI4200s and went go dust off mines. Turns out that I have the exactly opposite situation of vutt.
In my case the MSI is the one with 64MB, faster memories (3.6ns) and better Phillips TV chip with VIVO. The Asus has 128MB, 4ns memories and worse TV chip (output only).
I quickly installed the MSI on my XP setup and it overclocks fine, I can get it to 300Mhz with the stock cooler and it seems to be stable running HL2.
I don't have the stock cooler for the Asus, the passive cooler on it right now is not ideal for overclock, so I didn't even bother.

Not sure if that counts as a retro activity but today I successfully installed (on an external drive) and tested MacOS Sonoma on my 10 years old MacBook Air. That little thing is really well built and in those 10 years I never had a problem with it besides replacing the battery once and a ripped cable on its magsafe power supply unit.

It is commendable that Apple offered continuous software update for 10 years, now thanks to the OpenCore Legacy Patcher community I can upgrade it to the latest version of MacOS and enjoy a few more years (I hope 3 years) of updates.

At the time of purchase I chose to max out the specs in terms of CPU, RAM and HD capacity, to "make it last longer", and I do not regret it, that was money well spent.

Not much activity today but:

- secured a pretty rare Xbox 360 - it has the original (although in need of SW repair) "Blades" dash that they originally shipped with. Xenon revision but hey, I'll take any kind that has blades, honestly 😀
- packed an ASUS K8N for shipping w/ 80GB HDD and a Radeon 9550 (ASUS A9550GE - not the best, not the worst either.)
- cleaned up a QDI KinetiZ 7B for comparision against my ECS K7VZA R3.0 (aka Shuttle AK12A). Gonna be a fun evening tomorrow, especially since one's the original KT133 chipset (non-A) and the other is a 133A.

Pino wrote on 2024-05-01, 20:34:

...
I don't have the stock cooler for the Asus, the passive cooler on it right now is not ideal for overclock, so I didn't even bother.

Have you used it with passive cooler?
Asking because I as well do not have original cooler only heatsink. I'm trying to find good balance.
It runs well with el cheapo 2 slot Chinese cooler, but it's overkill. Now I'm experimenting with original heatsink and brutalized Noctua cooler (pic below). I kind of like aesthetics and it's also quiet, but clearly less powerful. Double sided sticky tape mounting technique is also ... ... lets call it prototyping phase 😉
Seems to be working OK. I mean if you were able it run passively then I should be safe with low rpm Noctua fan.

Yes, it can run with this passive cooler, but it gets toasty.
While this passive heatsink it's not original from Asus, I got it from a third Ti4200 that was dead, so it's supposed to be enough for a ti4200.

You should be more than fine with your custom solution / noctua fan, even the original cooler heatsink combo at my MSI card is really light, not much aluminum on it.

I can confirm that the GF4 Ti 4200 can run with a passive heatsink, like the GF3 Ti 200. But it does get pretty hot, and probably isn't the best thing for it in the long run.

Even a small fan makes a big difference.

Can also confirm that a Ti 4200 can run with a passive heatsink. I'm using a Zalman fanless cooler, which is basically two giant heatsinks connected via a heatpipe.

Also have a couple 140mm intake fans that help push air towards it for good measure.

There are Dell oem versions of the gf3 ti200 and the gf4 ti4200 that are passively cooled. They do get toasty, but I haven't noticed any problems so far.

For the first time ever I built myself a socket 478 Prescott rig for testing stuff. Cpu runs at 2.8ghz, on a Gigabyte 8ipe1000 board. Using the copper core intel box cooler, the load temps are fine, going to about 55C. The load power consumption is not so nice, about 165w in Prime95.

Mostly I want to see just what it's like, I avoided the P4 like the plague back when it was new and I was building my gaming rigs.

I for one stick with Northwood HT if I want a high end P4. Prescotts and Gallatins aren't worth the cooling hassle and the SSE3 instruction set IMO.

For 775s it's usually Cedar Mill and Presler - Prescott and Smithfield are instantly discarded. Note that this applies to the boards that do not allow any other CPUs besides CM and Presler.

I like the advantage of having SSE3 for dual booting into XP. Modern processors already run pretty hot, so the added heat isn't as much of a surprise now as it was then.

I've come to really appreciate the ULV C2D line of processors recently. They were designed to be halfway between a netbook and a laptop, totally underpowered for anything current at the time, but very capable of running software on XP up to 2004. Thanks to the popularity of retrofitting iPods that used similar tiny 1.8" IDE hard drives, there are adapters to use MSATA SSDs. I'm using a Dell D430 with a 512gb MSATA drive and a SlimSCSI PCMCIA card to work as the control interface for an old E-MU sampler and an M-Audio Ozone keyboard/audio interface.

In the last month I picked up a Dell Latitude XT, very early touch-screen convertible using the same ULV processor but slightly better, with Radeon X1250 graphics and able to handle an extra gig of RAM. The Radeon uses shared memory, but 256mb instead of the 64 the Intel in the D430 uses and using 32-bit XP with 3GB means that's not usable RAM anyway. So it's very capable of playing games up to the 2004-ish range, and as a C2D it performs better than many higher rated single core processors of that era. Tablet XP is definitely crap, but it does work, so I've also been using it to read comics, either with locally with a viewer or with MyPal browser with uBlock Origin.

Laptop cost me $20, had the RAM and drive adapter lying around, picked up a larger MSATA drive but that was because I wanted to load lots of software. Already had power supplies. Worst problem was that the ZIF drive connector is delicate, and the XT doesn't have any way to actually steady the drive, so the ribbon connector to the adapter falls out unless you find a specific way to tape it down. Pulling the battery reveals the drive, so if you want to swap batteries it might unsettle the drive, meaning you have to re-tape it. I broke the ZIF IDE connector on the motherboard of the first $20 XT by having to pull it so many times. Luckily I found a motherboard for sale for $20 and swapped them, everything works, but I am not messing with it again unless I buy a new battery. The one I have still holds a charge for an hour or so, at least.

Ah yes. Today I finally got to organizing my processor and memory collections. I have inventory of all my CPUs now, in and out of systems. I got over 60 including the ones in systems. I didn't think I had so many until I sorted them all. Pretty cool to see them in one place, too.

I've spent the last couple of days trying to solve why my ratbag system with the Amptron M930LR has been locking up. I first noticed it when I installed and played Max Payne on it, and I assumed that it was too aggressive memory settings. So I tried different memory, then moved on to trying a different video card. I reinstalled Windows 98, but the problem kept coming back. My hunch was that it was the motherboard, an old, cheap board that I've recapped. I decided to try and install Windows 2000 to see how that went. First off, it gave me a blue screen with "This BIOS isn't fully ACPI compliant" which confused me because I had installed Win2K on this board before with no issue? I disabled ACPI in the CMOS settings and moved on, but then it would get past the point where you entered the key, then restart and begin setup again, never getting past that point. I tried a different CPU, no change. I pulled the board out of the case, and inspected it, but I found nothing.

Wanting to leave no stone unturned, I decided that I should try a different hard drive (I had also tried a different SATA to IDE adapter). I had no confidence that this would make any difference, it was a Kingston SSD, one of the newest and best quality parts of the system. And that was the problem! With another SSD, Win2K installed with no problem! I did a search on the ACPI issue, and I found a forum where a guy with a dual Slot 1 system had this same issue; it once had Win2K but now when he tried to install it again, it said the BIOS was too old but it was supposed to be ACPI compliant. Someone else suggested that this can happen if the printer port is disabled in the CMOS, and his was disabled. He enabled it, and the problem was solved, so I tried that as well and it also solved my issue.

This is the first time I've had an SSD fail, and it behaved as you would expect a from component that's overheating, so I now know to suspect them when I have such symptoms.

Today I did some mild overclocking on the 2.8ghz P4 Prescott, and using the box cooler and default voltage, it went up to 3.73ghz, P95 and 3dmark2000 stable. System power consumption at load went up by about 20W. Load temps are about 60C. That's basically a free 33% overclock, I honestly wasn't expecting it to go that high with stock cooling and voltage. But I'm not going to push it any higher, because I don't want to blow up the mobo VRM 😁

Shponglefan wrote on 2024-05-02, 23:15:

Can also confirm that a Ti 4200 can run with a passive heatsink. I'm using a Zalman fanless cooler, which is basically two giant heatsinks connected via a heatpipe.

Also have a couple 140mm intake fans that help push air towards it for good measure.

Good that you've got some air blowing over it, the VRM cooling is probably the bigger issue now but you've got that solved. I discovered recently prodding around with the thermal cam that my FX5200's ISL6529 voltage regulator controller gets up to 80c in normal operation. Not the MOSFETs, the controller itself was finger burning hot and that seems like it might actually be normal? I swapped out the chip which made no difference and the ISL6529 moved to a package with a thermal ground pad on the bottom for thermal reasons.
Seems to just be a facet of older voltage regulator controllers but I did temporarily remove my voltmod to make sure I hadn't messed up

The last few days I've been stressed out / sad because my 'expendable' Aopen AX6BC motherboard became expended. In hindsight it was a comedy of errors, I saw Necroware / scorp's video on modding early S370 boards for coppermine ability and thought I'd try the same mod (relocating the RST pin) on a celeron slotket with this board. Then there were no POST codes and I thought I had damaged the chipset.
So I swapped to another W29C020 EEPROM fresh from ebay since that's the last thing I damaged and that continued the problem. I got the 2nd board out of storage and testing that - it won't come out of reset! Oh no! Both boards are dead by my hand????
Then I put the 2nd board's EEPROM in backwards and it melted the award holographic sticker ergh. Gotta work on how to determine pin1 position when refitting chips.

Now, I like the AX6BC because it's a pretty run of the mill coppermine capable 440bx board but nothing special, however... two of them, dead? Well, they came from a scrap lot and they have a history so nothing's that simple.
BTW I probably have somewhere in the region of maybe 10 or a little more slot1 boards but this is the only one where I have a pair of them and they're both capable of doing everything I want i.e. have AGP, ISA, coppermine support, so they've become important now. Oh and they stopped being cheap and plentiful a few years back.

On the 1st board I eventually decided to clean the socket with some card sprayed with IPA and swap back to the most recently working BIOS chip, and it worked again??? slot connectors are finnicky it looks like. The new EEPROMs which are chinese remarked type chips turn out not to work with this board which means these ebay special Winbond W29C020 chips fail at the very job they were intended for. Hope they work on something else.
Seems like the slotket dirtied the slot1 contacts to stop it working, then I swapped out the EEPROM for a bad one which stopped the board working with anything else.
This was all worked out pretty quickly as soon as I pulled out the oscilloscope, I knew the processor VCC was good but the scope let me see that clocks were all working fine and the computer was trying and failing to read data from the BIOS chip. Somehow I always feel that whenever I pull out the oscilloscope, I quickly reach an unsatisfying but meaningful conclusion / understanding of what's going wrong.

The 2nd board which was stuck in reset, well when I got it, Q7 had gone pop and I don't know whether I ever tested it. I took the neighbouring Q5 transistor off since they're marked the same and tested it as an NPN.

The same Q7 transistor gets pretty hot on the working board so idk if it's poor design or should be a PNP or what, but it'll get a heatsink. My understanding so far is that this circuit controls the ATX power on / power off maybe? The 2nd board could not be powered off after powering on via ATX power button.
Where I'd soldered the working Q5 transistor back in maybe I didn't do a good job? Reflowing the solder with fresh flux and suddenly the board is no longer stuck in reset and can POST with its original BIOS chip, so I'm back to two working AX6BC boards, hooray!!

Doing IRL stuff in the interim sure did replenish my energy for working on these though. Super glad they're both working again and I don't have to use one my fancy boards like my recently acquired Abit BH6 for general testing. In the end the faults were just 1. dirty slot1 contacts, 2. a bad chinese remarked EEPROM chip and 3. a soldering error on a transistor

I know this will sound not in context but fired up the old XP box and played some BF1942 and BF Vietnam single player after not doing so in about a decade. Long ago mid 2000's thought I was good but guess not 🤣
Recapped a board and fixed a bad bios flash too. Thats all folks....

Today I connected my IBM E74 CRT monitor to my Macintosh FrankenMac 8600 and played some Starcraft.

mc.jesus wrote on 2024-05-04, 04:46:

Today I connected my IBM E74 CRT monitor to my Macintosh FrankenMac 8600 and played some Starcraft.

Welcome!

Replaced two caps off my recapped Soyo 6BA+IV.

I forgot Sanyo WF caps tend to bulge in storage, oof. Thankfully, it was just two 1800uF 6.3v I had used near the CPU and the rest of the board got Rubycon MBZ, MCZ and Panasonic FL/FJ.