TPU's Nostalgic Hardware Club

[Ruru]

What a nostalgic feeling, installing XP from a cd

Will be testing different parts, but let's start with Duron 1.2GHz, 2GB DDR & Radeon 9700 Pro. I probably end up with XP2700+ & X800 Pro, but let's have some fun first with something 2002ish

I just hate those old VIA chipsets since they just don't like anything else than SATA 1.5Gbit HDDs, luckily those first SATA 3Gbit drives can be ran at 1.5Gbit/s mode. I've had the same problem before so it was pretty easy troubleshooting.

And it also fits perfectly under my TV and PS3!

Flashing BIOS was also a journey itself, I tried burning few bootable CDs with no luck, but using another HDD with a FAT32 partition, I put the BIOS file and flasher there and booted using FreeDOS from USB drive.

 

[Susquehannock]

Bad caps. I find the whole 'capacitor plague' fascinating and have been following it from early on.

Last I understood was the talent responsible for Abit's well regarded overclocking boards left for DFI.

From what I remember, buzz at the time was that he left because he did not like how Abit was being run. Not long after, the company was caught for bad accounting practices artificially boosting sales numbers by having several fake companies at same address. Likely an attempt to stop bleeding from the bad cap suits.

I think ABit was a victim of the bad-caps thing like everyone else back then, perhaps it just hit them harder.

Agreed. Seems that Abit got most of the bad press since they admitted to using bad caps very early on. As did IBM. Many companies avoided the issue all together. Worse yet was Dell. They were caught lying to customers about failures due to bad caps and were fined $300-million for it. Their reputation now ruined, Dell fell from top PC maker to a distant third in short order and have not completely recovered.

I am not a foil hat type, yet some do believe the 'capacitor plague' was an intentional move by certain entities. Don't quote me on any of this. If memory serves the so called bad formula was stolen in China, then moved through Chinese division of Taiwanese manufacturers. A long game play to undermine smaller companies and eliminate competition? Lenovo did buy IBM's computer division at a discount not long after news of the plague broke. And now they are world's largest PC maker. Points to ponder.

 

[Ruru]

I remember things been pretty much as you remember. That sucked hella lot, since Abit was an awesome manufacturer.

 

[Susquehannock]

I remember things been pretty much as you remember. That sucked hella lot, since Abit was an awesome manufacturer.

Sure does. Abit was great. Maker aimed squarely toward the overclocking enthusiast. Weren't they the first to introduce soft BIOS adjustments - no more DIP switches? I have a pair of Abit KR7A boards. Early one made with known (after the fact) bad caps, and a later revision with better caps.

Another maker I really miss is Soltek. Think I mentioned all this several pages back. Great customer service. They would send re-cap kits free of charge regardless of board model or proof of purchase. Got one for my FRN2 board but never did the switch. Plus they continued to revise their nForce2 boards where most other makers simply dropped it. Last revision the layout was reconfigured to use better branded caps of different values. Reminds me that I need to get that box back together.

Speaking of nForce2 boards ... surprised nobody has snatched up this EPoX nForce2 with Barton. Guess because it's 8RGA and not an Ultra 400. The EPoX 8RDA, along with DFI Lan Party are ones I would really like to have.

 

[dorsetknob]

Last and most common Chipset that used dips or jumpers was 440bx ( it was 440bx board that let me overclock my P2 300 to 650Mhz with additional cooling fans) "oh heady Days"

 

[biffzinker]

Speaking of nForce2 boards ... surprised nobody has snatched up this EPoX nForce2 with Barton. Guess because it's 8RGA and not an Ultra 400. The EPoX 8RDA, along with DFI Lan Party are ones I would really like to have.

Nice find, almost looks like my EPoX 8RDA+. I think I remember I had it paired with a Athlon XP 2500+ that I ran OC as a 3200+

 

[Ruru]

Last and most common Chipset that used dips or jumpers was 440bx ( it was 440bx board that let me overclock my P2 300 to 650Mhz with additional cooling fans) "oh heady Days"

I remember overclocking via jumpers also on my Asus P2B, P3-500 @ 575 with its passive heatsink. Didn't have the skills to push more (read: didn't know that it needs more voltage)

 

[lexluthermiester]

( it was 440bx board that let me overclock my P2 300 to 650Mhz with additional cooling fans)

How the bloody hell did you do that? You must have been using a ton of voltage! Or did you mean 450mhz?

 

[biffzinker]

How the bloody hell did you do that? You must have been using a ton of voltage! Or did you mean 450mhz?

That had to been a mistype. No way a 300 MHz Pentium II is going to scale up to 650 MHz without some extreme cooling.

 

[lexluthermiester]

I've heard that being shelved for a long time is bad for motherboard capacitors and they'll end up bulging/leaking, but I'm not so sure about that theory.

It's not the storage aspect, it's more just the chemistry of the time was imperfect and it was a serious problem. It's not difficult repair however. You really should get a bit more ram for it unless you're planning on running a 9X kernel version of Windows.

That had to been a mistype. No way a 300 MHz Pentium II is going to scale up to 650 MHz without some extreme cooling.

That's what I was thinking. The most I ever saw, even with high voltage, was about 500mhz. I'd imagine the only way someone would get 650 is with LN2, and even then you'd have to chill everything else as well because of how all the clocks were linked.

 

[dorsetknob]

Early P2 Celerons were known to clock higher
P2 and Celerons of this era had unlocked multipliers (you set the speed/multiplier/fsb by dip/jumpers)
and Certain Batches were Sold at lower Speed (Same SL code and Stepping as P2 450 and Same Voltage requirments ).
Shop that i bought it from i made them check the Sl code for the one i Specificly wanted and in the End i got them to order the one i wanted.

It pissed them off that i would not buy one they had in stock

 

[Mr.Scott]

It is indeed possible, but I believe there is a little exaggeration going on.
There are PII 300's in the 600 range on air alone at HWB but you have to take those with a grain of salt because adequate proof is not there.
I personally have a few 300's in the 600 range, but they are on the chiller.

 

[dorsetknob]

It is indeed possible, but I believe there is a little exaggeration going on.

3 x 80mm fans and 2 x 1000w hair dryers blowing cold air ( very noisy )

 

[lexluthermiester]

Early P2 Celerons were known to clock higher

True, but that's because the L2 cache was on the die. The P2's however had external L2 cache and could not clock as high...

3 x 80mm fans and 2 x 1000w hair dryers blowing cold air ( very noisy )

The highest I ever got a P2 300 was 549mhz on a 183FSB. It could get to Windows but wouldn't benchmark anything or run games. Had to get a leafblower to keep the thing cool enough. My wife at the time thought we were all crazy! The best stable OC of a P2 was 500mhz on a 166FSB, but only with lots of really noisy high-speed fans. Liquid cooling would have been fun back then!

 

[havli]

How about direct clock-to-clock comparison of NV30 and NV35?

It is interesting to see how much faster GTA VC is on the second generation FX. This is not some random mistake or a bug, for example FX 5700 is also twice as fast compared to FX 5600 when running 4xAA + 8xAF. Hard to tell what is the reason though, other games are running more or less as expected (only minor improvement).

 

[agent_x007]

It's probably linked to Ultrashadow, twice the capacity with 256-bit width memory bus (more memory bandwidth), and "shader tweaks" :

Source : LINK

 

[Ruru]

FX 5800 cards are hella expensive, but I've been thinking of Quadro FX 2000, they're FX 5800's Quadro variants.

 

[8BitZ80]

It's not the storage aspect, it's more just the chemistry of the time was imperfect and it was a serious problem. It's not difficult repair however. You really should get a bit more ram for it unless you're planning on running a 9X kernel version of Windows.

RAM is the major limitation with the 815 chipset, the maximum is 512MB. Should be enough but yeah I'd normally aim for at least 768MB.

 

[lexluthermiester]

RAM is the major limitation with the 815 chipset, the maximum is 512MB. Should be enough but yeah I'd normally aim for at least 768MB.

Oh damn, that's right. Forgot about that. I'll just shut up then. Carry on! LOL!

but I've been thinking of Quadro FX 2000, they're FX 5800's Quadro variants.

This. The Quadro's are excellent cards for the money! They generally came with more ram as well.

 

[biffzinker]

@AsRock, and maybe @Trekkie4 is this something you would be interested in? Just showed up over on Hackaday.

No, you aren’t looking at a 30 year old Teac graphic equalizer that somebody modified. The MWA-002 Network Music Player created by [GuzziGuy] is built entirely from new components, and easily ranks up there with some of the most gorgeous pieces of homebrew audio gear we’ve ever seen.

Internally, the MWA-002 is powered by a Raspberry Pi 3 running Mopidy to play both local tracks and streaming audio. Not satisfied with the Pi’s built-in capabilities, [GuzziGuy] is using a Behringer UCA202 to produce CD-quality audio, which is then fed into a TPA3116 amplifier. In turn, the output from the amplifier is terminated in a set of female jacks on the player. Just like the stereo equipment of yore, this player is designed to be connected to a larger audio system and doesn’t have any internal speakers.

Scratch Built Media Player Channels 1980s Design

No, you aren’t looking at a 30 year old Teac graphic equalizer that somebody modified. The MWA-002 Network Music Player created by [GuzziGuy] is built entirely from new components, and easily…

hackaday.com

 

[Grog6]

Guys, a note on the bulging caps; the bulges are due to the gas production of the dielectric reacting with the aluminum plates.

The aluminum plates are etched to increase surface area, and only one side has a oxide layer; that's done in a "forming" step, that makes it a capacitor instead of a firecracker.

The reaction of the dielectric causes gas generation, but the bubbles also form in between the plates, decreasing capacitance.
The lowered capacitance makes them run hotter, and the leakage increased, so it's a degenerative process.

Much like the strip club, If you see bulges, run.
Or recap the mobo.
(Strip club, def run.)

 

[Ruru]

PhilsComputerLab & RandomGaminginHD are my favourite youtube channels what comes to old hardware

HT was kinda ahead of its time when it was introduced in Pentium 4.

 

[biffzinker]

HT was kinda ahead of its time when it was introduced in Pentium 4.

I don't imagine Windows XP was properly optimized for Hyper Threading as well even with the necessary service pack installed.

 

[lexluthermiester]

I don't imagine Windows XP was properly optimized for Hyper Threading as well even with the necessary service pack installed.

Actually and oddly enough it was. It was already optimized for dual CPU setups so it was a seamless transition. Even Windows 2000 was seamless.

 

[MaxiPro800]

Gonna leave this spicy photo here:

Tahiti core.