I have to vote for Socket 7 boards from ASUS or Gigabyte with the Intel 430VX, HX or TX chipsets. They just wont die and I never encountered a DOA board.

Also they are very stable and have no fuss about them.

I think some of the reasons modern computers seem to die more, and seem cheaper/less robust are:
A: Simple scale. More computers, more failures- even if less per hundred.
B: More complex, a push by consumers and corporations to push 'prototype' technology to get a one-up. See the S7 and it bursting into flames- computers in this decade are designed with a 1 year, maybe even less, life-cycle in mind. The Apple II was sold for a decade, and it was par for the course that a flagship unit wouldn't be replaced for at least 5 years. The C64, the IBM PC, these things were all built to last and proved and refined the technology they used during their lifespan. Nowadays you're lucky if you're using the same SOC a year from now in your flagship thing, Intel changes chipsets yearly, etc. The cycle of lab to consumer has become incredibly fast, but that means there's simply less time to test. Combine that with the above massive scale, and problems that would have never been seen 30 years ago- if you sell 2,000,000 ships and 1 in 10,000,000 have a flaw that lights then on fire, it's possible nobody will ever know. Sell twenty million, and that changes- are seemingly all over the place.

This is also why we're stuck in the asinine push for thinner phones. Companies are out of legit good ideas and need something to justify the 1-year lifecycle, so tiny incremental improvements or releasing stuff that should be prototype (first generation 4G was very much this), become huge selling points. The end result is the consumers get more cutting edge stuff that doesn't last as long and I don't think this is a /purposeful/ planned obsolescence thing, it's simply a result of how things became. When your corporate structure is based around greater YOY profits, you get stupid eventually.

On a more on topic note: 440BX, obviously, but the Dell XPS based on that which my dad bought in 1997 is still 100% working despite having been left outside on the curb before I had the sense to bring it back in. Thank god.
I have a Pentium III 933 that has outlived several boards and whole system builds, which was once overheated due to a poor heatsink fitting and is slightly bent. Works 100% fine.

Further note: Retro hardware used sparingly for retro stuff will be used far less on an hours per day basis than if it was your primary work/recreation machine. Using it as a primary machine for a thousand hours a year versus using it as a hobbyist machine for a hundred tends to make it last longer and seem invincible- even if it turns out it's going to die at hour 205. That's two years to you, only a few months back in the day.

vetz wrote:

I have to vote for Socket 7 boards from ASUS or Gigabyte with the Intel 430VX, HX or TX chipsets. They just wont die and I never encountered a DOA board.

Also they are very stable and have no fuss about them.

The ASUS P/I-P55T2P4 is not a particularly reliable board (I've come across a few dead ones). The VX97 is better in that regard.

It's no bug, it's a feature. 😉
I'm afarid we haven't mentioned another reason for unreliable modern hardware yet: Planned obsolescence.
Normally, I wouldn't bring up something like this, because it is a rather polarizing subject.
However, there are certain products which do make me belive manufacturers really do take advantge of this technique.
Like modern printers or electric toothbrushes. They do include weak plastic gears and other prone-to-fail components,
which aren't much cheaper than quality parts. Saying this, I'm not stating that PCs are belonging to this same category, but it could be possible.
I mean, there are even some software firms which do release free upgrades to an older product, just before the release of a new OS.
This updated version of the old product then refuses to run on the new OS, even though the non-updated version runs just fine..
Even more, I watched a discussion of a developer asking in a forum on how to make his software incompatible with new OSes (he wanted the "compatibility tab" to fail). 🙁

kanecvr wrote:

vetz wrote:

I have to vote for Socket 7 boards from ASUS or Gigabyte with the Intel 430VX, HX or TX chipsets. They just wont die and I never encountered a DOA board.

Also they are very stable and have no fuss about them.

The ASUS P/I-P55T2P4 is not a particularly reliable board (I've come across a few dead ones). The VX97 is better in that regard.

Yeah, but the T2P4 tended to have the nuts overclocked out of it.

gdjacobs wrote:

kanecvr wrote:

vetz wrote:

I have to vote for Socket 7 boards from ASUS or Gigabyte with the Intel 430VX, HX or TX chipsets. They just wont die and I never encountered a DOA board.

Also they are very stable and have no fuss about them.

The ASUS P/I-P55T2P4 is not a particularly reliable board (I've come across a few dead ones). The VX97 is better in that regard.

Yeah, but the T2P4 tended to have the nuts overclocked out of it.

... I doubt many people overclocked these back in the day - at least not here (where I'm from), where PCs were so expensive in 1995-1999 one cost as much as one years pay - trust me - no one would do something to risk destroying something worth that much.

I usually come across these in unsuspecting beige AT boxes from time to time - found 5 so far - 3 are dead. The PCs were intact and the rest of the components worked fine, sans the PSU in one of the machines with a dead board. The owners either don't know the machine's history, or they swear it was working last time they checked (usually lots of years ago).

They seem to come in several versions too - some have a dallas RTC, others have a CR2032 coin cell battery slot - some have PS2 while others have the PS2 header removed.

Seems to be that the bad cap plague was mostly late Slot1, Socket A, Slot A, and Socket 370. Maybe some early P4 systems as well. I rarely see bad motherboards after the P4 era unlesss it was taken out by a cheap PSU imploding. I also think socket 7 boards escaped this mostly.

Cheap PSUs we have an issues with in every generation. SlotA and Socket as said ran hot in general and would completely fry if the CPU fan died. PC-Chips and Amptron had these motherboards that had everything onboard including a 56k modem and they dropped like flies.

Jo22 wrote:

However, there are certain products which do make me belive manufacturers really do take advantge of this technique.
Like modern printers or electric toothbrushes. They do include weak plastic gears and other prone-to-fail components,
which aren't much cheaper than quality parts.

"Aren't much cheaper than", but cheaper all the same. If you can shave off a few cents per part, to save a dollar or two per product, that becomes huge at the millions-of scale, if not in unit-price, then potentially in supply lead time and opportunity cost.

For something like an electric toothbrush, which is already under a heavy price squeeze due to stiff competition, being a little cheaper to manufacture is a big deal.

I don't really buy that "planned obsolescence" is some kind of malevolent conspiracy, I think it's just a response to our buying habits and market competition.

Don't know if productions costs are an issue. The consumer usually has to pay the multiple of that anyway.
I don't think he or she would care if something already expensive costs 10 cent more or less.
I mean, it isn't the first time that manufacturers play bad tricks on their customers.
For example, there was this car company..

Anyway, I'm the least one to believe in conspiracies. But this behaviour would make sense for coldblooded marketing-people.
Why produce a product that lasts, when you can sell it again and again ?

Another reason against quality products is our throw-away society.
Repairing stuff is out of fashion, it just isn't cost-effective. Replacing stuff is much quicker and less costlier.
Just think about all the repair technicans in the service department that could be fired..

10 cents per unit results in a million dollars or so that quarter and a bonus, so you bet your butt people do it.

Okay, folks, I got it. For the survival of a company it is absolutely necessary to implement cheap
low-quality parts into their products' core elements. If they wouldn't do that, they would face bankruptcy.

Edit: Changed by a change of mind. 😀

Jo22 wrote:

Okay, folks, I got it. For the survival of a company it is absolutely necessary to implement cheap
low-quality parts into their products' core elements. If they wouldn't do that, they would face bankruptcy.

Edit: Changed by a change of mind. 😀

Or have a blend of the two. You get what you pay for. Save the highest quality parts for the most expensive models.

But for me, if I designed computers, I would do like they did in the 90's - cheaper systems were slower and had less features but were just as highly reliable as the expensive high priced performance series. I just don't have the heart to dis my consumers. If I can't stand behind my products, I don't deserve to sell them. Even if I was the only one out there who cared about the cunsumers, it would inspire other companies to change for the better due to me taking the majority of their sales, thus risking them going out of business if they don't change their ways.

nforce4max wrote:

Every generation of hardware had its failings, if it wasn't something like shitty power supplies or rotting caps it was always something else like the barrels of cancer during the 386/486 era that leaked eating away at everything. Modern stuff fails more often because too many corners are cut quality wise while being so much more complex. Doesn't help that the skill level of builders these days is lower than past eras as they let things like vrms get far too hot for far too long so they end up losing $500-800+ dollar graphics cards or doing stupid things that kill their SSDs ect. Yea those cheap planar TLC nand SSDs are cheap for a reason....

Personally I've settled into playing with older hardware and keeping things to a minimum when I need to deal with something modern like a new build ect. I don't like modern laptops much and despise the shitty ultra thins!

looking at it all wrong, today a computer is cheap enough to throw it away ever 2-3 years and just buy the newest fastest thing. No need to worry about that it will work in 10 years, its not going to be used that long for 90% of people anymore

SRQ wrote:

The C64, the IBM PC, these things were all built to last and proved and refined the technology they used during their lifespan.

Well, I agree on the IBM but no so much on C64. They were pretty much thrown together using whatever parts were available at that time. Some had issues with mem chips, some had issues with other chips. I've had about a dozen C64's go through my hands over the years, and around 30-40% were somewhat faulty or had lower quality parts used unlike some other otherwise identical C64.

IBM's from the 80's and early 90's seem to last forever though. Obviously they have some week spots like floppy drives, but all the essential parts are still mostly functional after all this time. Many clones suffered a failed motherboard in just a few years and were scrapped. The clone 486's I've had around half has had a faulty motherboard.

sf78 wrote:

IBM's from the 80's and early 90's seem to last forever though. Obviously they have some week spots like floppy drives, but all the essential parts are still mostly functional after all this time. Many clones suffered a failed motherboard in just a few years and were scrapped. The clone 486's I've had around half has had a faulty motherboard.

In my opinion, IBM PCs are decent, but no more than that.
I've seen various good clones that were designed more cleanly, and are at least as reliable as the IBMs. I have three 8088 machines, all from 1987. A IBM PC/XT 5160, a Philips P3105 and a Commodore PC20-III.
Of these three, I'd say the Commodore has the nicest design. The motherboard is fully integrated, so it does not need any expansion cards for operation. CGA/Hercules, COM port, LPT port, floppy controller, XT-IDE controller, and even a mouse port are on the board.
The case is designed in 2 layers. The motherboard is at the bottom, then there's a metal sheet on top of that (aside from the part where the expansion slots go), and on top of that are the PSU and the floppy and harddrives.
So very solid construction, good shielding, and very clean and compact design overall.

The Philips is similar, but simpler. It doesn't have the graphics integrated on the board, and no mouse port. Everything else is on board (including XT-IDE). Unlike the Commodore, it does not use a single-chip solution for the chipset, but it uses the same discrete 82xx chips as the real IBM does.

All three systems still work fine, even the original HDDs in both the Commodore and the IBM 5160 (the Philips has no HDD).

sf78 wrote:

SRQ wrote:

The C64, the IBM PC, these things were all built to last and proved and refined the technology they used during their lifespan.

Well, I agree on the IBM but no so much on C64. They were pretty much thrown together using whatever parts were available at that time. Some had issues with mem chips, some had issues with other chips. I've had about a dozen C64's go through my hands over the years, and around 30-40% were somewhat faulty or had lower quality parts used unlike some other otherwise identical C64.

IBM's from the 80's and early 90's seem to last forever though. Obviously they have some week spots like floppy drives, but all the essential parts are still mostly functional after all this time. Many clones suffered a failed motherboard in just a few years and were scrapped. The clone 486's I've had around half has had a faulty motherboard.

My family's first computer was an IBM Aptivia with a 166mhz chip, biggest heap of trash ever owned, it went back to radio shack every 2-3 weeks for 2 years to get repaired, was never stable, never worked right and was the least realiable POS I've ever seen. So even IBM has its faults. My buddy got an IBM Aptivia with an Athlon 650 in it that likewise never really worked. MY guess is the Aptivia line was utter garbage.

SquallStrife wrote:

Jo22 wrote:

I don't really buy that "planned obsolescence" is some kind of malevolent conspiracy, I think it's just a response to our buying habits and market competition.

I strongly agree for the most part, but there are company's out they that use planned obsolescence, like apple and HP with their warranty's that cover each part differently. The place I work at mostly uses HP desktops and we had 10+ HDD's fails in HP all in ones, only one of witch failed under warranty, the rest failed about a week or 2 after the warranty was up. It's the same with apple motherboards in there laptops. A lot of times the motherboard fails shortly after the warranty is up.

M$ sort of does it too. Anything computer related with a support life span could be concentered planned obsolescence.
It's not truly planned obsolescence, but they are planning it's end, sort of.

But most people these days buy cheap junk that's prone to fail one way or another.

If you want stuff that is more reliable, you can always buy Enterprise class systems, they tend to be more reliable, just cost a premium

candle_86 wrote:

If you want stuff that is more reliable, you can always buy Enterprise class systems, they tend to be more reliable, just cost a premium

Ture, but they don't really cover the home user needs.
Bossiness class stuff seem more in place as a home system, but there only a small step up for the most part.

candle_86 wrote:

My family's first computer was an IBM Aptivia with a 166mhz chip, biggest heap of trash ever owned, it went back to radio shack every 2-3 weeks for 2 years to get repaired, was never stable, never worked right and was the least realiable POS I've ever seen. So even IBM has its faults. My buddy got an IBM Aptivia with an Athlon 650 in it that likewise never really worked. MY guess is the Aptivia line was utter garbage.

Quite so, I believe that after 386-era the quality went down pretty fast when they tried to move from business to consumer market.