Is there any cpu line that uses higher quality metals?

[pazsion]

I've been recycling cpus for a bit, and i noticed that most, even new ones.. are gold plated aluminum or zinc.. or worse gold/copper alloy coated aluminum..

The internals are thin hairlike copper coated aluminum..

no wonder they wear out so quick..

Does anyone have experiences otherwise? And what was it?

 

[etk]

internals (boundry wires) are generally gold...die is silicon. Wear out? boundry wires don't really wear out then way interconnets do (electromigration) sometimes vibration will get them though.

 

[Herr_Koos]

Wear out? I've never heard of a CPU "wearing out", much less had it happen to me. Please explain?

 

[IH8U]

Internals are typically gold or platinum (or other metals with low/no impedence), no aluminum there. The Aluminum is in the heat sheild/spreader, the pins are gold plated aluminum or steel. The only "wear" on internals may be due to burn-out from heat, or voltage (same result tho) frying the interconnects (think of solder, and high heat).

 

[geekapproved]

I've never seen a cpu wear out either. No idea what your talking about.

Yes there is a new line of 24K Gold cpu's out. They are $11,500 each and only 1ghz. If you recycle one you get $3.

You would be better off robbing graves for teeth.

 

[etk]

Transistors do wear out (hot carrier injection, bias temp instability, or shorts) but I don't think that is what he was getting at...seeing as the OP has not come back, I would say to just close the thread.

 

[pazsion]

wow, so an op is expected to carry out a live chat in a forum post? XD etk thats quite rude..

I have one of those on my site if you like, but my netbook is down for maintence.. so i'm not in there while gameing..

voltage sent through any material wears out that material..through electro migration.

Or electricity's ability to atomize the metal as it passes through it, electromigraton also takes into account metal transfer as electricy jumps points like interconnect, transistors, capacitors and transformers /resistors.

aluminum, zinc are absolutely terrible for this. 2-5 years of use.. visible pits in the material can be seen without magnification.

But yes that cpu would still function, but data loss, spikes and low voltage..even shorts. would persist. with excessive heat due to the increase in resistence to the data flow. the motherboard should also be considered.. which is mostly aluminum, zinc and copper. also usually a lower grade material then say ram and cpu. I don't know why, it's just an observation..that why i'm asking here..

iridium, platinum tho expensive, would be best. 2nd would be silver, then copper/gold alloys. Gold by itself although good at resisting electromigration and corrosion.. isn't as good as copper or silver for condutivity and retention of that energy as it travels, as say copper even.

but copper wears faster then gold..

11,000 dollars?

what is it?

Even considering all this.. and organic, fiber optic materials... durability and resistence to moisture.. humidity corrosive things.. is my basic concern. And ability to take extreame voltages and speeds.

prices, with less and an oz of rare/metal material used in cpus.. is probably not governed by materials used, but more targeted marketing.. manufactureing and pocket lining..

but that of course is my opinion.

 

[jsc]

Well, everyone is entitled to his own opinion.

Not counting a random failure, a CPU operated inside its voltage and thermal limits will, in all likelihood, become technically obsolete before before it "wears" out due to electromigration or any of the other failure modes you mention.

The only time electromigration becomes a practical factor in failure modes is when you overvolt way past the manufacturer's recommendations.

 

[fazers_on_stun]

^ Clearly a better metal for preventing electromigration would be Unobtanium ..

 

[pazsion]

lol, yea for most people who are enthusiast's or power users, one year old pc's are "obsolete"

If it doesnt play some new game, or use some uber software at light speed.. these people consider that pc junk.. it's a terrible mentality, and hurts the industry. The leading reason better materials are not used, and when they are its more then expensive..the idea is to turn a profit before these people crap on it, and call it obsolete and not buy it, and persuade others not to.

=/ nuts i was hopeing someone here would have word on organics being used, in "consumer grade" pc's. Like in olcd's, its rumored that this is being tested in quantum cpu's with some good results!

 

[jimmysmitty]

Obsolete and worn out are very different terms. Obsolete is well, within 6 months to a year with PC technology. Worn out is like a HDD which tend to die due to mechanics. Most CPUs these days will last longer than they are useful.

As for the metals, they have advanced. Intel is even using Hafnium for their HK/MG technology.

 

[ElMoIsEviL]

Electronics are engineered to fail (or should I say engineered to last up to the end of their warranty periods).

3 Years Warranty on a CPU indicates that it is expected to last at least 3 yrs. Anything above that goes beyond its life expectancy.

 

[etk]

wow, so an op is expected to carry out a live chat in a forum post? XD etk thats quite rude..

but that of course is my opinion.

Apologies, it just seemed like the kind of thread that would be abandoned, but it appears I was mistaken.


On topic:
CPU's at stock voltage will last far beyond their warranty period. 20+ years 24/7 use would not be an unreasonable expectation at reasonable temperatures and voltages. It's just a probability curve, like anything else, and elevated temperatures and voltages increase the chance of failure. If increased reliability is desired, undervolting combined with a good cooling system would be the way to go, along with running at a slower clock frequency, since many problems with chips will increase their switching time, rather than preventing them from switching at all. Any super important device would be designed with a fault tolerant parallel architecture anyway, regardless of how reliable any single device is.

 

[Dekasav]

In my years of PC work, I've never seen a CPU fail, but I have seen failures out of PSUs, RAM, and motherboards. I think CPU failure is the last worry I have on my PC.

 

[jimmysmitty]

In my years of PC work, I've never seen a CPU fail, but I have seen failures out of PSUs, RAM, and motherboards. I think CPU failure is the last worry I have on my PC.

CPUs failing are very rare. Normally they have to have been caused by humans or another component like a PSU blowing or a mobo catching fire.

RAM dies when low quality memory chips are used. Super Talent is a good example. Super Talent uses multiple vendors for their memory chips. I have found some with Hynix, crap that normally goes into OEM PCs, and some from Samsung/Micron (the best used in Corsair and other high end memory). I have never had Corsair fail me.

Mobos also depend. If you buy a cheap one to save and put money into the CPU, you will have a fail come along. They use lower end caps and do bad solder jobs. Thats why I only buy high end Asus mobos.

PSUs, same as mobo and RAM. Cheap ones tend to use lower grade components and fans. Most PSUs that go bad are because the fan dies and so the parts over heat. I try to only sell Corsair CX430s. Haven't had one go bad yet and we have been using them for over a year in our builds.

As I say with PCs, you get what you pay for. Thats why I can't understand people who buy a $50 dollar mobo and ask why it dies.

 

[abdussamad]

Electronics are engineered to fail (or should I say engineered to last up to the end of their warranty periods).

3 Years Warranty on a CPU indicates that it is expected to last at least 3 yrs. Anything above that goes beyond its life expectancy.

No it means the company wants to limit its liability to 3 years. A cpu will easily last 30 years. CPUs don't fail.

 

[abdussamad]

As I say with PCs, you get what you pay for. Thats why I can't understand people who buy a $50 dollar mobo and ask why it dies.

Most computers sold on the planet have $50 mobos. And they all work fine day in and day out.

 

[archibael]

Electronics are engineered to fail (or should I say engineered to last up to the end of their warranty periods).

3 Years Warranty on a CPU indicates that it is expected to last at least 3 yrs. Anything above that goes beyond its life expectancy.

Not... exactly correct.

CPUs are engineered with quality in mind such that 99.99xxxx% will survive at least Y number of years of continuous use, where xxxx and Y are closely held secrets of any given company's Q&R department. I cannot tell you what either number is, but I can tell you that Y is rarely as low as the warranty.

Mortality is a statistical distribution, of course, so your device is likely to last beyond Y years, it's just that once you hit that age your chances start to go up significantly since your vendor probably didn't put too much effort into finding out what happens after that point.

As was pointed out upthread, running the device higher than its stock voltage accelerates most failure modes and you can thereby pull the tail end (or the "fail end", if you like) of the distribution in closer than Y years... but again, it's statistical. Your part could be overvolted for a long time and last Y+3 years, or it might drop to Y-2.

Only Q&R know... and they ain't telling.

 

[abdussamad]

Back to the topic at hand. I think the OP is using the word "CPU" as it is used by laymen. That is he's calling the system unit a CPU. In other words he's referring to the case. Computer cases are not what they used to be. They are much flimsier than in the past. I think its got to do with the high price of steel.

 

[jsc]

No. Actually, aluminum, although easier to work, is more expensive than steel. Steel is heavier. Compare an Antec 900 case with the similar sized CoolerMaster 922.

If aluminum was cheaper than steel, we'd be seeing a lot more aluminum car bodies on the road.

Most computers sold on the planet have $50 mobos. And they all work fine day in and day out.

And they are in commodity systems running largely undemanding applications. Among other things, there's a reason that they all have locked BIOS's.

Mobos also depend. If you buy a cheap one to save and put money into the CPU, you will have a fail come along. They use lower end caps and do bad solder jobs. Thats why I only buy high end Asus mobos.

That's why I buy Gigabyte motherboards and don't buy ECS or AsRock motherboards.

 

[pazsion]

Back to the topic at hand. I think the OP is using the word "CPU" as it is used by laymen. That is he's calling the system unit a CPU. In other words he's referring to the case. Computer cases are not what they used to be. They are much flimsier than in the past. I think its got to do with the high price of steel.

um, lol? no?

Also my intent was to find the best quality materials, available to anyone, not a particular group, or marketing label/target.

And not to scrap it.. I only scrap things that fail multiple tests and can not in any stretch of the imagination be repaired or re-used.

Yes a cpu is the longest surviving thing on a pc usually. regardless of overclocking, as long as it's cooled, And voltages are stable and within the margins. (insert clip of screaming p4's here)
But also keep in mind it will always work as long as electricity can get to where it's sent to. Without error correction, and other stuff... these pits in the metals of the cpu... would render them utterly useless. Just because it works, doesnt mean it's not worn down. Tho benchmarking alone doesnt always tell the whole story, or the condition of the cpu.. strange voltages and excessive heat are usually the best indicators of of a worn out cpu or any electrical device. If you wanted to, and you could set a voltmeter for the range needed, you could test each pin for resistence. But i've found very little documentation to refrence these to. And as yet i wouldn't recommend sending volts through a cpu, outside of normal operating.


Also the things i'm considering in higher quality, or more pure metals.. is system latency, and clarity of signal. If you do somehow turn the thing up..some things begin to mess up. If it were gold used instead of aluminum or copper.. You'd probably have a higher fault tolerance.

my experience with aluminum/copper and gold/copper metals, suggests this is true.. lower latencies, less artifacts, less ?lost code and strings? and various other issues.

If it were possible to have a 1ghz perform like a 6ghz..simply by useing better metals.. or other material.. we should do that. Likely the ability to use less power will follow.

i'ma google this gold intel chip..wish me luck XD

A good example of how crappy metals messes with quality of computeing... take a cpu that contains iron... and try to do anything with it.. like a Pentium mmx.. I hated these... and they recieved so many complaints.. i think they were pulled after about a year of service? replaced with a pentium proII mmx? where i live humidity was also high year round.. i found it turned the chips green and orange =/ but they worked