Faulty Chips Might Just be 'Good Enough'

Ritalin16 writes "According to a Wired.com article, 'Consumer electronics could be a whole lot cheaper if chip manufacturers stopped throwing out all their defective chips, according to a researcher at the University of Southern California. Chip manufacturing is currently very wasteful. Between 20 percent and 50 percent of a manufacturer's total production is tossed or recycled because the chips contain minor imperfections. Defects in just one of the millions of tiny gates on a processor can doom the entire chip. But USC professor Melvin Breuer believes the imperfections are often too small for humans to even notice, especially when the chips are to be used in video and sound applications.' But just in case you do end up with a dead chip, here is a guide to making a CPU keychain."

Already commonplace with RAM chips

[no parity]

And for a long time so. "Audio RAM" is the euphemism.

Old?

[Borgschulze]

Didn't I read this a few days ago... seems old... Like I already read it...

Caveats

[karvind]

Testing ICs is an exponentially hard problem these days. One-third of the cost is devoted to it. Thus it may be a good idea to test the chip for only the applications it is needed (in some restrictive environments) and if it passes, it can still be deployed. It will ease some of the economic hammer on the manufacturing these days.

Xilinx offer EasyPath [xilinx.com] option by testing for a customer-specific application. Customers use EasyPath customer specific FPGAs to achieve lower unit costs for volume production once they know their design is fixed and no longer requires the full programmability of an FPGA.

Faulty Chips

[p0rnking]

I'm sure I read something, a long long time ago, that mentioned that Celerons were "faulty" versions of the Pentiums (and a comparison was made that the Durons were made as Durons, and weren't chips that were taken out of the garbage bins)

Ati and nvidia do this already...

[mobiux]

Usually thier LE and SE models have certain branches and pipelines already disabled. Usually these disabled pipelines are damaged in some way.

Re:Faulty Chips

[Anonymous Coward]

A Celeron was a Pentium III with a part of bad cache. The half with the bad section was marked as such, which is why the Celeron always had half the cache of a P3. They also ran them as much slower bus/core speeds. I'm not sure what the newer "P4" celerons are though. Probably the same deal.

Re:oh yes, by all means

[slughead]

My friends work in a warehouse where they resell Compaq and HP parts to companies.

They mainly sell old stuff, almost all of it's used.

There's this company that they currently get most of their inventory from, let's call them company X.

Company X sells used parts too, they just do rigorous testing before they send them to customers, so a lot of it is marked "defective".

When company X marks something "defective", they pay to have it shipped to my friends' company. It's actually cheaper to do that than to recycle the parts, so my friends' company actually pays just a few dollars for a thousand pounds of equipment.

My friends personally go through all of the components, and put them through the extensive refurbishing process of blowing the dust off and inserting them into static bags.

They test it "good enough".. which entails making sure the computer boots up with that RAM and CPU. Maybe a 1 minute memory test on occasion. All in all, about 10% of everything they send out is worthless, and will be sent back by the customer in a week.

Re:Already commonplace with RAM chips

[gl4ss]

and in counterfeit ram ;)

don't forget that.

but the real reason for disposal i think is that throwing away at that early saves money from the manufacturers, like, it's much cheaper to throw away one chip than to throw away a tv that doesn't work good enough to be sold.

however.. what would be the good solution? maybe build the chips redundantly so that it wouldn't matter if one gate didn't work?

Re:i486 SX vs DX?

[bulliver]

I remembered reading something like that so I dug out an old book of mine, "Upgrading and Repairing PCs" by Scott Mueller (2000):

The 486SX chip is more a marketing quirk than new technology. Early versions of the 486SX chip actually were DX chips that showed defects in the math-coprocessor section. Instead of being scrapped, the chips were packaged with the FPU section disabled and sold as SX chips.

Re:This is completely bogus.

[blanktek]

Actually he is a professor of electrical engineering systems. http://poisson.usc.edu/Breuer.html [usc.edu] But I think there is a lot of misunderstanding here about what is trying to be done. And it doesn't have to do with killing your RAM and your Counterstrike game.

There is another article here with some extra details. http://www.isa.org/PrinterTemplate.cfm?Template=/C ontentManagement/HTMLDisplay.cfm&ContentID=42102&F useFlag=1 [isa.org] I supposed what he is doing is trying to devise NEW methods to allow chips to work properly if they have errors. That is why he is getting the big grant money. For example in data transmission if you miss a bit it can be filled in with parity checking. I am of course guessing that it could be done this way. But the point is that it is not some conspiracy to trick you into buying crappy videocards. Firms know very well that the market will prevent that or they don't get to produce.

Re:Not quite

[Formica]

This is a fair amount of research going on for this; it's known as adiabatic logic [google.com].

Here's a short paper on how it's clocked:
Charge Recycling Clocking for Adiatbatic Style Logic [berkeley.edu]

Formica

Audible distortion and metadata

[tepples]

Second, even if all the bits of the sample are wrong, an answering machine probably samples at 8k Hz. If one sample has the wrong value, then the pop will be 0.125 milliseconds long, so not really that bad.

A single sample error will sound like the click in this wave [jk0.org]. But many digital answering machines use lossy compression optimized for the periodic sound of the human voice. A bit error in one of those may spread out over a whole speech packet, producing audible pops like in this wave [jk0.org].

In addition, even if the audio storage is lossy, there would need to be either a second certified defect-free part to hold metadata where in memory each message starts and ends, or an error-correcting code applied to the metadata.

Re:Sounds like Radio Shack parts

[Anonymous Coward]

There are good places to buy parts from, just probably not a store down the street:

• Digi-Key [digikey.com] (USA and Canada). Flat cheap shipping, sell most things in small quantities.
• Mouser [mouser.com].
• Jameco [jameco.com], although Jameco have crappy shipping to canada.
• Newark [newarkinone.com].

As well, there are a few off-axis surplus places (allelectronics.com, for example) that have super deals on things compared to the big suppliers, but less selection. Do you know a good surplus place? Add it to this thread!

Re:Ati and nvidia do this already...

[wannasleep]

These kind of techniques are very popular in memories. As a matter of fact, the "virgin yield" of a memory is close to zero. Because memories contain so many million devices that the chance of having a broken one is high. To overcome this problem, redundancy is added, i.e. some more rows and colums are fabricated. Then, the defective rows/colums can be disabled and the redundant ones are swapped in. The yield after this operation is called "repaired yield". There is a delicate trade-off with redundancy: if there is too much, the chip is too big and starts consuming unecessary area on the wafer. If it is too low, then your yield is low as well.

Re:Sounds like Radio Shack parts

[NormalVisual]

I could think of a couple of things to try - balanced signal lines (more expensive), or filtering. If all the PA is being used for is voice and/or crappy Muzak-type stuff, you should be able to safely cut off everything under about 200 Hz with a simple RC circuit.

Re:Already commonplace with RAM chips

[Anonymous Coward]

exactly this is already beeing done on RAM chips from nearly all big manufacturers. You can blow some fuses and viola a 265M chip becomes a 128M and has one address bit less.
also all DRAMs have a big amount of redundancy on their dies, that is activated instead of defective cells after testing. you can be sure that right after wafer processing not a single DRAM chip build with current technology has all cells perfectly working. this is only accomplished by replacing defective cells by redundancy.