rasz_pl wrote on 2022-05-30, 04:17:
Socket3 wrote on 2022-05-29, 18:17:
sram chips, some older dram chips, some chipsets
Semiconductors usually dont die on their own unpowered in storage. They are made from inert materials. What mostly happens is people pulling stuff that was put in storage because it was broken in the first place, or damage happens while powering on due to other factors.
SRAM and old DRAM is know to go bad. There are entire threads on VCFED about this - particularly DRAM and SRAM chips used on XT class machines and older.
While you are right about semiconductors being mostly inert, the chips themselves are made of many other materials like plastics, organic substrates and various metals. Any of these on their own or in contact with each other might degrade or interact with each other and/or air over long periods of time (in some cases 40+) years, causing the aforementioned chips to go bad.
Traces of certain chemicals used in the manufacturing process may have also become embedded inside the chip packages, chemicals that could theoretically interact with other components causing them to go bad over long periods of time. I've seen examples of PCBs themselves degrading over time - like one of my LS486e mainboards, witch due to manufacturing defects had air bubbles trapped inside the PCB around the cache chips. They look like air bubbles under a magnifying glass, but x-rays reveled tiny cylindrical cavities, witch indicates that those "bubbles" contained trace amounts chemicals that reacted very slowly with surrounding material - the substrate itself as well as the embedded copper, causing shorts and broken traces and rendering the L2 cache inoperable.
I've had this particular board for quite a while now - over 10 years, and over the las 4 I've noticed strange discoloration on the back of the motherboard under the L2 cache sockets. I didn't pay any attention to it until last year when the board became unstable and eventually stopped detecting L2 cache altogether.
Semiconductors and semiconductor packaging technologies are a relatively new, and we are only now seeing how very long periods of time affects their integrity and functionality.
As for "was put in storage because it was broken in the first place" this may be true for some, but I'm very meticulous when it comes to storing my hardware. Everything is categories, stored in anti-static bags of the appropriate size and in case of long term storage even accompanied by silica gel packages. Defective components are stored in separatly, and are labeled with date tested, confirmed or possible defect and probability of repair as well as a parts list required for said repair. Components witch I deem unsalvageable are stored in big plastic bins, marked with a permanent marker and either sent off to be recycled (I trade them for other parts at e-waste centers) or used as parts donors (ICs, sockets, MOSFETS, Diodes, capacitors, etc). Despite all these precautions I've had loads of known good hardware go bad in storage.
The worse offenders are electrolytic capacitors - be it SMD or trough hole types. Just now I pulled one of my Abit AN7 boards from storage and noticed it has 2 bulging 3300uf capacitors, one just south of the ram slots and another near the southbridge, as well as two caved in capacitors - both near the CPU socket. The board powers on normally but is unstable in windows. Same happened to an Asus P4-P800-E. Both these boards were stored in their original boxes, in anti-static bags, together with their manuals and i/o plates. I have pictures of the boards before placing them in storage so there's no mistaking the fact that the capacitors bulged while the boards were stored.
Second is EPROM and EEPROMs. For UV-EPROMs it's easy to imagine that the label on top of the glass window used for UV erasing could become damaged or was not completely opaque, letting some amount of light in witch slowly over very long periods of time could have corrupted the contents.... but I store my boards either in individual boxes, metal cabinets with drawers or black plastic boxes with lids... there's no way light could have gotten in. And what about electrically erasable programmable ROMs and Flash EPROMs? I've had a few of these fail in storage - particularly on 2003-2004 Asus and 2001-2002 Abit motherboards using PLCC chips.
As for "or damage happens while powering on", I've only experienced this with Tantalum Capacitors, MOSFETS and some diodes. Sure, I've had electrolytic caps blow up in my face at power on, but that was due to the fact that they were shorted - probably because they "dried out".