Now I'm lost. If ringing were potentially a problem with using polymer caps instead of electrolytic due to lower ESR, and now you say that the ringing is mostly resonated by the ceramic bypass caps, I would guess(tm) that the ceramics would be more important and the only problem with comparatively higher ESR electrolytics is the head produced by the ripple current (I^2*R).
That sounds like the older Mac Mini's, which are plastered with MLCC caps on the bottom of the motherboard. These MLCC caps are difficult to find which one's are shorted, and even more difficult to replace since hitting one with a soldering iron will produce enough of a thermal shock to crack the ceramic. I have to use solder paste and a heat gun to work with them. If your Mac Mini won't turn on, that's the likely problem:
I don't see much of those on PC desktop motherboards. However, they're common on laptops, Chromebooks, ITX, micro ITX, mini ITX, nano ITX, pico ITX, and such. It's somewhat of a tradeoff. The boards that do NOT use MLCC caps, tend to use electrolytics that are susceptible to bad-cap type problems. Those that don't have room for cheap electrolytics, use MLCC and other low inductance packages, which are susceptible to soldering and rework problem.
Moral: You can't win.
These daze, the +12v supply is only used for running fans and a few odd floppy/CD/DVD/etc drives. Without those, the typical computah would be quite happy with only a +5V and possibly a +3.3V power supply.
ATX Power Supply Spec 2.4 says 50 mv p-p for the 5v lines and 120 mv p-p for the +12 lines. However, that's not the spec for the CPU VDD line, which wants about 25mv pip:
The lowest voltage aluminum electrolytic I could find is 6.3V. (There may be lower voltage ratings, but I couldn't find any among the aluminum electrolytics on Digikey). However, there are porous carbon electrode caps, and super capacitors with 0.5V ratings. I don't think the computer industry will pay the price for super caps. Conventional wisdom is that both aluminum and polymer electrolytics need to be derated in order to operate properly at high temperatures. I wouldn't run them at rated voltage unless they were properly cooled.
Yeah, I have some of those monsters back in the closest somewhere. The extra power connector was because the PCI or PCI-e bus connectors could not handle the obscene amount of current required by those video cards. The card needed external power to work. That's still a problem with video cards that burn several hundred watts. Various video cards: Scroll to the bottom of the list for the cards that require 450 watt and cost $1,200.
Not exactly. The minimum the ESR meter could do was 0.03 ohms which is reflected in some of the flat line ESR graphs.
The original commentary associated with the test is at: Start at the top with my posting.
The temp versus ESR curves did vary substantially, but I made no attempt to attribute the variations to any particular characteristic (or manufacturer). Yet another project.
Yes, but I have a genuine lame excuse. If you look at the PCB silk screen markings, you'll see that the white stripe marker and arrow are actually the + polarity, instead of - as is marked on the capacitors and on every other motherboard except for these. I just matched up the white stripes, which was wrong. Also, please pretend not to notice that I accidentally installed 85C caps instead of 105C.