Here's an unrelated thought I had last night with regards to my Phat console. Rather than rip the NEC TOKINS off the board, has anyone every tried paralleling up capacitors (a smaller amount of capacitance of course) to the NEC TOKINS, boosting/topping up their capacitance rather than completely replacing it with new?
Yes, that's not how it works. People think of capacitors as a battery, discharging when there are voltage drops. Where capacitance is like how much energy they have stored. That's not a bad analogy, but it's only 1 of the 2 important jobs.
The second, most important job, is to provide a low impeedance path to groud for switching noise. When a voltage regulator switches on and off to provide the voltage, it does so with really fast rise times. There is a little bit of extra momentum and thus overshoot causing a bery brief voltage spike and dip. That rings and stabilizes at the new voltage. Then it drops back down before the next pulse. That overshoot is at the switching frequency, but the ringing it also produces extends into harmoic frequencies above the switching frequency. That switching ripple and noise has to be given a low impeedance path to ground, so that instead of entering the Processor and interfering with signal discrimination, it "bypasses" the processor and follows the path of least resistance.
Capacitors have an impeedance vs frequency curve that is not linear. At a certain frequency there is a sweet spot where it's impeedance to a specific frequency is super low. If you carfully selcet the capacitors so that your switching noise and harmonics line up with the lowest impeedance of your capacitor array, then they will filter the ripple and noise. Bypassing the processor to ground instead. That's called the combined frequency response curve of the filter.
The combined frequency response curve changes based on the caps chosen and their individual characteristics. Of you combine the wrong ones, it can lead to antireaonance peaks that produce noise spikes and make the filter worse, not better.
This is why you want to select the capacitors carefully to reduce the noise/ripple (attenuate) specific to your filter.
The tokins were attenuated to the filter. Adding polymer caps in parallel changes the freqency response with unknown effects that could be worse, even if it appears to work (which is does BTW).
I know somoeone that added two 470uf tantalum parasites and it worked. It failed about 2 years later requiring more caps. The problem is you're putting a band aid over cancerous skin lesion. That may stop the bleeding, but the cancer is still there, getting worse.
You need to replace the bad cap. And if one is bad, the rest were exposed to the same temperatures and time that caused the 1st to exceed it's enduance specification. They will all begin failing in the coming years.
So if you want to keep opening the console, placing strain on the BGA each time you have to break suction on the paste, flexing the motherboard, to solder on a couple more parasite band aids every few months to years, that's up to you. But I offer my reccomendation that you remove all the tokins and replace the filter array with one designed to best approximate the previous one. Only this time, having a brand new 2000hr @105C endurance rating. That way you only have to open the console once and risk damaging the BGA physically just that one time.
