Jonnyc55 wrote: ↑14 Feb 2024, 13:55
Another thing that changes over time, especially since windows 10, is that a lot of their updates will probably include more rampant data mining technology. The very act of monitoring all aspects of a process is a hindrance to that process.
Good point, but I've tried some things in this regard before. For example, I created a custom Windows ISO with ntlite, stripping it of all telemetry/data collection functionality. I've tried the LTSB versions of Windows 10, I've tried Windows 7, and I've tried some custom made Windows images such as AtlasOS or KirbyOS. Nothing changed in terms of desync problems at all.
LOJesmire wrote: ↑14 Feb 2024, 14:02
Often when you just flip the switch of the power strip the pc ist not fully disconnected from power. Obviously you can’t turn it on, still the pc is kind of connected to power line.
Got it. Some power strips have singular polarity switches. They only break the circuit at one place. Depending on which direction you inserted the plug in the outlet, it will either break the neutral or the live side of the circuit. If you happen to break the neutral, the computer still sees 240V coming in, but you can't turn it on because there is no closed circuit.
I've looked this up before and the power strip I use uses a dual polarity switch. It breaks the circuit at both points.
LOJesmire wrote: ↑14 Feb 2024, 14:02
I’d be surprise if the feeling is still bad, as I have otherwise the same symptoms as you.
As part of my internship I've been to training courses far away from home where I had to stay for 2+ weeks at a time. While I was gone from home, I disconnected the plug from the wall. And of course when I came back and tried to play, it was immediately bad. As I said before, I don't get the "creeping back in" effect of this problem. It's there all the time. Nothing I do changes this.
LOJesmire wrote: ↑14 Feb 2024, 14:19
Capacitors do not fully charge when the pc is turned on. They only need a certain level. But their current charge varies depending on the circuitry and operating status of the PC.
Yes, absolutely. They aren't fully charged all the time, and keeping them fully charged (meaning that you wouldn't allow them to discharge) would bypass their functionality. But i think it's safe to assume that motherboard manufacturers don't like to overspend on components so they'll probably use capacitors which are just about "large" enough to do the job. Which probably means that they'll become near-fully charged when the VRM is in it's "on" phase, and near-fully discharge when the VRM is "off".
For caps, you have to think in extremely tiny time dimensions, as basically any VRM on a modern motherboard has a default switching frequency of either 300KHz or 500KHz. These caps charge and discharge millions of times within a second. Anything hooked up to a cap (e.g. a CPU, GPU, etc) are low-resistance loads, enabling them to discharge incredibly fast. Caps aren't a thing that charges once you turn the PC on and then hold their charge for 2 hours, slowly discharging, then need to be charged again. They absolutely can hold their charge for that long, but usually they have low-resistance loads hooked up to them, meaning that they discharge within very, very short time frames. You can discharge any cap virtually instantly by shorting its output together. The discharge time of a cap is directly dependent on the resistance between it, so a dead short (<1 Ohm) discharges it very, very fast. And a computer component, from a caps point of view, is a very low resistance load, roughly about 0.1 Ohm for a CPU at idle.
LOJesmire wrote: ↑14 Feb 2024, 14:19
So maybe as long as they are not completely empty and get new “clean” power or the remaining charge falls below a certain level the issue persists.
Not quite sure what you mean by this.
IMO by the time any capacitor comes into play they do receive perfectly clean power. The only time I could imagine a cap which doesn't receive clean power would be the bulk cap of the PSU. Although, if you're not using some garbage tier power supply, it should feature plenty of filtration and EMI protection even before any electricity reaches the bulk cap. PSUs from reputable brands have lots of filtration which should be able to account for every possible form of "dirty" electricity there is. By the time the voltage actually gets to the components, it has been through at least the bulk cap, multiple stages of caps in the PSU, caps on the motherboards VRM input side, caps on the VRM output side and caps around the CPU/GPU die itself. Even if some form of "dirty electricity" crept in from the mains, it
should be filtered out at some point before reaching the components. I like to imagine it as a swiss cheese. Even if "dirty" power could enter through the filtration, it would be caught by the bulk cap. If not, it's caught by the smoothing caps, etc... The odds for all of these "holes" in the swiss cheese to line up perfectly so that some form of "dirty power" can reach all the way to a critical component seem astronomical.
I have a hard time believing that caps and/or electricity really is the main culprit. But this is just my 2 cents, and obviously there is complex electrical engineering involved which goes beyond my general understanding. I'm still hopeful about somebody figuring all of this out eventually.
Thanks for all the contributions, people.