 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
98.247.145.109
In Reply to: RE: You're on right path. posted by Tony Lauck on October 03, 2013 at 08:38:06
It makes sense to me power would be going down with voltage and for a cmos circuit the capacitor energy storage first order approximation is a good choice to model switching energy.
Your post helps me pull my head out more a little and realize there is probably no way I have a good argument about whether its generally going to affect E or B fields strength more than the other though. heheh
dv/dt's do go down a hair with lower vcore but at the same time currents shrunk too since as you point out the energy to accomplish the state switching is proportional with v^2, ie less current also means smaller di/dt's.
In other words difficult to say whether a reduction in potential for common mode or diff mode noise problems. sorry.
I don't have a kill a watt for some reason and am generally not really currently too worried about pc emi and so probably wont be messing with my vcore. Figured maybe someone whos tried it may have seen CPU temps come down from using less power.
Follow Ups:
As I understand it, as vcore is lowered the noise margins on logic signals inside the chip degrade. Eventually, the chance of data corruption becomes greater and greater. When it gets high enough then the computer will become unreliable. If you are lucky, there will just be slight glitches that do no damage, but there can be blue screens, etc... Worse, if data corruption happens when the operating system is running there can be file system corruption.
I would mess around only on a system dedicated to audio playback. I would assume that the computer is going to become unreliable at some point, even if it runs long enough to enjoy a listening session. In particular, I would assume that at any moment it will corrupt or "eat" programs and data files, including software and music library that are under its control. Since I also use my listening computer to do transcriptions and restoration of old analog tapes, this rules out my computer for this kind of hardware tweaking. Having worked with computers since 1960, I am quite familiar with unreliable machines. I have little interest revisiting data corruption scenarios from my past.
A kill-o-watt is a nice toy and might end up paying for itself if you have a large power bill. Mine already has.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
tony, we have a local version of the kill-o-watt which i use to check power consumption on my gear. those high end video cards really eat a large amount of electricity when in full load !
it's also the reason i sold 3 7850s, 1 7790 and 1 7730. lowering the vcore, disabling cores and lowering clock speed all affect this. your thing about data corruption due to too low a vcore, will this apply even if stress tested and found completely stable ?
I don't know how the error rate jumps with voltage. Others may know, but it may depend critically on the chip design. If it works like falling off a cliff, then one can find the limits easily and back off. However, if it is more gradual then one may go from one glitch a minute to one glitch an hour to one glitch a day, etc... and if no errors have been observed for a day one can not safely conclude that all is well. There may be some literature available, but it's been almost 20 years since I worried about eliminating data corruption. There are various causes for data corruption: thermal noise, shot noise, deterministic noise, EMI, alpha particles, ...
Back in the 80's those who were serious about data integrity were seeking corruption rates of less than 10^(-18) per bit moved, but with much faster machines and so many more chances to roll the dice this is no longer good enough. Of course, if the bits being moved are "just audio" then perhaps the occasional error doesn't matter, assuming one subscribes to the Sony/Philip's mentality when the CD was designed.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
fmak. i find winddows 8 a bit too digital for my taste. and you have not heard my modded windows 7 :).
my only quirk is i want to drop down the noise floor in my modded win 7 without removing internet functions. the in between instruments is good and there is air and space but the "blackground" is not as black.
gear in use is a xonar stx as preamp driving a bryston 2b and lsa standard speakers. (yes i ditched the bryston bda2. it's now in storage on a floor board next to me).
i just tried windows 8... it sounds good ! i don't know why before it sounded a digital. must have been the amp. rega *blech*
back with windows 7... i just wondered why they made usb dacs or dacs of any sort. they always sounded so dismal. now with windows 8, i don't wonder any more.
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
