 
|
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 |
81.159.134.211
In Reply to: RE: cMP - the open source high-end Memory Player posted by cics on December 30, 2007 at 05:42:01
Hi guys, I going to replace switched regulators on MoBo . First done, with good result.
Strangely for me, current on 1.85 volt line is 2.1 A , but memory KVR400D2N3/512
should take only 0.8Amp.!
MoBo +5 volt line, has a current draw now 3.7Amp., 1 Amp. more . LT1083 turn to heat about 6 WA ( 2.1A * 3,15V ) ,so I install heat sink outside, on bottom extension,.
to my low-profile computer case. Temporaly use Intell CPU sink.
Oscilloscope show excelent result, will send picture.
Many Thanks to JackWong for his pioneering!
Gene.
PS. Picture showing my old MoBo .
Follow Ups:
The L should be removed as it may create a current loop via the MOSFETS.
Jack , I did remove " L" , same consumption on +5.0 v MoBo.
Could you please take look on picture and comment on purpose of ( ? K3918 ) , for
me, look like, it connected in opposite direction.
Thanks, Gene.
Hi Gene,
The PWM controller (the small IC on the downside of 2R0 ) does not provide enough current, so a set of MOSFET driver (K3918s on your MB) is required. Each power phase uses a driver IC, two MOSFETs and one choke (2R0 in this case).
You are pretty much done by cutting both the legs of the K3918s and desoldering 2R0. The current consumption should be a liitle bit more than before, as linear regulators are less efficient than swtiching ones. Note that it may need to supply current to the RAM and other components.
I believe the problem you've got may lies with (a) improper grounding connection for LT1083, and/or (b) inappropriate input point taken from the MB.
(a) You have two ground connections, one would suffice, better still the ground of the LT1083 should be connected to the main ground of the ATX port of the MB.
(b) I'm not sure the input of the LT1083 is correctly connected on the circuit you provided. For me, I would connect the input to the ATX 5V (or 3.3V for less heat disspation), or use a seperate linear power supply for better isolation.
You may try to lower the supply voltage to 1.8V or below for better SQ.
Good luck with your attempts.
Jack
Hi Jack,
Thanks for yours schematic and explanation .
Total power consumption on 5 volt MoBo line , did increase actually on power
dissipation on LT1083 - 6 WA , 3 volt between in -out needed for LT1083 for good
performance . In my cMP 3.3 volt line feeded by small battery , so I need feed
from 5 volt.
Will try take from MoBo 5 volt , and reduce to 1.75 v .
Still waiting for delivery rest of LT1083 !
Gene.
Jack I appreciate that you are probably busy and can't post here all the time and you probably get a lot of queries on pc stuff but I really learn a lot when I read your posts. Are there any other forums on which you post so one can follow your various projects?
we missed you!
I never thought 15 words could make me so happy!
Jack, very nice to see you back here. I hope whatever has kept you away has abated and we'll get to share more of your wisdom and daring.
Anything you've done that's worth sharing?
Greg in Mississippi
Everything matters!
Thanks Jack, tomorrow will post some schema for this circut, as I trace it , comment will be usefull.
Could you please tell me your view on using same approach for CPU ?
Tnanks, Gene.
Most excellent sir!
Thanks for sharing. Send completed pix when you can.
Greg in Mississippi
Everything matters!
Greg, Thanks, but do not expect nice pix . I lost count of my revision on this chassis !
Gene.
First, I just unsoldered only one K3918 , upper on picture and it work , oscilloscope showed only linear ( no digital spike) .Maybe posible even without unsoldering , just a feed 1.85v and K3918 will be stoped ?
Gene.
Hi Gene! This is very interesting. But what about sound changing.
http://cmp2-mihaylov.narod.ru/
Hi Mihaylov,
I think bass better , more controlled at least that is what on my mind , after first 10 second listening , clearer. I am very happy with improvement , did ordered 6 more
LT1083 after listening, now I realise , there no need so many. Did find one more " dirty"
spot on MoBo plus for CPU , but it will be chalenge. Still not figured out about CPU , help will be usefull.
Picture show as it was before.
Gene.
This picture show what I got now. It is in fact, mostly internal noise of my Oscilloscope
on this range with some impedance, reading show 1-2mv , limit of my PICO.
Gene.
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 ]
