Bitcoin Forum
November 16, 2024, 08:46:04 PM *
News: Check out the artwork 1Dq created to commemorate this forum's 15th anniversary
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: MSI 5850 (ref) Volterra problem  (Read 1679 times)
2GOOD (OP)
Hero Member
*****
Offline Offline

Activity: 547
Merit: 531


First bits: 12good


View Profile WWW
December 13, 2012, 09:25:04 PM
Last edit: December 13, 2012, 10:36:18 PM by 2GOOD
 #1

Hi,
I bought this card as faulty several months ago, First run - it starts, but there was a very strange noise from the mobo's speaker. Then I noticed that one of the voltera's Capacitors was blown away (see pic1). I had to do a bridge for re soldering the new Capacitor as the pcb path was missing anyway this was done and now the card works without the creepy speaker noise from the mobo. The problem is that when you load the card VRM 1 starts overheating to extreme while VRM 2 and 3 stays cold, but eventuaaly gets heated by the PCB and cooling plate. At stock freq/voltage VRM gets above 140C in no time and card shutdown itself. I managed to get it to stable values for mining at the lowest 0.875 V and 430Mhz (177MH/s  Grin) the VRM stays around 109C and that way the card is stable and mining for more than 4 months without a glitch. It can go higher freq, but the VRM gets too hot, more than ~120C

Quote
Device [1]: Cypress [Radeon HD 5800 Series]
        Current core voltage: 0.8750 V
        Presets: 0.8750 / 0.8750 / 0.8750 / 0.8750 V
        Core power draw: 68.81 A (60.21 W)
        VRM temperatures: 109 / 58 / 50 C

IMO there are two possible scenarios here
1. VRM 2 and 3 are not working at all and the power is driven by VRM 1 which can't handle the full power requirements of the chip.
2. VRM 1 is shorted or just faulty and thats why it overheats.

What is your oppinion, now that the GPU mining is going to end I have more time to experiment with this.





Not actual pictures

Edit: I remember now that same Capacitor (pic1) was desoldered for VRM 3 too.


P.S.
===
This is not a cooling problem Wink

ssateneth
Legendary
*
Offline Offline

Activity: 1344
Merit: 1004



View Profile
December 14, 2012, 01:27:58 PM
 #2

68 amps draw for 0.875v and 400mhz core is an EXTREME amount... my 5970 cores only draw 53 amps with 1v and 755 core.

2GOOD (OP)
Hero Member
*****
Offline Offline

Activity: 547
Merit: 531


First bits: 12good


View Profile WWW
December 14, 2012, 07:16:32 PM
Last edit: December 16, 2012, 12:52:55 AM by 2GOOD
 #3

68 amps draw for 0.875v and 400mhz core is an EXTREME amount... my 5970 cores only draw 53 amps with 1v and 755 core.

Good point, so maybe the problem is N.2 i.e. VRM 1 is shorted or faulty, maybe I'll try to isolate it and see what happens. Does anybody know how to do that, to cut power supply to that VT1157SF

Edit: after some more serching I found almost the same story:
Quote
Hello
I have problem with my Radeon 5970. On the master GPU have VDDC Phase #2 Temperature very high, but all other temp. is good. When load GPU to 99% temp. rise more 120 and crash my system. VDDC Current is ~ 80A load and ~12A idle. On the slave GPU all is good. VDDC Current is ~ 54A load, and temp. is good. I change one VT1157SF from PCB (U72) but this not fix my problem. When I remove (U72) card work, but available only one GPU (slave), and not crash when load 99%. Anyone known where is located sensor "VDDS Phase #2" for master GPU? Maybe Master GPU is broken? I need fix my card but can't locate where is the problem. If anyone have information which parts more I need check please help.

Regards
Sakura



Remember remember the 5th of November
Legendary
*
Offline Offline

Activity: 1862
Merit: 1011

Reverse engineer from time to time


View Profile
December 22, 2012, 11:11:54 PM
 #4

My 5850 is pretty much the same. High VRM temps, high AMPs whereas the 5870 is waaay below these temps/amps.

BTC:1AiCRMxgf1ptVQwx6hDuKMu4f7F27QmJC2
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!