Overclock.net › Forums › Intel › Intel CPUs › Did we ever come a conclusion on safe SB voltages?
New Posts  All Forums:Forum Nav:

Did we ever come a conclusion on safe SB voltages? - Page 8

post #71 of 93
The limit for Sandybridge was set at 1.52v by Intel.

The only reason know one "really" knew was because they couldn't understand how the volt limit could raise while nm lowered from 45 to 32. But Intel doesn't play around; they come up with some hard core chip technologies that would allow for such a possible feat.

I for one believe that the limit is 1.52v for the average chip. I think you would be fine as long as your chip never went above 1.5v.

If I could actually get anything stable even at 1.5v I would run it at that, but I can't seem to even accomplish that with this crappy chip.
post #72 of 93
I don't have much to add. The only thing I will say is my chip has went from being 4+ months dead stable 24/7 folding at 4.8Ghz/1.39-1.41v to now needing upwards of 1.43 to get the same OC stable. Even dropping down to 4.7 isn't even stable at 1.41v(the previous 4.8Ghz stable voltage) and it seems to continually need extra voltage as time goes by. This of course could be degradation or simply an unstable OC from the beginning, but I find it hard to believe such a long period of rock solid stability of 24/7 folding would not have show this instability much earlier.

I am now having to run at 4.6Ghz at 1.39-1.41v(haven't updated my sig), the same voltage that kept 4.8Ghz stable for several months. I have changed nothing during this time either, same BIOS, all other BIOS settings, etc same during all this. Not going to say what is causing it, but something is causing extra voltage being needed as time goes by.
post #73 of 93
Fitting this thread is around tonight... I had been folding on my 2600k at 1.495 for the last couple of weeks, 2 months @ 4.8 before that (1.46) and 3 months at 4.7 before that (1.39). I went to go put my 570 back in this evening, as well as update the bios and now my system will not boot into an os at anything higher than 4.3 and 1.39v. I have reverted back to the old bios to no avail. I don't know what happened, but either my cpu or motherboard is screwed up now.
 
TC folding
(7 items)
 
 
CPUMotherboardGraphicsRAM
2600K Asus Gene-z MSI 570 Ripjaws X Series 8GB 1600MHz 
Hard DriveCoolingOSPower
OCZ Vertex 3 XSPC Raystorm Block + 180mm Radiator Windows 7 Ultimate/Ubuntu 10.10 Corsair 750TX 
Case
Silverstone TJ08-E 
CPUMotherboardGraphicsRAM
2600K Sabertooth P67 ATI 4650 GSkill 2133 2x2GB 
CoolingOSCase
XSPC Rasa, RX240 Ubuntu 10.10 Fractal Midi 
  hide details  
Reply
 
TC folding
(7 items)
 
 
CPUMotherboardGraphicsRAM
2600K Asus Gene-z MSI 570 Ripjaws X Series 8GB 1600MHz 
Hard DriveCoolingOSPower
OCZ Vertex 3 XSPC Raystorm Block + 180mm Radiator Windows 7 Ultimate/Ubuntu 10.10 Corsair 750TX 
Case
Silverstone TJ08-E 
CPUMotherboardGraphicsRAM
2600K Sabertooth P67 ATI 4650 GSkill 2133 2x2GB 
CoolingOSCase
XSPC Rasa, RX240 Ubuntu 10.10 Fractal Midi 
  hide details  
Reply
post #74 of 93
Thread Starter 
Quote:
Originally Posted by JoshHuman View Post
Fitting this thread is around tonight... I had been folding on my 2600k at 1.495 for the last couple of weeks, 2 months @ 4.8 before that (1.46) and 3 months at 4.7 before that (1.39). I went to go put my 570 back in this evening, as well as update the bios and now my system will not boot into an os at anything higher than 4.3 and 1.39v. I have reverted back to the old bios to no avail. I don't know what happened, but either my cpu or motherboard is screwed up now.
I've seen similar things happen during a BIOS update. I'd recommend clearing the BIOS via the mobo pins/button, and then trying to re-apply your OC. That usually works for me. It seems that sometimes the OC/BIOS update messes with something and a reset fixes it up.

My BIOS update process:

1. "Hard" reset BIOS.
2. Run Memtest for a few minutes to see if everything is still hunky dory.
3. Update BIOS using DOS/in-BIOS option (never Windows).
4. Boot computer, and make sure BIOS is updated.
5. Hard reset BIOS.
6. Run Memtest to see if everything is still hunky dory.
7. Re-apply OC.
8. ????
9. Profit!

EDIT: One more thing. I also look at the in-BIOS readings of different voltages to see how "different" they are from my setting. It seems that sometimes BIOS updates fix the vDroop/Drop that occurs for different measurements, or artificially adds an offset to the "set" voltage to make it appear that way. If my re-applied OC volts show a different in-BIOS reading, then I'll adjust to compensate.
Edited by Nick2253 - 10/31/11 at 9:37pm
post #75 of 93
Quote:
Originally Posted by Nick2253 View Post
I've seen similar things happen during a BIOS update. I'd recommend clearing the BIOS via the mobo pins/button, and then trying to re-apply your OC. That usually works for me. It seems that sometimes the OC/BIOS update messes with something and a reset fixes it up.

My BIOS update process:

1. "Hard" reset BIOS.
2. Run Memtest for a few minutes to see if everything is still hunky dory.
3. Update BIOS using DOS/in-BIOS option (never Windows).
4. Boot computer, and make sure BIOS is updated.
5. Hard reset BIOS.
6. Run Memtest to see if everything is still hunky dory.
7. Re-apply OC.
8. ????
9. Profit!

EDIT: One more thing. I also look at the in-BIOS readings of different voltages to see how "different" they are from my setting. It seems that sometimes BIOS updates fix the vDroop/Drop that occurs for different measurements, or artificially adds an offset to the "set" voltage to make it appear that way. If my re-applied OC volts show a different in-BIOS reading, then I'll adjust to compensate.

I pulled the CMOS battery around 4 times with no luck. Maybe I didn't pull it for enough time.
 
TC folding
(7 items)
 
 
CPUMotherboardGraphicsRAM
2600K Asus Gene-z MSI 570 Ripjaws X Series 8GB 1600MHz 
Hard DriveCoolingOSPower
OCZ Vertex 3 XSPC Raystorm Block + 180mm Radiator Windows 7 Ultimate/Ubuntu 10.10 Corsair 750TX 
Case
Silverstone TJ08-E 
CPUMotherboardGraphicsRAM
2600K Sabertooth P67 ATI 4650 GSkill 2133 2x2GB 
CoolingOSCase
XSPC Rasa, RX240 Ubuntu 10.10 Fractal Midi 
  hide details  
Reply
 
TC folding
(7 items)
 
 
CPUMotherboardGraphicsRAM
2600K Asus Gene-z MSI 570 Ripjaws X Series 8GB 1600MHz 
Hard DriveCoolingOSPower
OCZ Vertex 3 XSPC Raystorm Block + 180mm Radiator Windows 7 Ultimate/Ubuntu 10.10 Corsair 750TX 
Case
Silverstone TJ08-E 
CPUMotherboardGraphicsRAM
2600K Sabertooth P67 ATI 4650 GSkill 2133 2x2GB 
CoolingOSCase
XSPC Rasa, RX240 Ubuntu 10.10 Fractal Midi 
  hide details  
Reply
post #76 of 93
Quote:
Originally Posted by JoshHuman View Post
I pulled the CMOS battery around 4 times with no luck. Maybe I didn't pull it for enough time.
Reset BIOS to default and run on stock settings for a few days. Try overclocking again after 3-5 days and see what happens.

I thought I fried a chip because I couldn't get any stable overclock. Let it sit off for a day, then ran it stock for a few days. After that I could hit the same overclocks as before. YMMV.
Scorpion
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 2500K Gigabyte GA-Z68X-UD3H-B3 MSI GeForce GTX 580 Lightning Corsair Vengeance 8GB DDR3-1600 CL9 
Hard DriveOptical DriveOSMonitor
OCZ Agility 2 120GB Samsung SH-S223L DVD-RW Windows 7 Professional 64-Bit AOC e2243Fwk 21.5" LED 
KeyboardPowerCaseMouse
XArmor U9 Mechanical Corsair HX850 Corsair Carbide Series 500R Logitech G500 
Mouse PadAudio
Nova Gaming Winner 3 Realtek ALC889 Codec 
  hide details  
Reply
Scorpion
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 2500K Gigabyte GA-Z68X-UD3H-B3 MSI GeForce GTX 580 Lightning Corsair Vengeance 8GB DDR3-1600 CL9 
Hard DriveOptical DriveOSMonitor
OCZ Agility 2 120GB Samsung SH-S223L DVD-RW Windows 7 Professional 64-Bit AOC e2243Fwk 21.5" LED 
KeyboardPowerCaseMouse
XArmor U9 Mechanical Corsair HX850 Corsair Carbide Series 500R Logitech G500 
Mouse PadAudio
Nova Gaming Winner 3 Realtek ALC889 Codec 
  hide details  
Reply
post #77 of 93
A lot of people on this thread are making the erroneous assumption that because Intel said that the maximum VID for a Sandy Bridge processor is 1.52v, 1.52v is also the maximum safe voltage that the CPU can be given. This is not true because VID is not the same as core voltage. VID is the maximum core voltage in a transient before vdrop and vdroop.

Intel actually specifies the details of this for Sandy Bridge in the VRD 12 specification but I can't access that because it's private. They have however, published the socket 775 VRD 11 document, which is probably similar:
Quote:
> To maintain processor reliability and performance, platform DC voltage regulation and transient-droop noise levels must always be contained within the Vccmin and Vccmax socket load line boundaries (known as the load line window).

> Equation 1: Vccmax Socket line line: VCC = VID - (Rll*Icc)

> Equation 2: Vcctyp Socket load line: VCC = VID - TOB - (Rll*Icc)

> Equation 3: Vccmin Socket load line: VCC = VID - 2*TOB - (Rll*Icc)
Where VCC = core voltage, Rll = resistance of it, ICC = core current, VID = VID, TOB = Tolerance.


http://www.intel.com/content/dam/doc...very-guide.pdf

Or in other words, VID is the maximum voltage in any transient, BEFORE vdroop AND vdrop, which are by design. Enabling LLC and setting voltage in BIOS to 1.52v, or messing with the load line, means you're breaking the spec.

VCC != VID

VCC != VID

VCC != VID

VCC != VID

The more I post the more they are different.



In another document intel specifies that the load line slope is 1.7 milli-ohms for Sandy Bridge processors. I assume this is the same as Rll in the previous datasheet. According to intel, vccmax for sandy bridge is therefore 1.52-Icc*0.0017. Maximum core current is supposed to be 112 amps. Intel usually also disclaims that these specifications are only valid under normal operating conditions, since we are running these things with way better cooling yet at massive overclocks it might be entirely invalid. We also don't know WHY intel chose these values. TDP? Lifespan? Or what?

http://www.intel.com/content/www/us/...datasheet.html

It is truly idiotic when someone sets load line calibration to a very high-setting, goes to 5.2ghz at 1.52v (load) then comes on a forum encouraging other users to do the same claiming they are within Intel spec. They have well and truely in practically every single way possible blown right past the intel spec. They just didn't actually read the intels guidelines or didn't understand them.


To find the actual max voltage that's fairly safe we can only rely on user reports and opinions of those who have tested a lot of CPUs, 'cause intel has in reality said nothing. The user reports might be also be unreliable and untrue. Juan Jose (ASUS Technical Marketing Specialist), wrote here http://hardforum.com/showthread.php?t=1578110 That the max 24/7 voltage for an overclock he could recommend was 1.425 volts. Most people say max voltage for an overclock should be between 1.38 and 1.40 volts. If overclocking then you're probably best keeping it below somewhere in the range of those values depending on how conservative you are.

My 2500K is at 1.35v.
Edited by Dodgson - 11/1/11 at 3:44am
Ma System Brah
(7 items)
 
  
CPUMotherboardGraphicsRAM
i5-3570K @ 4.4 GA-Z77X-UD3H MSI GTX 570 OC 8GB DDR3-1866 
Hard DrivePowerCase
Intel X18M G2 160GB Corsair HX620 Antec P280 
  hide details  
Reply
Ma System Brah
(7 items)
 
  
CPUMotherboardGraphicsRAM
i5-3570K @ 4.4 GA-Z77X-UD3H MSI GTX 570 OC 8GB DDR3-1866 
Hard DrivePowerCase
Intel X18M G2 160GB Corsair HX620 Antec P280 
  hide details  
Reply
post #78 of 93
I'm curious to those who say they need more Vcore now after running a lower one 3 months ago. What did you let you max temps get to under load when you were stressing? Are you the ones that said it was fine to stress up to 85c?? Or did you guys let you cpu hit 90c?? I have been one that will not let my cpu go over 75c PEAK and 70c avg during stress. Since May my chip has been stable at 1.392v @ 5ghz and has not needed any more Vcore at all. I have played with max booting at 54x and 1.55v. I've ran stress at 5.2ghz for 6hrs at 1.475v but never hit 80c. Now when running LinX AVX for a bunch of runs trying to win that highest Gflops, I have seen 90c but haven't let it run the full 20 runs.

I still have not needed a Vcore bump. So with that said, I wonder if it's people running real high temps on air or small water setups VS how much Vcore has been applied. I think that if you can manage 1.500v or less with 70-75c or less, you would not have any problems.

Anyone add to this???
Red X9
(20 items)
 
  
CPUMotherboardGraphicsGraphics
4790K @ 5ghz 1.350v 75c max load temps Asus Maximus VI Formula Evga Gtx Titan X @ 1500mhz / 7800mhz +112mv EK Acetal / Nickel / Nickel back plate 
RAMHard DriveHard DriveOptical Drive
Corsair Vengence Pro 32gig 2400mhz c11 red (4x8... 6x 256gig (1.5tb) in Raid 0 , OS / Gaming Drive 2 x Liteon 256gig (512gig) storage drives Asus Dvd-RW 
CoolingCoolingCoolingOS
2x 360mm Alphacool XT45 / 1x 480mm Alphacool Mo... EK X-Res 140 D5 vario Pump/Res combo + EK X-top... EK Supreme HF cpu / EK Acetal / Nickel with Nic... Windows 8.1 64bit 
MonitorMonitorKeyboardPower
Asus ROG Swift 27" 144hz with G-Sync Samsung 65" 7150  Logitech G510s Corsair AX1200 
CaseMouseMouse PadAudio
Thermaltake Core X9 custom painted Red and Black Logitech G602 Wireless Corsair Aluminum Core 2 sided  Bose Companion 5 
  hide details  
Reply
Red X9
(20 items)
 
  
CPUMotherboardGraphicsGraphics
4790K @ 5ghz 1.350v 75c max load temps Asus Maximus VI Formula Evga Gtx Titan X @ 1500mhz / 7800mhz +112mv EK Acetal / Nickel / Nickel back plate 
RAMHard DriveHard DriveOptical Drive
Corsair Vengence Pro 32gig 2400mhz c11 red (4x8... 6x 256gig (1.5tb) in Raid 0 , OS / Gaming Drive 2 x Liteon 256gig (512gig) storage drives Asus Dvd-RW 
CoolingCoolingCoolingOS
2x 360mm Alphacool XT45 / 1x 480mm Alphacool Mo... EK X-Res 140 D5 vario Pump/Res combo + EK X-top... EK Supreme HF cpu / EK Acetal / Nickel with Nic... Windows 8.1 64bit 
MonitorMonitorKeyboardPower
Asus ROG Swift 27" 144hz with G-Sync Samsung 65" 7150  Logitech G510s Corsair AX1200 
CaseMouseMouse PadAudio
Thermaltake Core X9 custom painted Red and Black Logitech G602 Wireless Corsair Aluminum Core 2 sided  Bose Companion 5 
  hide details  
Reply
post #79 of 93
I haven't had to bump vcore yet either, and I've ran it at over 1.4v since I got it.
    
CPUMotherboardGraphicsGraphics
Intel Core i5 2500K P8P67 PRO NVIDIA GeForce GTX 470 NVIDIA GeForce GTX 470 
GraphicsRAMRAMRAM
NVIDIA GeForce 9800 GT G-Skill A-Data G-Skill 
RAMHard DriveOptical DriveOS
A-Data Crucial M4 64GB + 1TB F3 Spinpoint $155 LS/DL DVD RW $?? Windows 8 64-bit "Epic Registry" Edition 
MonitorPowerCase
ASUS 21.5 1920x1080 2ms $135 CORSAIR HX850 $120 Mother Earth $free 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
Intel Core i5 2500K P8P67 PRO NVIDIA GeForce GTX 470 NVIDIA GeForce GTX 470 
GraphicsRAMRAMRAM
NVIDIA GeForce 9800 GT G-Skill A-Data G-Skill 
RAMHard DriveOptical DriveOS
A-Data Crucial M4 64GB + 1TB F3 Spinpoint $155 LS/DL DVD RW $?? Windows 8 64-bit "Epic Registry" Edition 
MonitorPowerCase
ASUS 21.5 1920x1080 2ms $135 CORSAIR HX850 $120 Mother Earth $free 
  hide details  
Reply
post #80 of 93
Quote:
Originally Posted by Hambone07si View Post
I'm curious to those who say they need more Vcore now after running a lower one 3 months ago. What did you let you max temps get to under load when you were stressing? Are you the ones that said it was fine to stress up to 85c?? Or did you guys let you cpu hit 90c?? I have been one that will not let my cpu go over 75c PEAK and 70c avg during stress. Since May my chip has been stable at 1.392v @ 5ghz and has not needed any more Vcore at all. I have played with max booting at 54x and 1.55v. I've ran stress at 5.2ghz for 6hrs at 1.475v but never hit 80c. Now when running LinX AVX for a bunch of runs trying to win that highest Gflops, I have seen 90c but haven't let it run the full 20 runs.

I still have not needed a Vcore bump. So with that said, I wonder if it's people running real high temps on air or small water setups VS how much Vcore has been applied. I think that if you can manage 1.500v or less with 70-75c or less, you would not have any problems.

Anyone add to this???
Mine did. I required 1.425v for 4.8GHz when I first got it. Now it takes 1.45v for 4.8GHz. My first chip did the same thing (required 1.375v for 4.8, 1.40v for 4.8 later).

However once this "burn in" (as I like to call it) happened, the chip has been rock solid since then. My "burn in" happened about 2 weeks into owning the chip. That was back in March. I've been rock solid stable since March at 4.8 and higher, with no more vcore needed than normal.

I don't know why, but it seems to be a trend that after a few weeks to a month or so, these chips tend to want a tad more voltage, but once that happens...it doesn't happen again.

And I've kept my chip under 1.52v as stated by Intel. My overclock right now uses 1.525v under a light load and 1.488v for full load (I never see it over 1.525v).
Junkyard
(18 items)
 
NAStradamous
(12 items)
 
HTPC01
(16 items)
 
CPUMotherboardGraphicsRAM
Intel i5 2320 @ 3.1GHz MSI B75MA-E33 HIS HD7970 3GB @ 1050 / 1700, 1.17v Patriot 2x4GB DDR3-1066 
Hard DriveHard DriveOptical DriveCooling
Crucial MX100 256GB WD Green 2TB DVDRW Corsair H50 Push / Pull 
CoolingOSMonitorMonitor
NZXT 120mm fans strapped to stock 7970 cooler Windows 10 Pro Dell 3007WFP Dell 2007FP 
MonitorKeyboardPowerCase
Dell 2007FP Cheap Toshiba (R.I.P. Ducky) PC Power & Cooling Silencer MKIII 950w Cheap Garbage (free) 
MouseAudio
Logitech G602 Yardsale Special 
CPUMotherboardGraphicsRAM
Intel Core i5 3330 Asus H61 MSI GTX 750Ti Low Profile 2x4GB Hynix DDR3-1333 
Hard DriveHard DriveOptical DriveCooling
60GB Kingston SSD 1TB Seagate Samsung DVDRW CoolerMaster Gemin II 
OSMonitorKeyboardPower
Windows 8.1 + Steam Big Picture / Plex HT 50" Insignia LED 1080p Logitech Rosewill Green 530w 
CaseMouseAudioOther
Silverstone HTPC Logitech Pioneer 2.0 tower speakers + 135w RMS Rosewill MCE Remote 
  hide details  
Reply
Junkyard
(18 items)
 
NAStradamous
(12 items)
 
HTPC01
(16 items)
 
CPUMotherboardGraphicsRAM
Intel i5 2320 @ 3.1GHz MSI B75MA-E33 HIS HD7970 3GB @ 1050 / 1700, 1.17v Patriot 2x4GB DDR3-1066 
Hard DriveHard DriveOptical DriveCooling
Crucial MX100 256GB WD Green 2TB DVDRW Corsair H50 Push / Pull 
CoolingOSMonitorMonitor
NZXT 120mm fans strapped to stock 7970 cooler Windows 10 Pro Dell 3007WFP Dell 2007FP 
MonitorKeyboardPowerCase
Dell 2007FP Cheap Toshiba (R.I.P. Ducky) PC Power & Cooling Silencer MKIII 950w Cheap Garbage (free) 
MouseAudio
Logitech G602 Yardsale Special 
CPUMotherboardGraphicsRAM
Intel Core i5 3330 Asus H61 MSI GTX 750Ti Low Profile 2x4GB Hynix DDR3-1333 
Hard DriveHard DriveOptical DriveCooling
60GB Kingston SSD 1TB Seagate Samsung DVDRW CoolerMaster Gemin II 
OSMonitorKeyboardPower
Windows 8.1 + Steam Big Picture / Plex HT 50" Insignia LED 1080p Logitech Rosewill Green 530w 
CaseMouseAudioOther
Silverstone HTPC Logitech Pioneer 2.0 tower speakers + 135w RMS Rosewill MCE Remote 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Did we ever come a conclusion on safe SB voltages?