Autochthon Posted July 6, 2005 Posted July 6, 2005 I would think the sensors can give a reasonable relative value. I just wouldn't trust them for absolute or high precision values. your 12V isn't varying that much less than 1% is very good. Share this post Link to post Share on other sites More sharing options...
higherøx Posted July 6, 2005 Posted July 6, 2005 I would think the sensors can give a reasonable relative value. I just wouldn't trust them for absolute or high precision values. your 12V isn't varying that much less than 1% is very good. So do you think it's good? If the answer is "yes"... Then it's OK! Thanks for your opinion. Share this post Link to post Share on other sites More sharing options...
stealth17 Posted July 6, 2005 Posted July 6, 2005 So do you think it's good? If the answer is "yes"... Then it's OK! Thanks for your opinion. anyone can tell you yes or no. ill say both how is that? no one will know for sure until you get a DMM and check the actual. to be an overclocker its a must. it will be the best $30 at radioshack you will ever spend but if the sensor on the 12v rail is correct, then yes its fine. but still check with a DMM becuase it could be better, or worse. no way of knowing for sure Share this post Link to post Share on other sites More sharing options...
timpanogos Posted July 6, 2005 Posted July 6, 2005 yep relative needs a TO Share this post Link to post Share on other sites More sharing options...
Autochthon Posted July 7, 2005 Posted July 7, 2005 relative VARIATION If the sensor is 2 Volts low then it's still bad but that is made pretty clear by many many posts saying check with a DMM and he acted like he was concerned with the variation. And I'll concede that the sensor could be bouncing around while the voltage was steady though that seems odd to me. It seems a more likely possibility that the variation is smaller than the sample points outputed by the sensor but what do I know Share this post Link to post Share on other sites More sharing options...
Autochthon Posted July 8, 2005 Posted July 8, 2005 looks like support for the idea that this fluctuation is is on the vdimm line and that this voltage may be a incorrectly scaled vdimm reading. http://xtremesystems.org/forums/showthread.php?t=68114 Share this post Link to post Share on other sites More sharing options...
flexy Posted July 8, 2005 Posted July 8, 2005 autochton, NO. Please dont confuse the VTT issue with the Vdimm issue ! The VDIMM issue occurs as soon as you set your VDIMM too close to your 3,3V rail....has been mentioned earlier also. Thats the reason why it is always better to have either) VDIMM 0.2 or 0.3V BELOW 3,3V rail (so it doesnt drop) or) Mod/increase your 3,3V rail a notch so it doesnt drop There should always be a GAP at least 0.25V between actual VDIMM and what the rail supplies ! ---> It has NOTHING to do with the VTT fluctuations because those occur no matter whjat Vdimm is at. EMC also clearly showed in his readings that VTT is dropping like a MoFo while VDIMM, VREF etc. all stay STEADY. The VTT issue *seems* to be an issue happening solely in the voltage regulator chip RT9173A...as said measuring all the inputs to the voltage reguator shows they look fine...just the OUTPUT of RT9173A (the VTT) drops...so it MIGHT be a load issue or even bad parts. They do a great job over there and i really, really hope they come up with a easy mod etc. especially for the VTT issue. Share this post Link to post Share on other sites More sharing options...
Spotpuff Posted July 8, 2005 Posted July 8, 2005 Used OCCT to successfully bench my new 2.4 GHz oc (priming overnight, ran 4.5 hours stable, it's still going while I'm at work) and the graphs OCCT generated are nowhere near as fluctuating as the ones that people are posting. I am using an OCZ Powerstream 420. I'm disappointed that my board's HTT doesn't seem to want to go over 280 and that the CPU doesn't seem to want to go over 270, but at 267x9 (3/4 divider, ram is going at around 200) the thing so far seems rock stable; passed OCCT stress test, and 2 hours of torture, not to mention 4.5 hours of prime that's probably still going right now. I hope people get to the root of this though; more OCs are a good thing Share this post Link to post Share on other sites More sharing options...
Autochthon Posted July 8, 2005 Posted July 8, 2005 Flexy, I wasn't confusing the two. I had suggested early on that the mystery voltage may be related to Vtt (because It was near 1/2 vdimm) then as shown by my later couple posts was leaning towards an improperly scaled Vdimm as what i'm actually seeing. I was commenting on how when I changed the vdimm voltage from 3.2V to 3.1V the magnitude of the fluctuations had dramatically decreased. Possibly because I had improved the buffer between 3.3v rail and Vdimm (from ~ 0.15V to ~0.25V). If you check out a few other posts I have scattered around here I think you'll see I have repeated mentioned how the gap should be at least 0.2V. Now this speculation about these fluctuations could still be completely wrong as I would need to put a reliable measurement on Vdimm and test while observing this "mystery voltage" output simultaneously. Edit: It occures to me I should elaborate. I thought I had a 0.25V buffer originally as I had tweaked my 3.3 rail to about 3.45V. But when I actually measured my Vdimm using a DMM I found it was measuring 3.31 at a bios setting of 3.2V. By lowering my bios Vdimm to 3.1V. I got closer to the 0.25V buffer I had thought I had. I was commenting on how this appeared to be indicated by the increased stability in the unidentified voltage I was messing around with. Share this post Link to post Share on other sites More sharing options...
timpanogos Posted July 8, 2005 Posted July 8, 2005 maybe we should try and split this thread into a vtt thread and a vdimm thread ... I wonder if we give ag a list of post id's if he can/will split it Share this post Link to post Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now