Do you know how many strokes you did? I will be doing this as soon as I get home to see if I can get this OC stable.Quote:
Originally Posted by sbinh
Printable View
Do you know how many strokes you did? I will be doing this as soon as I get home to see if I can get this OC stable.Quote:
Originally Posted by sbinh
Don't rely on number of strokes because it depends on how hard each person push pencil on the resistor.Quote:
Originally Posted by PanteraGSTK
In my case, I don't know how many strokes I put on that resistor.... But one thing I know that I brought it down from 130kiloOhm to 77.6kiloOhm.
If it is too much, erase it and reapply ... :D
You need a Digital multimeter to measure size of resistor, vcore ...... You can buy one at RadioShack ...
I've got a DMM, I just wanted to know if there was a specefic drop in Ohm's with each stroke. So 77.6kohm is about where I should aim?Quote:
Originally Posted by sbinh
Yup .. I got 3 values 77.6, 78.3 and 82.4 and 77.+ seems to be best out of those 3. If you are able to get around 77-77.6, that would be good.
Play around with it, you might find a better spot :D ....
pencil mods will just end up going wrong.
they are too volatile and unreliable.
the voltages/ohms/amps/watts just fluctuate all over the place.
extremely dangerous.
and after my last experience with Shamino's pencil mod's, i wont ever trust another one of his mods. (they dont even test them), they rely on us to do it and fry our hardware.
so you take that for whats its worth.
How? You blow off excess graphite and put tape over it. How is it going to "end up going wrong"?Quote:
"pencil mods will just end up going wrong."
I have an x800 with pencil mods that are 3 years old. It's been going right for 3 years.
Using a multi-meter, measurements are exact. My vdroop mod is reliable, if you ask me. I used to have vdroop and now I don't.Quote:
they are too volatile and unreliable.
My volts stay where I put them, thanks to the vdroop mod!Quote:
the voltages/ohms/amps/watts just fluctuate all over the place.
If the mod were to corrupt (somehow) I would have vdroop again. No big deal. Not even slightly dangerous.Quote:
extremely dangerous.
By the way, what mod did you have go wrong?
Sorry to say that, but if you really feel the need to modify your hardware, you need to be aware of the fact that something can go wrong. And even "gurus" like shamino can't test every single mod on their own. They get asked for mods and do the best to help the community. If it goes wrong, that's bad, but when doing a mod you need to keep in mind that things like that can happen anytime. If you can't bare the risk, then don't modify. Or learn how to work out mods on your own in order to know what you are doing, giving you some kind of security while performing the mods. ;) Don't misunderstand me, no offence taken. I understand your point, but you also have to respect the work that, in your example, shamino puts into working out these mods. Nobody's perfect.Quote:
Originally Posted by Lestat
if you cant test a mod for validity then you god damn well better not be posting it on your freakin website as a valid mod.Quote:
Originally Posted by celemine1Gig
Kuunak even told me not to do the pencilmod on the DS3 for vdroop as it would FRY the motherboard.
when i asked him then why the hell is the mod still being shown on vrzone, he stopped talking to me and hasnt said a word since.
the point is if you can test it dont post it as a valid mod.
all your doing is fvcking with peoples minds and hardware.
Well anyway the pencil mods that are in this thread DO work. I've tried the vdroop and vcore mod and they both worked perfectly.Quote:
Originally Posted by Lestat
I think like anything on the internet you can't just take it at face value. You really have to dig down deep and find out as much as you can before doing any kind of mods. You can't just do something because someone else told you. If you're going to do a mod you have to make sure you understand the risks you take. It always sucks when something screws up but in the end there's no one else you can blame but yourself because you're the one who did it without finding out what you're really doing. I have to admit myself that I took a risk when I did the mods but I did cross reference the chip with other boards before doing it. Also I knew that if the board died on me I could get it replaced without costing me too much.
I did the vdroop mod and it worked very well. I need to redo it again because I can't stop messing with it. ( still .02v droop). Has anybody done a vmch mod with a pencil?
Pantera did you do the 2 locations or just 1 location ?
i cant take the droop anymore.
when you hit about 1.5v the droop is insane 0.10v atleast
so that means 1.4v load.
but my board is working so good i dont want to !$%^ it up.
*EDIT*
oops. those two pictures are of the same thing, just the bottom picture is rotated.
Are you measuring with a DMM? PC Probe's results are pretty much BS. Even after the vdroop mod, with 1.5v selected in BIOS and showing on my DMM, PC Probe gives me 1.44v idle, 1.32 load.Quote:
Originally Posted by Lestat
Before the mod, I was getting around .05 ~ .06v droop at load, which is still pretty awful, but it's nothing close to what PC probe was reading.
After the mod, I get only .02v droop at the most. It increased stability a lot for me. I don't like pencil mods either, but it is working until I have time to put the VR on there.Quote:
Originally Posted by Lestat
holy crap the vdroop on my board is TERRIBLE. With a BIOS setting of 1.4v and +100mV OFF I get (measured with a DMM):
(at 129.5k ohms before mod)
1.337v idle
1.264v load
Droop of .073V
I took the resistance down to 77.3k ohms and here's what I get:
1.355v idle
1.285v load
Droop of .07v
So that essentially did nothing for vdroop...it helped by like .003v, which is well within the error for the multimeter. The only good thing it did was increase voltage by .02 so its closer to what I set in bios.
Now what? Should I take the resistance down even more or ....?
Raise the resistance up a little bit ..... like 77.6 to 77.8
I'm just curious what .2ohms is going to do? I tried 70.6ohms and there was very little difference from 77.6ohms (i dont remember the exact values, just that there was little difference). Going from 77.6 to 77.8 is going to be virtually impossible with a pencil... I'm not skilled enough to solder on the 1M resistor in parallel.
Sorry for confusing .. what I meant was that you need to raise it from what you have (77.3 to 77.6 or 77.8) ...
Yup .. it takes lots of time and effort ... Erase ... apply ... erase ... then apply ... :D :D
Yea...I wish I could solder cause this would be so much easier! One day I'll actually get off my ass and practice on some old boards.
Why is everyone getting different resistance values though? 28.5k ohms, 33k ohms, 105k ohms, 77k ohms...?.
Edit: Okay so I've tried 70k ohms, 77.3 k ohms, 77.8 k ohms, 96.2 k ohms, 109.5 k ohms, and 129.5 k ohms and there is very little difference in vdroop... almost 0.07V every time. What are the risks to going down to 28.5/33 k ohms as others have had success with?
I'm pretty confused with what decreasing the resistance of this actually does???
edit 2 and 3 and 4: more results...
1.343v idle 82.7k ohms
1.283v load 82.7k ohms
.06vdroop
1.345v idle 77.6k ohms
1.285v load 77.6k ohms
.06vdroop
1.350v idle 66.7k ohms
1.295v load 66.7k ohms
.055vdroop
1.351v idle 57.3k ohms
1.307v load 57.3k ohms
.044vdroop
1.3431v idle 41.6k ohms
1.310v load 41.6k ohms
.033vdroop
1.361v idle 31.5k ohms
1.328v load 31.5k ohms
.033vdroop
1.362v idle 27.6k ohms
1.331v load 27.6k ohms
.031vdroop
1.356v idle 20.7k
1.330v load 20.7k
.026vdroop
1.348v idle 14.9k
1.327v load 14.9k
.021vdroop
1.348v idle 12.4k
1.327v load 12.4k
.021vdroop
1.357v idle 9.75k
1.338v load 9.75k
.019vdroop
So it looks like I'm stuck at ~.02vdroop...not as good as I had hoped
Bump...I could really use some help :(
From my own experiments, the mod depends on voltage level that you set.
When you increase vcore, you might need to re-apply the mod.
Like I mentioned on previous posts, vcore was at 1.500v ... I tried to raise vcore up few notches last weekend .. guess what, vdroop changed from 0.003 to 0.04 (back to nornal) ....
In your case, like couppi said, you might only need to take that resistance down to 90-100K
well i finally got around to doing the crazy pencil mod and,, well. i'm shocked
1.4 bios = 1.42 bios and windows (idle)
Orthos Gromacs load 100% = 1.44v
Prime load = 1.42v
for a grand total of ZERO DROOP. or 0.02v INCREASE if its 100% both cores.
i went back and forth over the resistor maybe 8 times with a pencil
It would be interesting for an electrical engineer to analyse this circuit to let us know what its actually doing. Varying the pencil resistance by less than 10% makes no sense. Its ohms law, not voodoo. It behaves is regulated and predictable ways. Eg; 77.6 verses 77.2 would make negligible difference.Quote:
Originally Posted by Kobalt
Kobalt what you have done looks good. It seems that vdroop improves as resistance decreases. If we took it just on these figures, you'd expect that dropping the resistance to 0 would give use no droop. However, what if this resistor is part of a feedback circuit or part of some other aspect of power regulation? Without knowing its full function, something else could be dangerously affected.
Quote:
Originally Posted by dexx
I took it all the way down to 5k ohms I believe and vdroop was still .02v. I couldn't even go down farther no matter how hard i pressed with the pencil. It doesnt really matter though cause this was my friends computer and its now in a safer environment :p:.
Quote:
Originally Posted by dexx
Have a thorough look at this document, provided by Intel, and especially the loadline figures and numbers, starting on page 13 for the different processor types:
http://download.intel.com/design/pro...s/31321402.pdf
It shows that vdroop is simply put a function of Icc (the current, that the CPU draws). The higher the current draw the higher the voltage drop (relative to the CPU's standard voltage; have a look at the graphs). AFAIK Intel wants all mainboard manufacturers, using their chipsets, to keep to these loadline specs. They need to stay within the specified loadline window over the whole scale from 0 to the max specified current draw in Amps. If they don't, I guess their circuit is either not certified to be VRD compliant, or perhaps they won't even get the chipsets, if they use Intel chipsets at all. I honestly don't know for sure. Could ask next week, but don't know if I'll get an answer at work. Perhaps I'll have to ask someone from the development section.
Now that explains why all Intel motherboards have the droop function implemented from the start. It has its purpose for sure. The droop resistance is chosen in order to keep the circuit compliant with the loadline sheme over the whole scale (for all compatible CPUs!).
It controls the max possible current draw at a certain operating point.
The higher the droop resistance, the lower the possible current, the higher vdroop. And respectively: The lower the resistance, the higher the possible current draw and the lower vdroop. It's that easy. BUT, what you have got to keep in mind:
1.) If you change your CPU Vcore, the CPUs current draw will increase, which will influence vdroop. You can't adjust it once and think that it'll keep vdroop at close to zero in all load situations. You have to adjust it for each operating point of the circuit.
2.) The boards are designed with vdroop functioning as shown in the Intel specsheet. If you now alter the resistance and allow more current to flow at a certain operating point, you normally would have to increase the decoupling capacitance to help keep the vcore stable at that point. It's because Vdroop is also there to help the motherboard manufacturers save some money. It takes off some strain off the voltage regulating circuit, thus helping them to cut costs, for example for expensive caps.
And concerning Kobalt's problem:
His vdroop doesn't decrease, even when lowering the droop resistance to a value far lower than anyone else with this board. It looks like he has some current limiting factor elsewhere in his system/circuit. The droop function would allow for enough current to flow to decrease vdroop, but the cuircuit just doesn't seem to get enough current or can't deliever it due to a problem I don't know. IMHO, this seems to be the explanation why his vdroop doesn't get any lower than 0.02V. Unfortunately, neither do I have access to his very motherboard, nor am I in posession of at least the same model. So, IMHO his whole system would need to be thoroughly analyzed to find the "fault", causing this behaviour of the circuit.
Edit: BTW, because you mentioned dropping the droop resistance to ZERO Ohms. BAD IDEA! No current limit --> short circuit --> very likely dead motherboard or at least dead vcore regulation, if not even dead CPU, too. ;)
kobalt,
given you have tried a vast array of resistance i will go out on a limb here and say your not doing the mod right.
wrong location
bad POT
smoking to much POT, lol j/k