Undervolting Core Duo & Core 2 Duo by Pin Mod - PART 2

UPDATED 20th MAY 2009

I successfully undervolted via pin mod the T5300 by 0.2v . Without pin mod, and using RMclock the T5300 runs stable with a vcore of 0.9500v for all its multipliers.

The pin mod consisted on connecting ViD4 to the closest Vcc to force it to be High (1) all the time. After the pin mod those are my readings:

------- RMclock -- real vcore
06x --- 0.9500v --- 0.7500v
07x --- 0.9500v --- 0.7500v
08x --- 0.9625v --- 0.7625v
09x --- 1.0000v --- 0.8000v
10x --- 1.0375v --- 0.8375v
11x --- 1.0750v --- 0.8750v
12x --- 1.1000v --- 0.9000v
13x --- 1.1250v --- 1.1250v

If you want to read more about my experiments download the following PDF

 

I pinmod the Q9200 from 1.1625V to 1.0375V, at least it didn´t die. And i think the mod is successful because temp really lower than before. Don´t know about lower multis, i alwasy use full speed.

How do you know your default Vcores x10 is 1.2615V? Read from CPUZ?

 

I read it from RMclock "CPU info" panel.

Kaltmond why did you do your undervolt? what did you connect to what?

 

I think RMclock read the same like CPUZ, but i think it´s not correct. Try use Core temp to get the default VID, that should be the real one.

Under full load 4 cores create too much heat there, so i need it to run cooler.
I connect from 0011011 to 0100101, VID1/3/4 to VSS, VID 0/2 to VCC, and a long wire to connect VID5 to VCC.......Totally 6 copper wires......

 

6 wires that's impresive
thanks for the info regarding Core temp

 

i been looking online on pin modding.... and i find nothing... how do i unlock voltages... i am limited to min of .9 v on my t9400, and i am at .9v from 6x - 8x multiplier... im sure i can go lower.

 

I'm mildly interested in undervolting via pin mod (as a method to bypass Intel's blasted VID lock), but unfortunately there's not much I can contribute from my own knowledge. You might want to try contacting this guy, though; he seems to be pretty well read in the matter:

http://forum.notebookreview.com/showthread.php?p=4659239#post4659239

Also might want to try reading that RMClock forums thread if you haven't already; though some of the information might just be repetitive. It kind of just went over my head for the most part -_-

 

Commander Wolf thanks for your comment.
I've read everthing I could find on the internet about this subject as I've looking into doing it for the past year. I also exchanged few PMs with heinz2005 regarding this subject.

1- Can someone explain what I did wrong to end up rizing the Vcore instead of decreasing it?
2- Any good software to read the Vcore before pin mod? or is RMclock reading accurate?

 

A concept for a dynamic undervolting may be to connect VID0 to VID5.
The result should be a mapping to a range of 0,9375 down to 0,7125.

 

Here is a small electrical diagram of a typical CMOS-Output circuit (simplified):

I think at the mobile processors Intel use CMOS-Output to minimize the power drain.
So a pinmod to VCC or VSS would be a shortcut of one of the driver transistors. :-(
So I have to be very careful with pinmodding the VIDs.

IMHO the VID-Output goes directly to the voltage regulator chip.

What is the relevant specification for the Montevina voltage regulation?
"IMVP-6+. (or Intel MVP-6+). Intel Mobile Voltage Positioning – revision 6+."

If we do any pinmod (without soldering) we should take care of the driver transistor e.g. by using maybe
a resistor pulling the VID softly to another binary value.
The resistor can be sized according to the inner resistance of the activated CMOS gates.


Source: http://www.xilinx.com/support/documentation/application_notes/xapp329.pdf

 

Q: What is the relevant specification for the Montevina voltage regulation?
A: "IMVP-6+. (or Intel MVP-6+). Intel Mobile Voltage Positioning – revision 6+."

Here is a small electrical diagram of a typical CMOS-Output circuit:

I think at the mobile processors Intel use CMOS-Output to minimize the power drain.
So a pinmod to VCC or VSS would be a shortcut of one of the driver transistors. :-(
So we have to be very careful with pinmodding the VIDs.

IMHO the VID-Output goes to (the VID-Input circuit) the voltage regulator chip.

If we do any pinmod (without soldering) we should take care of the driver transistor e.g. by using maybe
a resistor pulling the VID softly to another binary value.

 

Hi All

I just got my hands on a T5300 (FSB 133). Guys how does it compar to the T2300 (FSB 166)?

Yesterday I successfully undervolted via pin mod the T5300 by 0.2v . Without pin mod, and using RMclock the T5300 runs stable with a minimum and maximum vcore of 0.9500v.

The pin mod consisted on connecting ViD4 to the closest Vcc to force it to be High (1) all the time. After the pin mod those are my readings:

------- RMclock -- real vcore
06x --- 0.9500v --- 0.7500v
07x --- 0.9500v --- 0.7500v
08x --- 0.9625v --- 0.7625v
09x --- 1.0000v --- 0.8000v
10x --- 1.0375v --- 0.8375v
11x --- 1.0750v --- 0.8750v
12x --- 1.1000v --- 0.9000v
13x --- 1.1250v --- 1.1250v

for 13x the Vcore set in RMclock and the real one are the same. This is because by modding ViD4 I lost all the Vcores between 0.9125v and 1.1000v including 0.9125v and 1.1000v.

NEXT:
Try to up the FSB from 133 to 166 or even 200 ... can't doing right now as the laptop (Ei-System / Advent) I'm using has an intel chipset

 

Naton,
when I had my T5300 in my E1505, I could run all multipliers at 0.950V, which was the lowest voltage rightmark allowed me to undervolt my processor

Any pictures, I may be interested trying it with my T7400
It would be nice to push the idle voltage from 0.950V down

0.750V, wow that is much lower
I found my T7400 to run cooler than my T5300 because the T7400 had a larger die, allowing a more efficient transfer of heat between the cpu and heatsink

K-TRON

 

Hi K-T
I will take a picture and upload it in the next few days.

I notice that the T5300 was running hotter than the T2300 for some reason. Maybe it is down to the size of the die as you said.

Before I did the VID4 mod I did a VID5 mod. VID5 = H allows to lower the Vcore by 0.4v. The computer started and I was able to access the bios but Windows refused to load.

For the T7400 the VID4 mod maybe too excessive because you will lose all the Vcores between 0.9125 and 1.1000v. I think you should use VID3 (i.e -0.1v) instead, since ViD3 gives a wider range of Vcores.

 

On my T7400
From the 9X multiplier up I need more than 0.950V for it to run stable

9x is 1.0000V
13 x is 1.0500V


Only multipliers 6, 7 and 8 can be set to 0.950V
So I am guessing the 6 and 7 multipliers can go down a bit more.
Probably not all of the way down to 0.750V

K-TRON

 

yes but by doing the VID4 as I did your x8 and x9 multipliers will end up using 1.1v as minimum Vcore. This because you won't have access to 0.95v anymore after the mod.

I think that the T5300 can go all the way to 0.75v and I suspect even below (i.e. 0.7125v) is because it uses an FSB of 133mhz only.

 

Hi

I made the following PDF to explain how to undervolt through pin mod. I also spent the whole weekend thinking about the pin mod and those are my conclusions:

1) VID4 mod as explained in the PDF is only useful to those of you who can undervolt with RMclock all the multipliers of their CPU to 0.95v.

2) Those that can't use the VID4 mod should try the VID3 mod (also explained in the PDF). In my opinion the VID3 is not as good as VID4 because if in theory VID3 allows to undervolt by 0.1v, in practice you can not go lower than 0.9125v. In other words, 0.0375v undervolt

Link to the PDF here

 

I haven't had the time to read the whole thing, but this is a nice guide. Will definitely see if I can try this out sometime. What're the chances it'll work for Socket P processors?

 

I'd love to to able to get my lower multiplier down to 0.750V, I only have 3 multipliers which are 6x, 7x & 8x and they all run at 0.950V.

I will watch these two threads and help out if I can (doubtful ).

Thanks

 

This mod will work but the pin configuration for the mod will be difference. So you need to lock to the intel documentation for socket P to identify the location of the different VIDs

 

You can try the VID4 mod and see if it suits you. Besides this Ei-system laptop I have a Gateway laptop with a T2500 CPU. Most of the time I use my Gateway with the multiplier locked at x6 (i.e. 1 ghz). So in this case it is worth it for me to do the VID4 mod on the T2500.

Now if I was someone who games a lot, and had my CPU running at full power all the time the VID4 is not worth doing. This is because with a mutliplier of x12 my CPU is stable with 1.05v before mod. After the mod the lowest vcore I'll be able to use would be 1.1250v. So technicaly speaking I would have lowered the Vcore for x6, x7, x8, maybe also x9, but I would have increased it for x10, x11 and x11.

 

I'm not entirely sure that the vid lock is in the CPU itself, although it very well may be. This is because on my X9000, the lowest VID I can get to (after unlocking VIDS in RMClock and CPUgenie), is 1.0V. I couldn't really find any other posts with X9000s to confirm if anyone can go lower. So somehow I think it might be the chipset PLL in my Dell. Just a thought

 

I can get my P8400 down to 0.8750V with RMClock.

 

Keep up the good work and experimentation. Giving you a rep for your hard work!!!!

 

Maybe I'm misexplained myself. If you look at the VID/Vcore table in the PDF you'll see that forcing VID4 to be 1 all the time means that there will be no combination of VIDs to allow the selection of Vcores between 0.9125v and 1.1000v

Now the minimum Vcore for the highest multiplier in a CPU (for example x13 for the T5300) has nothing to do with the chipset. It's the minimum voltage that the CPU need to run at a given speed.

For the minimum voltage used by the smallest multiplier (e.g. 0.9500v in the case of multiplier x6 for the T5300) I will agree with you it's not a lock in the CPU. And I don't think it's a lock in the chipset or the BIOS either. I mean if it was a lock in the Chipset or the BIOS my pin mod would have failed and the laptop would have refused to boot.

I think since I've managed to boot in the BIOS with VID5 = 1 (i.e. with 0.7750v for multiplier x13) I think the lock is maybe in the drivers of the chipset / CPU or else.

you have high chances to succeed with the VID4 mod. Note the pin configuration for socket P is different from socket M. So check the intel documentation before attempting the VID mod.

Thanks

 

So I noticed you never actually describe the physical pin mod in the guide? The holes and socket are all fairly small on Sockets M and P; what kind of wire did you use such that you were properly route it around and into selected pin holes and where did you get it? Any pictures?

 

The holes are similar by I thought their configuration was slightly different in the top left corner. In addition the location of the VIDs are not the same

I used an electric wire... not sure about its name. I took a wire of an electric cable used in a power extension.

I'll try to take and upload pictures of the phisical mod I did in my laptop in the next few days . Over the weekend maybe

 

Heres a good reference of the idea

http://www.notebookreview.com/default.asp?newsID=3226

 

Since thihs thread is about undervolting,
i'm going to ask a question regarding undervolting.
i know undervolting lowers the cpu temp, but does it also
decrease the performance of the cpu or is it same?
So any effect on cpu performance?

 

Straight from flipfire's UV guide:

 

thx Commander for quick response,
i learned something

 

Thanks to naton for the great PDF guide!

I will get my palm wet soon applying the mod to this yonah sillyron-m 430

 

Bravo, bravo!

Great PDF, reads as well as it should. +1

Your undervolt is very impressive for a T series, it makes me wonder why Intel still leave the voltages as high as ~1.2V.

My P series is a bit odd, CPU-Z shows a core VID of 1.038V when I have set RMClock to 0.925V

I want to try pinmodding but I'll save it for the holidays

 

This really shows how we have opposite sides of the spectrum.

You guys want to undervolt, but over in the Asus W90 forum we want to overvolt.

Esp the QX9300 owners so we can get about a 3.6+ghz overclock. Is it possible to overvolt the cpu the same way, and any leads/guidance for us?

 

Overvolt is possible, have done that with my X9000 before, but it´s really hot....

 

W90 has one of the greatest cooling units I have seen, it will handle it well I think, infact it overvolts it some on its own via default overclock software and it still stays cool.

 

I think they do this so they can sell their low voltage CPUs.

 

Hi guys
I haven’t posted photos showing the pin mod I did on the T5300 yet as I don’t have a camera at the moment.

A little update; I did a ViD4 mod on a Pentium Dual Core T2060 and I tried to do a Bsel mod on a Celeron M430.

T2060 before mode:
06x ---------- 1.0250v
12x ---------- 1.2125v

T2060 with Rmclock only
multiplier --- Rmclock
06x ---------- 1.0250v
07x ---------- 1.0250v
08x ---------- 1.0250v
09x ---------- 1.0250v
10x ---------- 1.0250v
11x ---------- 1.0250v
12x ---------- 1.0250v
I don’t have access in Rmclock to any Vcore lower than 1.0250v. I think this value must have be locked in the T2060 by intel.

T2060 with RMclock and ViD4 mod:
multiplier --- Rmclock --- Real Vcore
06x ---------- 1.0250v ----- 0.8250v
07x ---------- 1.0250v ----- 0.8250v
08x ---------- 1.0250v ----- 0.8250v
09x ---------- 1.0500v ----- 0.8500v
10x ---------- 1.1000v ----- 0.9000v
11x ---------- 1.1125v ----- 1.1125v
12x ---------- 1.1125v----- 1.1125v

Celeron M430 and Bsel1 mod:
The M430 default fsb is 133mhz. This corresponds to the following Bsel combination: Bsel 2 1 0 = L L H. To force this CPU to run at 166mhz the Bsel combination should be Bsel 2 1 0 = L H H. What’s next is based on what I’ve found on the internet a while ago:

1/ I connected Bsel1 to its nearest Vcc to force it to be High.
My laptop booted with the M430 at its default speed.

2/ I connected Bsel0 and Bsel1 together with a copper wire thinking that this will result in Bsel1 = Bsel0 = H.
My laptop turned ON but nothing displayed on the screen. I had to keep my finger on the power switch for 4-5 seconds to force my laptop to turn Off.

3/ I tried to insulate the Bsel1 pin on the CPU.
I couldn’t find anything small enough to do that so I did the next step instead.

4/ I broke the Bsel1 pin from the M430, and connected the Bsel0 and Bsel1 on the socket together with a copper wire.
The result was the same as ‘2’.

5/ I put some super glue on where the Bsel1 pin was in the CPU because I thought that it was still touching the copper wire. Bsel0 and Bsel1 still connected, I switched my laptop ON.
The result was the same as ‘2’.

6/ I removed the copper wire from the socket and tried the M430 with the broken pin.
My laptop and the M430 worked normally. Breaking the pin didn’t damage the CPU.

I’m not sure why my Bsel mod failed when many claim that they have succeeded doing it. The chipset in my Eisystem 3103 is intel i943GML.

If you know what I did wrong let me know.

 

you overvolt to make your laptop run faster. My problem with overvolting is the extra heat generated by the overvolt. My laptop is ON 6 - 8 hours a day, and most of the time it sits in my laps. Therefore, I'm keen on having my laptop as cool as I can, and this why the biggest laptop I'm willing to use on a daily basis is a 14" laptop.

If I had a 17" laptop like yours sitting on a desk I would definitly overclock... well if I can of course since most laptops can't be overclocked (because of limited BIOS or unavailibility of overclocking software).