BSEL Mod on a socket P explained with photos

are you asking me about the the serial number for your motherboard?
This should be above or around the memory slot.

You can also know what type of chipset you have by readin the print on the top of the chipset. But for that you need to remove the CPU heatsink first since the chipset is located under the CPU heatsink.

 

where is the place for the modding? in the cpu socket right?

 

sorry my english is not very good. i mean do i have to remove all the wire's cover or insulator to do the mod?

 

It's not your English. I simply didn't understand the context.

Yes the mod is executed within the CPU socket by connecting the holes with a wire.

I didn't use insulated wire. I used the wire from an IDE cable. I've also tried with an electric wire. You can try with an insulated wire if you wish but you need to remove the insulation of the bit that goes inside the CPU/Socket hole. You can remove it with a knife

 

i've tried to do the mod...it is difficult to insert the wire to the hole. besides, the hole is very small

 

That's true. The first time I did it it took me about 30 minutes to get the wire to stay in the holes.

But you didn't answere the most important question of all. Do you have the i940 chipset? or the i943/i945?

 

Hi guys can anybody answer these questions?
1)For how many seconds do you need to keep the wire connecting the pins?
2)If I do the -0.2 vmod can eventually apply the -0.1 vmod?
3)The author stated that voltages between 1.125 and 1.2v won't be affected by the vmod (at leats that's what I understood).
I'm waiting to receive a laptop mounting a t6500 as cpu, if all the voltages are comprised within these values does that mean I have no reason to do the mod?
4)Will the new voltages be effective without the need of software tool like rmclock? I mean: if previously the cou with multi 9x was using 1.25v ,would mean that now it will use 1.05 (given I applied the -0.2v mod) without usiing any software?(I plan to use linux..so no rmclock).

 

1) the wire has to remain inside the socket for the mod to work.
2) yes
3) it depends on the mod and on the CPU min and max vcore. What I stated in the first post is according to the CPU I used for the mod. Yours could be the same as it could be different.
The aim of the mod is to access (low) voltages that can not be normally accessible by an undervolting software such as RMclock.
4) Yes if the mod works your CPU will use 1.05v instead of 1.25
Check the posts from "Daytona 955i" in page 27 to 30. He did the mod on a laptop running Linux.

 

Do you know if these mods can be applied to the socket BGA956 which is for core Duo cpu with low voltages? (like intel su4100 or celeron 723)

 

Can BSEL mod the FSB on the su4100 via the PLL FSLx pins as described here. That's how m11x and Asus ULx0VT achieves their 1.3->1.73Ghz overclock.

 

Thanks for the reply,I have a laptop lenovo ideapad u350 with celeron 723, chipset gs45 (4500M) and 2gb DDR3.
The mod you posted are for laptop with ddr2, do you think the fact the mine mount ddr3 doesn't make a difference?

 

If you apply the FSLx+BSEL mod to your celeron 723 and it works, then it would be running 1066Mhz FSB. As long as your DDR3 is 1066Mhz, then can attempt this mod with some level of confidence.

 

What is not clear to me is that the method will involve;
1-flashing the DDR2 RAM to PC2-6400 spec so has a 400Mhz SPDtable, but since my laptop mount ddr3 1066 , am I required to follow or jump this step?
2-Can you post an impage of the type of resistor I should apply? I searched over the internet but I 'm not sure about the exact type I should use. Also , do you know a tutorial on how I should phisically apply it? On the thread you posted unfortunately there are not practical explanation about that, not even a picture with the job done.

 

I'm not an expert but I would say that you don't have to mess with the memory timing since you're using DDR3

Check this thread http://forum.notebookreview.com/showthread.php?t=393027. It has a lot of info about PLL mod. Also you can post there if you need any additional support

 

hello Natan,

1. could you please attach the pdf guide again? i tried to access it but it would not work.
2. i have tried the vid4 mod (socket m, t5600) anyways but the system works unstable. although i have figured out the according voltages and did the prime test i get bluescreens after a couple of minutes. any ideas why?
3. maybe i did the mod wrong. i am not completely sure whether the cable also has to have contact with the vid5 pin...I have tried to make it pass between vid5 and vid3 (according to pictures for socket p) and put some nail polish above the cable in this area in order to serve as insulation (avoid contact with either vid5 or vid3). any better suggestions how i should do it?

thanks for replying,

friedemann

 

is it safe to replace Aspire 4315 processor Celeron M 550 533FSB to C2D 2.2GHz T6600 2M 800FSB ?
if yes, is it run in full speed?
thanks

 

1- The first page of this thread contains all the material that was on that PDF and more.

2- BSOD means that the mod was successful but that the voltage you've selected for your CPU highest multiplier is too low. After VID4 the range of VIDs accessible to the CPU is limited. With VID4 you get the following:
RMclock selected VID ---- CPU VID
1.1125 - 1.3000v -------- remain the same
0.9125 - 1.1000v -------- drop to 0.7125 - 0.9000v

So to get multiplier x11 stable you need to select 1.1125v in RMclock

3- As I said before BSOD means that the mod is fine but your CPU is not getting enough voltage. What you can do now is:
a. keep VID4 and select 1.1125v for x11, and 1.1000v or below for multiplier x10 and below.
b. try VID3 mod

Can you list the VIDs you have selected for every multiplier in RMclock before the mod?

in my friend's Aspire 4315 I did the upgrade from a Celeron M 530 to a C2D T7300. I've also tested a T2370 on this laptop.
in My Aspire 4315 I did the upgrade from a Celeron Dual Core T1400 to a C2D T5250

Based on the above I'm sure that the Aspire 4315 can be upgraded all the way up to a T9500 . This laptop should support the T6600 but I don't know if its BIOS supports it. The only way to find out is to test it or you can upgrade to a T7300 (2Ghz and 4MB L2 cache) insead. The T7300 cost almost half the price (about £35 in ebay) of a T6600 (about £65 in ebay).

 

thx for your quick reply naton!

1. i get the bluescreens although the voltages should be fine. here is what i managed to run (but very unstable)

1,0 ghz (multiplier 6)-------cpu vid 0,8v
1,16 (7)------------------------------------0.8375v
1,33 (8)------------------------------------0.875v
1,5 (9)--------------------------------------0.9v
1,66 (10)----------------------------------1.125v
1,83 (11) ---------------------------------1.125v

it should run stable with those voltges! the voltages before the mod are 0.95v all the way through up to 1,83 ghz.

i guess i did something wrong regarding the mod itself. this is why i would always get the bluescreens. i managed to connect vid4 and vcc. my question is if i have to insulate the cable for the part of the cable that passes through the first row of holes between the vid 3 and vid 5 holes of the socket? in your picture of the first page of this thread it looks like you have not used any insulation material. this means that the cable could either get in contact with the processor pins for vid 3 or vid 5 or both of them. is this desireable? if not, how do i insulate it? i have tried to put some nail polish on that critical part of the cable.

 

So you selected the following in RMclock:
x06 ----- 1.0000v
x07 ----- 1.0375v
x08 ----- 1.0750v
x09 ----- 1.1000v
x10 ----- 1.1250v
x11 ----- 1.1250v
right?

is you CPU stable with x06 @ 1.0125v? maybe all you need is to up the voltage a bit for multiplier x06 to x09.
is x10 and x11 at least stable with 1.1125v?

I doubt it since you've even isolated the wire with some vernish.
What are you using to stress your CPU? Orthos?

You're right I didn't use any insulation when I did the mod. That's said I had to redo the mod couple of time to get the wire in a position were it didn't touch VID3 or VID5. If the wire touches VID3 too the voltage will drop by 0.3v. This will make the system unstable. It it touches VID5 the voltage will drop by 0.6v. your laptop will likely refuse to switch on with VID4 +VID5

 

Thanks for this thread, guys.

I've been tuning the under-volting on my ThinkPad T60 with T7600 CPU, and I've been finding that at x14 (maximum) speed, I need 1.0625V to keep it stable (default for that multiplier being 1.175). At all other (lower) multipliers, it is totally stable at minimum settable VID (0.9500V)

So, I tried the -0.1V pinmod.

My findings are:

1) No software notices this, and shows standard VIDs (CPU-Z, Core Temp, NHC, etc.)

2) The voltage drop actually appears to be 0.05V, rather than 0.10V. I am inferring this from temperatures and stability (tested with OCCT, VIDs set with NHC). If it was -0.1V, then I should need 1.1625V VID set to keep it stable (pre-mod was 1.0625V). I actually need to set it higher, to 1.1125V for same temperature/stability as before at x14 multiplier (2.33GHz). This is 1.0625V+0.05V, not 1.0625V+0.10V as implied by the pin mod. Has anyone else observed this? Is this specific to the T7600?

3) All other (lower) multiplers, I can STILL run 100% stable at minimum settable VID of 0.9500V (which if my inferring is correct above, is actually 0.9000V now). It's stable and cooler. This implies a -0.2V pin mod might be useful (even if it turns out to be an actual -0.1V pin mod in my case). But I doubt I could ever manage to route the wire past all those pins without shorting them out. On the outer row, that is reasonably straight forward. But across rows? I just don't see how that won't short out things on the way.

 

Try enamelled wire, baring the ends.

 

With VID3 + Vcc some of the Vcores changes, some remain the same, and some are lost:
1.1125 to 1.2000 ----> remain the same after mode
1.0125 to 1.1000 ----> drops to 0.9125 to 1.0000
0.9500 to 1.0000 ----> remain the same
here the range between 1 and 1.1v is lost.

If the mod is successful, your T7600 will require setting RMclock to 1.1125 or more to run stable at x14. With such Vcore you should notice an increase in temperature compared at the initial Vcore of 1.0625.

After mod, using a Vcore of 1.1000 or less will crash your laptop. I had a T7200 before, so if I'm guessing well in your case without the mod x10, x11, x12, x13, and x14 require more than 0.95v to run stable. This in itself makes the mod obsolete in your case because after mod you are likely to end up overvolting you CPU to run it stable.

Sorry to say that but your inferring is incorrect. Selecting 0.95 after the mod is the same as selecting it before the mod. Take a multiplier that you know stable at 0.95 and see after the mod if it is still stable with 1.0125. If it is stable that would mean that you didn't do the mod properly.

you can use nail polish to isolate the wire. This should avoid any short out

 

Thanks, that makes sense. I did notice a slight increase in temperature, but it's still about 10C cooler than at the default 1.1750V. I just wasn't sure since there is no VID3 mod range table listed on the front page, only VID4.

Ah, that makes sense. I'll try 1.0125V at <= x12. My T7600 was completely stable at <= x12 at 0.95V. I am happy to sacrifice slightly higher voltage at x14 for further reductions at <= x12, since when on batteries, I limit it to <= x12, and 90%+ of the time in normal use, the CPU is idle anyway, so the savings at idle or low multipliers are (theoretically) more important since they are used more often.

Or that my CPU is actually stable at 0.9125V, which is plausible, since it is stable at 0.95V at x12. If it isn't stable at x12, it probably will be at <= x11. Since 1.1125 is stable, but I get an insta-crash below that, whereas before I was stable at 1.0625, that implies that the mod is, in fact, successful.

I found that the voltages required seem to drop quite dramatically, even just a couple of notches down the multiplier range (1.0625V @ x14 and 0.9500V @ x12).

An interesting idea. It's just that on laptops everything is just that much smaller, and I have to say I'm concerned with my ability to do this without shorting something out with my fat fingers. :-/ The outer pins were just that little bit safer.

 

naton, many thanks for your input. My findings are showing that the pin-mod was indeed successful.
At x12, I need 0.95V. Both 0.9500 and 1.0500 settings work fine. Setting to 1.0125 errors out in NHC pre-check, and 1.0250-1.0375 causes errors in OCCT almost instantly.

At x11, my T7600 is stable when I set it to 1.0125, which should be equating to 0.9125 now!

What I'm wondering now is whether it's worth trying to go for the VID4 mod. What range will be removed, and what will be replaced with default-0.2V? It looks very much like I could get away with further reduction of voltages at 6x-11x multipliers.

On a separate note, is there any software that can read the new voltages correctly? NHC, RMClock, CoreTemp, OCCT, and everything else I tried is showing the default VID equivalent values rather than actual voltages.

 

I've noticed that too. There seems to be a relationship between heat and voltage requirement. The cooler a CPU is the less Vcore it needs to run stable and vise versa.

Definetly worth a try. You seem to have an amazing CPU. With modded VID4 x6 to x10 would run at 0.9v and below -possibly x6 will be stable at 0.8-0.85v- and x11 to x14 would run at 1.1v

This was discussed quite few times in the forum and it seems that no software can read the real Vcore after mod. Few pages back a member of the forum used a voltmeter to read the voltage before and after mod. I'll dig the page and post back.

EDIT:
Read the posts of Daytona 955i in pager 28

 

OK, I think I have acquired some sufficiently thin insulated wire, will try it.

Unfortunately, I didn't quite grok what the conversion table for VID4 is. Which voltage ranges get changed to what with the VID4 pinmod? Can you clarify? This should really be edited into the first post in this thread...

 

1.1125 to 1.2000 ----> remain the same after mod
0.9500 to 1.1000 ----> drop to 0.7500 to 0.9000

With a VID4 mod the Vcore range between 0.9125 and 1.1000 is not accessable.

 

Hmm... That removes at least the reduced voltages for x14 (1.0625) and x12 (0.95) for me, and possibly even the x11 (coudn't test below 0.9125). I guess if x11 ends up working at 0.90V I could just about live with that if the reductions at x6-x10 are significant...

Only one way to find out.

I don't suppose there is a way to drop the top band (1.125-1.200) by 0.1V?

 

unfortunately you can't have -0.1v for the upper voltages and -0.2v for the lower voltages at the same time.

Most applications don't require a fast CPU. So you can lock yours to use up to multiplier x10 (@0.9000v) with the "performance on demand" profil, and multiplier x14 @ 1.1v with "maximum power" profile, and then switch between the two profiles as required.

 

Is there a table somewhere detailing the VIDs corresponding to voltages? I'm trying to do this on Linux and with the PHC functionality, I can only see VIDs, rather than voltages. For example:

# cat /sys/devices/system/cpu/cpu0/cpufreq/phc_vids
37 33 28 24 19

# cat /sys/devices/system/cpu/cpu0/cpufreq/phc_fids
14 12 10 8 6

The FIDs are multipliers, but I can't figure out what voltages the VIDs correspond to. and it's kind of important to know since I'm running pin-modded... Has anybody got some relevant info?

 

This?

 

That looks like it, but the figures don't seem to quite tally up. Here are the defaults according to the kernel driver:
$ cat /sys/devices/system/cpu/cpu0/cpufreq/phc_default_vids
37 33 28 24 19

That would make the default of 37 = 0100101 = 1.0375 V when NHC and RMClock think default is 1.1750. Since this table doesn't go that high on voltage, I'm guessing it's for a different CPU (I have Merom C2D T7600).

 

I've removed my pinmod for now because I couldn't quite figure out the correct VID translations, and I couldn't find a suitably thin insulated wire for the VID4 mod. I have, however, been doing some power drain testing, using the on-board sensors.

At idle, with screen at full brightness, the battery discharges at about 27W. Given the CPU is running at 0.95V, my TDP at x6 should be about 8W, not accounting for the fact that when it's truly idle, it goes down to a fraction of a W. Which means that at least 20W goes on the rest (including an SSD that uses 75mW idle, 150mW peak). That's with the Bluetooth and WiFi radios both off.

So it strikes me that I am right up against diminishing returns with further CPU power savings. Most of those other 20W seem to be used up by the screen and the NB/SB, and I don't think I can do much about those. :-(

 

I m new on this forum, but after read a lot about OC&OV in this thread, pinmod my XPS M1330 to FSB =1066 and reach 3324 MHz with a T9300 (ES) see : http://valid.canardpc.com/show_oc.php?id=1242127


Thank a lot for share you knonwledge

 

Hi people, i have a Sockep P T4300 @ 2100mhz, and i like to do the 200 to 266fsb mod, someone have a picture of the socket with the mod ? Sorry am new here, and sorry for my bad english.

My chipset is a PM45 + ICH9

 

Amazinnn ............ curious what PLL is in your XPSM1330? thanks

 

It's a ES T9300 CPU. So means no lowest-multiplier lockout when BSEL cpu pinmod it from 200/800Mhz->266/1066Mhz operation. Normal T9300 CPU will not let you do that.

 

thanks for info.. would it be same with X9000(ES) compared to X9000(normal) to pin mod from 200 to 266?

 

Probably want to read/try: http://forum.notebookreview.com/acer/489140-how-unlock-core-2-extreme-multiplier-windows.html first.

 

thanks again, ive already using ThrottleStop 253 ...actually pushed to 21x .....but doesnt increase to 266 .. so can pin mod 200 to 266 an use ThrottleStop?

 

Since you have a non-ES CPU I believe you'd get multiplier locked out when applying BSEL pinmod. Could be wrong. Could always revert to a FSLx PLL pinmod to get a 266Mhz FSB. Has been done on a 965PM chipset successfully. Likely running a 2.8Ghz CPU at 4.2Ghz as you are now is the the edge of cpu stability unless you have some crazy good cooling. Only advantage of increasing from 200->266Mhz FSB would provide faster RAM/video response with CPU multiplier pulled back to x15/x16 (4Ghz/4.27Ghz) from the current x21 to provide CPU stability. Is it worth the trouble? 4.2Ghz@200Mhz FSB is pretty impressive, if stable.

 

trying to run benchies @ x21 > but can only run Superpi1m an Pifast successfully, otherwise @ x21 is unstable....so hope on 266Mhz will help keep it stable just enough to finish other benchies is me goal.....cooling is great in XPSm1730 as long as heatsink internal fins are dust free & gpu,s an cpu fans on 100% speed ......winter here .....can get some nasty cool roomtemps
but if have to can tinker around with extra cooling

about the FSLx PLL pinmod ... need time to build up on me soldering skills(im still beginner) before attempting the impossible the PLL Legs have paper width gap between each other an going to require precise expert skill to mod > me usb microscope set up is going to be handy for checking

 

M1330 have SLG8LP550V, cut circuit and force pin 45 (BSEL1) to GND with a 5k6 resistor. this change to 266 MHz, normal socket wire mod not working.
First time run wit a t7300 ( see http://valid.canardpc.com/show_oc.php?id=1234099 )
Then app a VID 3 mod (VID3 to Vss) and run a T9300 (ES) as sowh on previus post.

 

Not, I try with original processor T7300, and work right with FSB to 1066,
I open circuit betwen PLL to chipset and CPU and pull down pin 45 (FSEL1) to GND w/5K6 resistor.
Prior to do this not working.

NOTE: RAM running at 880 MHz. I have a 800 MHz OCZ and working well and past all exaustive test.

 

Hi: today load Trottlestop and enable as Exterme CPU and reach multiplier x13.5. FSB=3590 MHz
See: CPU-Z Validator 3.1

 

What CPU is that?

 

XPSm1330 with a T9300(ES) + cas5 800Mhz ram + PLL MOD 200 to 266Mhz + ThrottleStop(v2.53?) = fantastic OCing

uXsPeSr..... do ya think throttlestop would work with T7300 (non-ES) + cas5 800Mhz ram + PLL MOD 200 to 266Mhz , would be really interesting to see if can set to x11 ,
anyway thanks heaps for posting, hoping to get similar result in XPSm1730 > prob is me PLL CY28547 may not support 266Mhz?

 

If you see in post#343 I use 266 MHz with a T7300@2660 MHz, but not test with throttlestop. The previus mentioned M1330 PLL is the same CY28547.
On M1330 schematics show CY28547, but pinout are same. Pull down pin 45 to GND with a 5k6 resistor and seprate cirrcuit betwen chipset and PLL (cut trace),
and not need VID mod. I can help If you upload picture of PLL and oposite motherboard face with good resolution.

 

Is a T9300 QADW (ES) stepB1

 

I tried the BSEL mod with a T7300 and a PM45 express chipset, as expected it was locked to the 6x multi:
http://forum.notebookreview.com/har...ock-core-2-extreme-multiplier-windows-15.html