BSEL Mod on a socket P explained with photos

Has anyone tried the FSB mod or the voltmod on a laptop with socket P and an nVidia chipset?

 

Well, I OC the Celeron 530 from 1.73 (13x133) to 2.16 Ghz (13x166), It couldn't boot at 200 FSB. I think it need a little more voltage, but this laptop gets hots easily so I think i will stay with that config. Anyway I am very satisfied with the results, the laptop runs like butter.

Thanks Naton.

And for chipsets Nvidia, I bet it won't block multiplier as ATI chipsets. Nvidia in general is very overclock pro-active. Even if isn't, you could always use Ntune.

Regards

 

Got my Celeron 900 to 2.93Ghz. Stock it is the latest stepping 45nm 3MB cache CPU with one core disabled running an 11x multiplier off of a 200Mhz FSB. Thankfully it lacks speed stepping. To get to 266Mhz FSB I had to rip off the BSEL 1 pin because I had no idea how to block it. Works great. I test out some other FSB settings using the traditional method just to see if it would work, but you can't go higher than 200Mhz on a laptop without blocking or removing that pin somehow. Connecting it to ground would cause a short and possibly fry the cpu. I still have a lot of hard work ahead of me. I have to lap the heatsink, which fortunately has a copper base. I also have to lower the voltage. The problem is that with this processor it must be done manually at the socket because it doesn't support speed stepping. The default is 1.175. I will have to rip out one or more additional pins unless someone can show me a way to block them from connecting in the socket, so I really can't go back if I make a mistake and take it too low. I can make any low to high reversibly, but I can only make a high to low one time and can't go back. =( I will try though, but not today, as I have to solve an engine problem on my car.

 

Thanks for the info. Now I need to find the right laptop .

 

I don't understand why did you had to break the pin off your Celeron. I read in few other websites that connecting the BSELs to the ground is dangerous. From my own experience it is not. I tried it with:

a T2500 (FSB 166 to 200 - problems: stable by temperature reaches 80c)
a T5300 (FSB 133 to 200 - problems: not stable and too hot)
a T7200 (FSB 166 to 200 - Problems: fast and stable by temperature reaches 90c )

I think you shoud have tried to do what's on the photo below instead or removing the pin to go from FSB 200 to 266.

I'm not sure about lapping. If you search the web you'll see that some peaple clame that they saw over 5c reduction after lapping both the CPU and the heatsink, but others claim that the reduction is only 2 to 3c. I'm more keen to think that it is only 2 to 3c.

DON'T BREAK THE PINS. Try to pin mod with wires instead as shown below.

VID6543210 for:
1.1750v is 0 011010. Your best option is to try 1.0750v
1.0750v is 0 1 00010

 

Mate

I did Mods in 5 different CPUs and 3 laptops and haven't fried neither laptops nor CPUs. I used copper wires for allof them, and have never broken any of the CPUs pins.

FSB mod:
Acer AS 4315:
Celeron M530 from FSB 133 to 200 (BSEL0 + Ground (Vss) and BSEL1 + Vcc)
Pentium T2370 from 133 to 200 (speed was locked at 1.2Ghz because of the intel chipset)

Gateway MX3103b:
Core Duo T2500 from FSB 166 to 200 (BSEL0 + Ground (Vss))
Core2Duo T5300 from FSB 133 to 200 (BSEL0 + Ground (Vss) and BSEL1 + Vcc)
Core2Duo T7200 from FSB 166 to 200 (BSEL0 + Ground (Vss))


volt mod
Ei System 3103:
Pentium T2060 min voltage from 1.025v to 0.825v
Core2Duo T5300 min voltage from 0.95v to 0.75v

Besides me others have tried and it worked for them too:

Tinselworm - link

netc0rd - link1 and link2

tarquini - link

DooMaster - link

Have you tried to remove the wires used for the voltmod and tested your CPU again?

 

Here is some input from me, I couldn't do it earlier, sorry.

T5750, wired for -0.2v. For all except the lowest one (6x) this is the ultimate stable voltage:



Here is how it was before:



Voltages aside I saw a little to no difference in the temperatures. Tomorrow if I have more time, I'll try how the T9500 would react on this

 

Yes, you are. There is no problem on the meroms, but I have (had) a penryn. You said that the people saying it would fry the cpu were wrong and that you have done it fine. You obviously haven't, but quelled my fears and convinced me to fry my chip.

 

WOW, now I'm a little bit concerned for my 1 month old Penryn

 

I pinmodded my old X9000, T9500, Q9200 before with direct contact to VSS and VCC, all are fine, none dead. Don't know why yours is so.

 

Trottel

Forget about me for a second and think for yourself. Volt mod does only rise or increase the voltage. Correct?

1- The highest voltage that a socket M or P CPU can reach is 1.5v, and this voltage can only be achieved if you connect ALL the VIDs (i.e. from 0 to 6) to the ground (Vss). I don't think you did that, and even if you did, your laptop would have started and worked for a bit before switching itself off to prevent your CPU from frying.

If your CPU is dead it is more likely that you damage it when you pulled the pin out.

2- All the people that said never connect to the ground where talking about connecting BSEL to a Vss. They never said anything about connecting a VID to a Vss.

3- As I said forget about me for a second and just google BSEL Mod. You'll find 100s of links showing people connecting both BSELs and VIDs to both Vcc and Vss. So it is SAFE

example: link


4- How is your laptop? I mean is it OK?

 

Overvolting or undervolting?
What's the highest temperature you reached under load?

My T7200 @ 2.4ghz on a hot day reaches 90c with RMclock (undervolt) and AS5 . too hot for my liking.

 

There is nothing to worry about. Undervolting with pin mod is the same as doing it using RMclock.

Listen there is something wrong with your mod on the T5750. After the mod:
voltages between 1.1125 and 1.3000v remain the same
voltages between 0.9500 and 1.1000v drop to 0.7500 and 0.9000v

So, if the minimum is 1v for x12 before the mod, this same value, which is 0.8v after mod, should make your laptop crash when running orthos.

did you use an insulate wire?

 

I'd put a little Kapton tape where the wire passes between the two pins.

As I said - 12x @ 1.1000v crashes, so do 11x and 10x and I found that weird too Maybe there is something wrong, after all. I'm still running T5750, so I'll check it, today or tomorrow and I'll let you know

 

Following the recent threads really made me worry about oc'ing or v modding...

it should be that short-circuit occured that caused the disaster...

damn... AMD should really catchup so that intel think twice to do any locks on any cpu!

 

dear trottel,

i am curious to what changes you did after the fine 200 - 266 mod that caused it to be dead ? a volt mod or 200 - 300 mod (unlikely) ..

 

Yes, I agree about the voltage. It couldn't die from over or under voltage on the laptop, assuming it is working properly. When I twisted the pin off, the solder point was fine and still on the chip. The processor also worked for the day or two after I did that, and I continued to try all of the different BSEL pin possibilities.

The laptop works fine after I put in a new processor.

Sorry I got mad at you, I obviously did something wrong, even if I can't figure it out. I'll try this again later today to see if I can make it work. I think 2.9Ghz, stock volts, and 65C at full load was pretty good, even if it was a single core.

 

So shiver me timbers, but the Celeron lives.

I got a replacement CPU today, and I put the T4200 back in its proper laptop, and the new processor in the Celeron's old socket. The new processor seemed dead. The laptop would power on and then just sit there. I tried the T4200 in that laptop again and it worked. Then I tried the new CPU in the other laptop and the same thing; it would power on and play dead. Then for some reason I tried the Celeron in the other laptop and it works. Really weird. The cpu that wouldn't boot in a laptop that another processor works fine in boots up and is stable in another laptop? IDK what is going on or how that happened. I'll conduct further tests tomorrow.

 

the Asus uses a x38 chipset or some thing if i remember correctly

EDIT : yep

i have a weird memory

 

I agree.

no problems.

Trottel do you mind updating or removing this post so it doesn't confuse the readers.

was your new CPU a Celeron M900 too?

just an idea, did you reset the CMOS battery before testing the new CPU in your laptop?

 

I'm a fool I was looking at the same picture when I said that your mod didn't work.
Your mod from an undervolting perspective has worked fine. Your temperatures from x10 to 12 should higher than before, and temperatures for x6 to x9 should be lower. Correct?

 

No, the new CPU I bought was a dual core Merom. I guess I'll try that, but it won't boot in either laptop. A Merom-based Celeron should be here real soon, so I'll try that and if it doesn't work, then I'll try these two chips in the only different socket P laptop I have access to, a Toshiba.

I really want to try the VID mod again, but I'm scared to even though there seemed to be no permanent damage. That was extremely weird.

 

From what I learned after posting in "
What Notebook Should I Buy?" section is that none of the socket P laptops uses ATI chipsets. Also from serching the net I couldn't any laptop that uses nVidia chipsets too. It seems that all the present socket P laptops use intel chipsets. intel.
That's to say that if you want to try the BSEL/FSB mod you need a celeron cpu. A dual core celeron should work too though I didn't test it myself because dual core celeron(s) use a 200 fsb and my friend's acer laptop (the one moded in the first page) has a maximum fsb of 200.

 

just wondering if this pin mod is applicable for t7250 on pm945

 

Wow great work guys.

I have some cpu's that i would like to overvolt using the Asus W90VP. Setfsb will oc the FSB 320 and over depending on the CPU.

Cooling is not a problem I have a custom setup so CPU temps dont go over 40oC under full load.

I have the following CPU's that I would like to increase the voltage say around 0.3volt:

T9400
T9550
T9600
T9800
T9900
X9100 E0
X9100 C0
QX9300
Q9100
Q9000

Whats the best pins to join

 

the volt mod will work in any system
the FSB mod will work in the follow two cases:
1- any chipset + celeron M
2- any intel cpu + none intel chipset

Go to the first post and use the diagram for undervolting socket P, but instead of connecting the VIDs to the closet Vcc (like I did) you have to connect them to a Vss.
VID3 + Vss = +0.1v
VID4 + Vss = +0.2v
VID5 + Vss = +0.4v

 

@nation, I have a T7300 and the PM965 chipset.

I want to bump my voltage up by 0.2v, should I just connect VID4 + vss as you say ^^^?

Also can I combine them, connect VID5 to Vss and VID3 to Vss will give +0.5v?

 

Max VID possible is 1.5V.

 

my default VID I believe is 1.0625V.
What would happen if I connected VID5 to Vss and VID3 to Vss.
would I just get 1.5V?

Edit: also to get +0.3V, would it be by connecting VID3 + Vss and VID4 +Vss?

 

default vid at 1.0625v means that:
VID6-5-4-3-2-1-0 are
--- 0-1-0-0-0-1-1

So basicaly:
VID 2 + Vss = no cahnge in VID
VID 3 + Vss = no cahnge in VID
VID 4 + Vss = no cahnge in VID

your only option is VID5 + Vss with lead to a VID of 1.4625v, which is a lot. I mean your CPU will be run very hot. You might even damage your CPU.

How about +0.2 v instead (i.e. 1.2625v). You can get that by moding both VID4 (-0.2v) and VID5(+0.4v) at the same time:
VID4 + Vcc = -0.2v
VID5 + Vss = +0.4v

I don't know what you're trying to achieve. If I am you, and have an unstable overclock I'll start by increasing the vcore by 0.1 first. To do that pyou have to mod VID3, VID4 and VID5 at the same time (assuming that your default VID is 1.0625)
VID3 + Vcc = -0.1v
VID4 + Vcc = -0.2v
VID5 + Vss = +0.4v

 

moral hazard

I'm not sure what would happen if the temperature goes above 100-105c. I think your CPU will slow down to keep it self cool. If the temperature reaches 125c your CPU/laptop will turn itself off.

If there is a (minor) default of fabrication in your motherboard or CPU the above described process may not work and you may end up with a dead CPU. That's to say if I'm you I will stay away from +0.4v.

 

thanks for the help.

I think this is the best for me:

The reason is that my notebook is unstable when I take my FSB anywhere above 272mhz. My CPU is 2.72ghz at that point and my ram is over 900mhz.

I'm not sure what's failing first (CPU or ram) since both are being stressed a lot. I had to increase my memory timmings a lot. BTW I have 800mhz DDR2 ram.

I'm going to get new ram and flash it to a lower frequency so that it will be stable. Then I'll just push the CPU to 3ghz. So +0.2V should be enough for me.

 

It is likely that 0.2v is enough but I not an expert in overclocking.
Out of curiousity what's the temperature of your CPU at 2.7+ GHz under load?

 

from memory it's ~88C, I'll run a proper test on the weekend though. Also I will be doing a cooling mod to my notebook.

Edit: the Tj max for my CPU is 100C.

 

So I'v just check RMclock and it lets me increase the voltage by almost 0.2V.
Also by my calculations I would need +0.3V or +0.4V to get to 3ghz.
So it's either:
VID3 + Vcc = -0.1v
VID5 + Vss = +0.4v

Or just:
VID5 + Vss = +0.4v

I know your advice is to stay away from +0.4v, but it's kind of hard to do +0.4v and -0.1v.

I'v got a photo of my socket, can you show me what to connect for -0.1V and +0.4?



Sorry for the bad quality image.

Edit: here is what I think it should be:

 

Is that pic your PM965? or just a pic from Internet? Seems to be odd.....

 

It's a pic I took of my actual socket. (not from the net).

If you want I can take one with better focus?

 

This is exactly what my PM965 looks like, without 2 pins in the corner.

 

myne also is without one pin in the corner, look at the bottom left of it.
This is strange.

 

while my CPU is actually missing two pins, here is a photo:

 

It looks fine to me.

For the socket and the missing pin(s), the one on the Acer I used in the first post has only one missing hole. I was able to insert a t2300 (socket M) inside it (inside a socket P), but did't power the laptop.
By that I mean that the socket M and P are similar; they have the same number of holes and the distance between the holes is the same. The difference is on the pins layout of the CPUs.

 

thanks for the confirmation. +rep

BTW I did do a test @2.7ghz with orthos running for 5min it leveled out at 90C.

I would have to fix the cooling for this notebook and also I need some new ram to get to 3ghz.
I will wait for the 4gb sticks of ram to come down in price. Or untill I get some ram that I can flash.
The only stick that I have which I can flash just got bricked. I tried to flash some really high timmings and now it doesn't work. I really need to find some good ram.

I made some rough calculations and I would need to run at least another 0.33V to get to 3ghz. So the only way is with +0.4V.
I'll decide a bit later, when I do the mod I'll post results here. (probably not going to be soon).

thanks again naton. great guide.

 

I knew that the T7300 is a good CPU. I was planning to get a laptop that can go to 1066mhz (FSB) and then pin mod the T7300 to run at 2.6ghz instead of its native 2ghz. Only problem is I can't find a laptop with socket P that used an nvidia or an Ati chipset. With intel chipsets the mod doesn't work.

what do you mean by flashing the ram?

 

You can use software to change the SPD of your ram.

you can either use SPDtool or thaiphoon burner.

Both are really great tools.

you can change the frequency, timming and voltage of your ram.
Be carefull with those tools, you have to make sure your ram can handle the new frequency or timmings. Also I don't suggest changing the voltage of your ram.

If you flash bad timmings onto your ram (i.e. the ram doesn't work), you will only have a small chance of saving it.

Here is a link to thaiphoon burner:
http://jungle.at.tut.by/download.html

If you want SPDtool, just google it


Edit: If you have any problems with SPDtool or thaiphoon burner (e.g. they dont "see" your ram or any SMbus error) let me know, I can help.

Also some ram has write protection enabled. If it is software write protection then it should be easy to get around. If it is hardware WP then you really can't do anything. I did try to remove HW WP on one of my sticks of samsung ram, but it didn't work. here is the thread if you're interested:
http://forum.notebookreview.com/showthread.php?t=408349

 

Why don't you try a PLL pin mod.
It works with every chipset.

I can help you out with it if you want.

 

hello, but p965 chipset ( i have a t9500) can go at 1066?

 

yes it can, but you have to modify your PLL (clock generator).

You may also have to use SPDtool to flash your ram to make sure it will be stable.

I can give you instructions if you want.

 

I need to get a socket p laptop first, and read again your thread that explains the PLL the mod. If I deciced to go for it I'll contact you if I need any info.

thanks for the links

 

Did the .1 undervolt with a p8400 in my msi 1651. I tried the .2 undervolt and the system wouldn't even boot up, might have been touching more than the two pins.

 

Yes it is likely that the wire was touching more than one pin. you should try an insulated wire like the one used in electric engines. if you use such a wire, remove the insulant at the ends (i.e. the bits that are inserted inside the socket holes). You can do that witha knife.