BSEL Mod on a socket P explained with photos

CPU-Z is notorious for false voltage readings. I wouldn't trust it.

 

If you do the wire pinmod of Voltage, no more software can read the REAL voltage, so CPUZ will not change.

 

any particular reason why I couldn't run rmclock?

 

I guess, you have to reset it to default voltages and starting from there on.

 

Please let me try again.

let say that your CPU is a t5300. It can run stable in RMclock with a Vcore = 0.9500v (multiplier 13x).

When you did the VID4 and left RMclock with the old sitting you actually had multiplier 13x running with a Vcore = 0.9500 - 0.2v = 0.75v. Windows didn't like it, and hence your computer was crashing.

This is the way you should do the mod:
1- find out the Vcore used by all your multipliers in RMclock
2- If all your multipliers are stable at 0.9500v then do VID4 else do VID3
3- UP the Vcore in RMclock to 1.25v for all the multipliers
4- Do the VID4 mod
5- Go to RMclock and try to find again the Vcores for all your multipliers.

 

I did the VID3 and VID4 VoltMod for Socket P. It did change voltages, however the changes were opposite to the expected in every way.

When mod was supposed to decrease a voltage chosen from RMClock by 0.1V/0.2V, it did nothing. Since you can't find out the real voltage anywhere if mod does change it, I had to compare temperatures to unmodded CPU temperatures. In this case temperature didn't change at all.

When mod wasn't supposed to decrease a voltage, it actually increased it by 0.1V/0.2V. I figured this out because temperature rose by level of 0.1V/0.2V depending on the mod.

For example, choosing 0.95V from RMClock with VID4 VoltMod, it didn't change the voltage, temperature was same as without the mod. Choosing 0.9V with same mod, the voltage rose to 1.1V, since temperature rose exactly to the level of 1.1V.

Did I do something "backwards"? Of course there is a very small chance that mod actually did decrease voltages and not increase, but it made temperatures rise for some other reason...however I measured temperatures accurately and they showed very clear correlation with supposed increased voltages.

 

I am interested in doing this mod on my Alienware M15X that currently has an Intel T9500. Looking at CPU-Z my current FSB is at 200MHz, how would I raise it to 266Mhz ??? Thanks for any info. God Bless

 

Same as going from 133 to 200 by using only the small wire (the wire in the second row).

 

True, it doesn't make sense to what the mod should have done and that was the problem. It doesn't really matter because, as you said, I probably didn't do the connection properly in VID4 mod because I didn't use insulation.

As for the VID3 mod: I measured temperatures again and the temperatures were a little different this time, so this mod didn't do things "opposite" as I thought first. But the mod did change temperatures compared to unmodded.

For example, choosing 1.00V in RMClock the temperature rose to 58C and then choosing 1.0125V the temperature fell to 54C. Without mod the temperature at these voltages were around 51C.

I did this mod with 2 CPU's 2 times each, and I understand it's supposed to decrease those Vcores voltages that are accessible after decrease. It didn't work as supposed and it's hard to figure out what the mod did actually do. I can post the rest of the voltages and their temperatures but they don't make much sense. Could someone else try this mod to see if it works for them as it is supposed to?

 

How would i do this for my t5800. Please help me thank you

 

The above is with VID3 isn't it.
Vcore ----------- VIDs -----------
--------- 6 -- 5 -- 4 -- 3 -- 2 -- 1 -- 0
1.0000v - 0 -- 1 -- 0 -- 1 -- 0 -- 0 -- 0

As you can see for Vcore = 1.0v VID3 is already set by default to 1 (check the table). Hence, the VID3 mod won't decrease the voltage when it is set to 1v in RMclock. In other words when you use 1v with VID3 mod the temperature before and after mod should be the same.

Vcore ----------- VIDs -----------
--------- 6 -- 5 -- 4 -- 3 -- 2 -- 1 -- 0
1.0125v - 0 -- 1 -- 0 -- 0 -- 1 -- 1 -- 1

Vcore = 1.0125v has a default VID3 of 0. So after the mod the VID6-5-4-3-2-1-0 combination becomes 010 1111 instead of 010 0111. If you check the table this is equivalent to 0.9125v (i.e. 1.0125 - 0.1v). Hence the difference in temperature.

I know it's a bit hard to get use to this as it took me a while to figure it out myself. The rule is:
1- Any Vcores that has a default VID = 1 will have the same voltage in RMclock before and after the mod
2- Any Vcores that has a default VID = 0 will see a decrease in the voltage after the mod

 

I definitely understand that, what it's supposed to do, I thought I made it clear in my previous post. The example is with VID3.

The decrease in temperature (in 1.00V -> 1.0125) could be explained with decrease in voltage as you said (-0.1V, what the mod is supposed to do). However, as I said in the example, the temperature when unmodded at 1.0125V was 51C and 54C when modded. How is 0.9125V going to give me 54C and 1.0125C give me 51C? That was the problem. The mod didn't decrease temperatures at all, but raised them.

Also, the mod is supposed to decrease the voltage to 0.8V when 0.9V is chosen in RMClock. When I choose 0.9V in RMClock the temperature is 53C. Without the mod the temperature at 0.9V is 48C.

There's no bias in my temperature readings, I let them settle down for almost an hour, ambient temperature was same all the time and I measured many times.

 

this mod works for me i clocked my Celereon M 560 2.13 (FSB 133) @ 2.66 GHz (FSB 166). Stress temp 67C.

http://valid.canardpc.com/show_oc.php?id=593991

 

I´m a bit confused...

I have a MSI PR-200 (MS-1221) laptop with:

Intel T7100 (1.8GHz, 800Mhz bus)
GM965 chipset

1) will the mod work with it or due the intel chipset it won´t?
2) does anyone knows the PLL for this laptop? (MS-1221, PR-200 Laptop)

 

i have a inspiron 1525 with a celeron m540(1,86ghz - intel GM965, socket P), could you send me pics of your mod?
i'm a bit confused, because i opened up my laptop, and my socket is P and is purple(dark pink or something like this, i've seen it a week ago), and all schemes i've seen are all in white sockets, like naton(thread openner) in the first page, that claims he did the socket P, but in google i've found that his pics look like a socket 478 or something like that, and celeron 530 is socket p
i'm confused!
plz help me
(sorry for the bad english, i'm from Brazil)

 

The maximum that your chipset allows is 800mhz so you wouldn't be able to do it. Even if you could do it, since it's a santa rosa chipset it would lock the multiplier to the lowest value.

You should try to use setfsb and clockgen, if they don't work then you will need to do something simmilar to this:
http://forum.notebookreview.com/showthread.php?t=393027

 

1- What thermal compound did you had before and after the mod?

2- Before the mod, what was the minimum vcore that you were able to use with you CPU highest multiplier before your laptop craches? what's the Vcore that mad your CPU crash?

3- Same questions as 2 but after the mod?

Please answere the above so we can figure out why your CPU run hotter after the mod.

 

1) It had the same thermal compound that the this notebook came with. It is very dried out and almost gone since I have switched the CPU many times. I tested CPU temperatures last time without pin mod and they were the same as I tested them for the first time when I hadn't opened this notebook, so worn out thermal compound didn't make temperatures rise.

2) Without mod the lowest Vcore I could use without notebook crashing was 0.9250V (Tested 10min with Orthos). 0.9125V made my notebook crash immediately.

3) With mod the lowest Vcore I could use without notebook crashing was 0.9125V (Tested 10min with Orthos). 0.90V made my notebook crash immediately.

 

Is no one going to try the Socket P pin volt mod so we can confirm if it works or not? It took me only 10 minutes to do it every time. Naton do you have any ideas?

I'm quite eager to get this working, maybe I could have a silent (and cool) notebook at last. I will be very thankful if I get this mod working! If it's not possible I guess I have to use Socket M CPU, since the volt mod supposedly works in it.

 

Nope, pin volt mod is not working for me too

 

Which mod did you try on your T9500 ??? I have one too on my Alienware M15X and would like to overclock it. Let us know. God Bless

 

I tried to undervolt my CPU, for overclocking you should try SetFSB (it's not working for me though)

 

Maybe the thread in my sig can help?

 

Hi and sorry for the delay.
I did some invertigations and I think the rise in temperature is because the wires used for the mod make the CPU sits unevenly. This leads to the CPU die not been in full contact with the heat sink, and hence the rise in temperature.

I think I found a work around it which necessitate a bit of DIY. I'll post some pictures tomorrow.

 

1- The T9500 has an FSB of 200Mhz. Right?
2- Does your laptop support FSB 266Mhz?

If both 1 and 2 are yes you can try to run your CPU using the 266FSB insted of 200FSB. since this is a 33% increase you may have to up th voltage.

for undervolting I need to know your default min and max Vcore. They can be read in the CPU tab of RMclock.
I'll also need to know the min Vcore for all the multipliers used by your CPU.

 

Naton, congrats for the guide, I like it very much. I'm gonna try it for a couple of notebooks, I will post my experience in a couple of weeks. In the meanwhile, the specs:

- Compaq C700, Celeron 530 1.73 Ghz, Chipset Intel GMA965
- Vostro 1310, C2D 1.8 Ghz, Chipset Intel PM965 (with discrete Geforce 8400M GS)

The first one, i'll try the 1.73->2.6 ghz mod , and for the Vostro, 1.8->2.4 Ghz mod. Wish me luck !

 

I will try the work around. I used IDE wire for the mod btw.

But what about the crash vcores? Why did 0.9125V crash my notebook without mod, but not with the mod?

Also, choosing 1.0125V when modded, the temperature was 54C. Choosing 0.90V when modded, the temperature was also 54C. Both voltages are supposed to actually be 0.1V lower when modded.

 

I have a good and a bad news. Since both your laptops use an Intel chipset you will be able to FSB pin mod the Celeron M only. The Core 2 Duo if FSB moded will run at 1.2Ghz instead (i.e. 200FSB x 6 multiplier).

 

I'm not sure are this is realy confusing:
1.1- With ViD3 1.0125v = 0.9125v which is about the same as 0.9000v. Hence more or less the same temperature (when rounded up).
1.2- have used the same multipliers for 0.9000v as for 1.0125v? or are you talking about a different multipliers?

2- you can use normal electric wire. It seems to work (see nest post).

3- It seems that I'm lacking some information. Can we start from the beginning:
3.a- what laptop are you using?
3.b- what chipset do you have (intel, ati, nvidia)?
3.c- what cpu are you using?
3.d- what is your cpu min and max voltage? (as shown in the 'CPU info' tab in RMclock)
3.e- your CPU minimum Vcore for every single multiplier when you're using RMclock.
3.f- in RMclock do you set the Vcores individually for every multiplier? or do you set the highest and the lowest multipliers only, and let RMclock adjust the intermediate multipliers?
3.g- after the mod was your CPU sitting flat on its socket? or was it a bit listed?

To make it easier for me, can you replay to all the above questions one by one. It would helpful if you just copy the above message and add your answers to it using a different colour (e.g. red)... thanks

 

Socket M FSB mod:

Experimenting with a Gateway MX3101b laptop and a C2D T5300


Experimenting with a Gateway MX3101b laptop and a Core Duo T2500

 

I see that Socket M hasn't the Mutiplier lock, but, even with RMClock cant change the multiplier?, deactivate Speedstep? Even with that problem, I would like to test.

 

the lock has nothing to do with the type of socket. The lock is related to the type of chipset used. In other words:

Intel chipset (i943, i950, i965 ...) = lock
Ati chipset = no lock.

Since the Celeron M uses only one multiplier, the mod works even with an intel Chipset based laptop.

hope this helps

 

Mutiplier lock depends on chipset. All intel mobile chipset 945, 965, 45 lock multiplier when you try bsel mod on cpu socket. It doesnt metter wchich socket have cpu- M or P. This particular gataway laptop have ATI chipset where locking multiplier not appear.

 

I'm not sure are this is realy confusing:
1.1- With ViD3 1.0125v = 0.9125v which is about the same as 0.9000v. Hence more or less the same temperature (when rounded up).
Isn't 0.9000V supposed to actually be 0.8000V in VID3? So there's a difference.
1.2- have used the same multipliers for 0.9000v as for 1.0125v? or are you talking about a different multipliers?
I used same multipliers.

2- you can use normal electric wire. It seems to work (see nest post).

3- It seems that I'm lacking some information. Can we start from the beginning:
3.a- what laptop are you using?
Thinkpad T61 model 7661
3.b- what chipset do you have (intel, ati, nvidia)?
Intel GM965
3.c- what cpu are you using?
Intel C2D L7700. I also tried the mod with Intel C2D T7100
3.d- what is your cpu min and max voltage? (as shown in the 'CPU info' tab in RMclock)
L7700: min 0.85V, max 1.0125
3.e- your CPU minimum Vcore for every single multiplier when you're using RMclock.
Without mod: 9x/1,8GHz=0.9250V, 8x/1,6GHz=0.8625V, 0.85V for the rest. I didn't try with the mod yet.
3.f- in RMclock do you set the Vcores individually for every multiplier? or do you set the highest and the lowest multipliers only, and let RMclock adjust the intermediate multipliers?
Individually.
3.g- after the mod was your CPU sitting flat on its socket? or was it a bit listed?
It was flat, didn't seem listed at all. Wire I used was very thin.

 

1.1- With ViD3 1.0125v = 0.9125v which is about the same as 0.9000v. Hence more or less the same temperature (when rounded up).
Isn't 0.9000V supposed to actually be 0.8000V in VID3? So there's a difference.
You're right 0.9v becomes 0.8v... I was locking at the wrong column in the VID table. Its 0.1v difference so the temperature should be lower
1.2- have used the same multipliers for 0.9000v as for 1.0125v? or are you talking about a different multipliers?
I used same multipliers.

2- you can use normal electric wire. It seems to work (see nest post).

3- It seems that I'm lacking some information. Can we start from the beginning:
3.a- what laptop are you using?
Thinkpad T61 model 7661
3.b- what chipset do you have (intel, ati, nvidia)?
Intel GM965
3.c- what cpu are you using?
Intel C2D L7700. I also tried the mod with Intel C2D T7100
3.d- what is your cpu min and max voltage? (as shown in the 'CPU info' tab in RMclock)
L7700: min 0.85V, max 1.0125
I'm I right by saying that the only Vcores that you can see in RM clock are values between 0.85v and 1.0125v? or can you see values above 1.0125v? if yes what's the highest voltage you have access to in RMclock?
I remember you said that you tried VID4 mod and it failed. Was it with the L7700? what happend with the mod; was the laptop booting into the bios and windows or wass it completely dead (I'm mean sor of dead)?
3.e- your CPU minimum Vcore for every single multiplier when you're using RMclock.
Without mod: 9x/1,8GHz=0.9250V, 8x/1,6GHz=0.8625V, 0.85V for the rest. I didn't try with the mod yet.
3.f- in RMclock do you set the Vcores individually for every multiplier? or do you set the highest and the lowest multipliers only, and let RMclock adjust the intermediate multipliers?
Individually.
3.g- after the mod was your CPU sitting flat on its socket? or was it a bit listed?
It was flat, didn't seem listed at all. Wire I used was very thin.
I used thin wires too when I did the FSB mod on the T5300. But because the wire had to be a 'U' shape the CPU wa not flat... it looked flat but was not flat.
I noticed that because when I applied the thermal paste into the CPU, and then put the heatsink back, the marks left by the paste on the heatsink were less than the size of the CPU die... hence the reason for removing a bit of plastic from the CPU socket.
Removing a bit of the plastic ensured that the wires were not in the CPU way. I noticed about 4-5c difference in the temperature after removing the plastic from the socket.

 

1.1- With ViD3 1.0125v = 0.9125v which is about the same as 0.9000v. Hence more or less the same temperature (when rounded up).
Isn't 0.9000V supposed to actually be 0.8000V in VID3? So there's a difference.
You're right 0.9v becomes 0.8v... I was locking at the wrong column in the VID table. Its 0.1v difference so the temperature should be lower
1.2- have used the same multipliers for 0.9000v as for 1.0125v? or are you talking about a different multipliers?
I used same multipliers.

2- you can use normal electric wire. It seems to work (see nest post).

3- It seems that I'm lacking some information. Can we start from the beginning:
3.a- what laptop are you using?
Thinkpad T61 model 7661
3.b- what chipset do you have (intel, ati, nvidia)?
Intel GM965
3.c- what cpu are you using?
Intel C2D L7700. I also tried the mod with Intel C2D T7100
3.d- what is your cpu min and max voltage? (as shown in the 'CPU info' tab in RMclock)
L7700: min 0.85V, max 1.0125
I'm I right by saying that the only Vcores that you can see in RM clock are values between 0.85v and 1.0125v? or can you see values above 1.0125v? if yes what's the highest voltage you have access to in RMclock?
I remember you said that you tried VID4 mod and it failed. Was it with the L7700? what happend with the mod; was the laptop booting into the bios and windows or wass it completely dead (I'm mean sor of dead)?
RMClock says max Vcore is 1.0125 and I can only see up to 1.0125 in SLFM and 6x-9x multipliers. In 10x/IDA multiplier, the range I can see is 0.85V-1.1375V.
I tried the VID4 with L7700, it booted into windows like without the mod. The wiring could have went wrong because I didn't use insulation.
3.e- your CPU minimum Vcore for every single multiplier when you're using RMclock.
Without mod: 9x/1,8GHz=0.9250V, 8x/1,6GHz=0.8625V, 0.85V for the rest. I didn't try with the mod yet.
3.f- in RMclock do you set the Vcores individually for every multiplier? or do you set the highest and the lowest multipliers only, and let RMclock adjust the intermediate multipliers?
Individually.
3.g- after the mod was your CPU sitting flat on its socket? or was it a bit listed?
It was flat, didn't seem listed at all. Wire I used was very thin.
I used thin wires too when I did the FSB mod on the T5300. But because the wire had to be a 'U' shape the CPU wa not flat... it looked flat but was not flat.
I noticed that because when I applied the thermal paste into the CPU, and then put the heatsink back, the marks left by the paste on the heatsink were less than the size of the CPU die... hence the reason for removing a bit of plastic from the CPU socket.
Removing a bit of the plastic ensured that the wires were not in the CPU way. I noticed about 4-5c difference in the temperature after removing the plastic from the socket.
Then I believe the wiring caused temperatures to rise. But even if I take 4-5C rise into account, it still doesn't fully explain how the temperatures vary. I can post the temperatures for L7700 and T7100 CPU's with all their voltages in VID3, but they are confusing/strange.

 

If you're using the same program to stress your CPU (e.g. orthos) and the room temperature is more or less the same, the CPU temperature should be about the same.

1- what happens with VID3 when you set the Vcore in RMclock for multiplier 9x at a value between 1.0000v and below 1.0000v?

If your laptop crashes while running orthos that would mean that the mod was done successfully. If doesn't crash, it will mean that the mod has failed (i.e. check the wiring).

2- I mad the following diagram. I'm not sure but I think this should explain the temperature variation you've experienced.

UPDATED: 04/08/2009

 

Before I answer your questions about voltages and crashing, could you explain why VID3=0 at different (opposite) positions than before and and different/opposite voltages become lower by 0.1V?





These pictures contradict each other. I thought 0.8125V-0.9V were inaccessible in VID3.

 

Well, your CPU the L7700 is a low voltage CPU. The spane of Vcores available to it is different from those available to T2000, T5000 and T7000 series. When you asked me for help I was not aware that you had an L series CPU.

It was a mistake on my behalve. I assumed that those that are using a low voltage CPUs (like yourself) don't need to do the mod. Also, I don't own a low voltage CPU so I've kind of forgotten that they use less voltage than normal CPUs. Last, I thought that most low voltage CPUs were soldered and not socketed.

The table above will work for yout T7100 but not for your L7700. For example, if I take the case of my T5300, the range of Vcore availabe to the CPU are between v0.9500 and v1.1750v. If you map this range using the VID3 table (above), you'll see that the lowest value that can be used by my T5300 with VID3 mod is 0.9125v (i.e. 1.0125 - 0.1). For your L7700 it's 0.8125v

Hope that this has ansered your question. It's all about mapping your existing Vcores according to the mod you're using and the CPU you have.

 

Ok, I thought the available Vcores were same to every CPU. My L7700 is BGA (soldering) CPU that has been modded into PGA (socket) CPU. I'm trying to achieve 0.75V, at least for the lowest multiplier. My T7100 actually has 0.9625V as it's lowest Vcore, not 0.9500V, I think it varies a little.

1- what happens with VID3 when you set the Vcore in RMclock for multiplier 9x at a value between 1.0000v and below 1.0000v?
L7700 VID3 9x crashes at 0.9V
L7700 unmodded 9x crashes at 0.9125V
L7700 VID3 8x crashes at 0.8500V
L7700 unmodded 8x crashes at 0.8500V

T7100 VID3 9x crashes at 1.0125V
T7100 unmodded 9x crashes at 1.0750V
T7100 8x doesn't crash at all


Here are some temperature readings, I hope they're not too hard to read:
--------------------------------------------------------------------------

L7700 VID3 ( Vcore-Temperature-Temperature without mod)

0.8500-49-47
0.8625
0.8750
0.8875
0.9000-50-48

0.9125-55-49
0.9250
0.9375
0.9500-56-50
0.9625
0.9750
0.9875
1.0000-58-51

1.0125-54-52

--------------------------------------------------------------------------

T7100 VID3 ( Vcore-Temperature-Temperature without mod)

0.9625-55-53
0.9750-55
0.9875-56
1.0000-57
1.0125-59-56

1.0250-69-57
1.0375
1.0500
1.0625-72-61
1.0750
1.0875
1.1000
1.1125-76-65

1.1250-69-66
1.1375-71
1.1500
1.1625
1.1750
1.1875
1.2000
1.2125-75> and fast rise in temperature

1.2250-75> and much faster rise in temperature

--------------------------------------------------------------------------

From crash voltages and temperatures it seems like the voltages actually rose when they were supposed to decrease. Other temperatures also rose, maybe because of the wire as you said. I tried to tuck the wire into a hole in the socket, but it didn't have any effect on temperatures.

 

are those volatges and temperatures for your highest mutilplier

 

They are for multiplier 6x (1.2GHz).

 

Hi Mate,

There was a mistake on the photos showing the VoltMod for Socket P. It took me a while to see it. I updated the corrected photos. They are in the first post (page 1).

I have to thank you because without your insistance I wouldn't have spotted the mistake.

That said, please try the mod again and let us know your new temperatures. This time they should be lower.

-------------------------------------------------
You were right with the mistake on the photos you were actually rising the voltage by 0.1v for the voltages below 0.9125 and above 0.9875. The is the mapping for the L7700 with voltage + 0.1v

1.0250 --> 1.1250
1.0125 --> 1.1125
1.0000 --> 1.0000 |
0.9875 --> 0.9875 |
0.9750 --> 0.9750 |
0.9625 --> 0.9625 | those remain the same
0.9500 --> 0.9500 |
0.9375 --> 0.9375 |
0.9250 --> 0.9250 |
0.9125 --> 0.9125 |
0.9000 --> 1.0000
0.8875 --> 0.9875
0.8750 --> 0.9750
0.8625 --> 0.9625
0.8500 --> 0.9500

-------------------------------------------------
L7700 VID3 9x crashes at 0.9V (this is not normal since v = 1.0000)
L7700 unmodded 9x crashes at 0.9125V
L7700 VID3 8x crashes at 0.8500V (this is not normal since v = 0.9500)
L7700 unmodded 8x crashes at 0.8500V

T7100 VID3 9x crashes at 1.0125V (this is not normal since v = 1.1125)
T7100 unmodded 9x crashes at 1.0750V
T7100 8x doesn't crash at all


Here are some temperature readings, I hope they're not too hard to read:
--------------------------------------------------------------------------

L7700 VID3 (Vcore-Temperature-Temperature without mod)

0.8500-49-47 (this is normal since v = 0.9500)
0.8625
0.8750
0.8875
0.9000-50-48 (this is normal since v = 1.0000)

0.9125-55-49 (this is not normal since v = 0.9125)
0.9250
0.9375
0.9500-56-50 (this is not normal since v = 0.9500)
0.9625
0.9750
0.9875
1.0000-58-51 (this is not normal since v = 1.0000)

1.0125-54-52 (this is normal since v = 1.1125)

--------------------------------------------------------------------------

T7100 VID3 (Vcore-Temperature-Temperature without mod)

0.9625-55-53 (this is not normal since v = 0.9625) - (different room temp. maybe)
0.9750-55
0.9875-56
1.0000-57
1.0125-59-56 (this is normal since v = 1.1125)

1.0250-69-57 (this is normal since v = 1.1125)
1.0375
1.0500
1.0625-72-61 (this is normal since v = 1.1625)
1.0750
1.0875
1.1000
1.1125-76-65 (this is not normal since v = 1.1125) (typo maybe... 1.1125-76-75?)

1.1250-69-66 (this is not normal since v = 1.1125) - (different room temp. maybe)
1.1375-71
1.1500
1.1625
1.1750
1.1875
1.2000
1.2125-75> and fast rise in temperature (this is normal since v = 1.3125)

1.2250-75> and much faster rise in temperature (this is normal since v = 1.3250)

 

You're saying that the middle Vcores (0.9125v-1.0000V with L7700) stayed the same and outer Vcores (0.8500V-0.9000V and 1.0125V) rose. There must be some logic behind that, but it was evidently the other way around (no difference in room temps or typos). Highest multipliers didn't crash at middle Vcores, which means that they rose. Note too that the middle Vcore temperatures rose much more than the outer temperatures. Same analogy with the T7100 too.

The Vcore range (0.9625V-1.2250V) I posted for T7100 was wrong. It is actually 0.1125V too high for all Vcores range is actually 0.8500V-1.1125V). This was because I believed CPU-Z that reported it wrong wrong and not RMClock that reported it right right, because 0.8500V didn't seem believable for standard voltage CPU.

------------------------------------------------------------------------------------------------------------------------

I tried VID3 with L7700 and T7100 with new wire connection and they both worked. They lowered voltages/temperatures at those Vcores that are marked inaccessible at the end of your PDF (0.8500V-0.9000V and 1.0125V-1.1000V), unlike in this picture you posted previously:

Vcores that crashed also confirmed this. Also, I didn't have any other rise in temperatures (due to wire causing listed CPU for example) anymore, at least with L7700.

Does this mean that all CPU's with different Vcore ranges lower at same Vcores in VID3?

---------------------------------------------------------------------------------------------------------------

I also tried VID4 and VID5.

With T7100 (0.8500V-1.1125V) and VID4 my notebook couldn't start booting, so I couldn't even get into BIOS. It started to make beeping noise. I tried the mod twice and used tape insulation (tape put above the wire onto socket).

With T7100 and VID5 my notebook couldn't start booting either. It didn't make beeping noise with this mod.

I think it's strange that both of these crashed before BIOS access, because VID4 should only crash with low Vcore like 9.5000V and VID5 should only crash with high Vcore like 1.1125V. 0.9500V is default Vcore for lowest multiplier and 1.1125V is default Vcore for highest multiplier in T7100, but does CPU's really use both when just starting to boot?

I also tried VID5 to L7700 (0.8500V-1.0125V), which shouldn't alter Vcores at all. I could get into BIOS, but it crashed when it started to load Windows. This might have been because it used IDA/10X multiplier (0.8500V-1.1375V.) when loading Windows. I could get into Windows after disabling Intel Speedstep altogether in BIOS so that it only uses lowest multiplier (with default 0.9500V) but then I couldn't change the multipliers/Vcores in RMClock. Trying to set CPU to use only lowest/highest multiplier in BIOS didn't work.

 

with the error (i.e. when you followed the first tutorial) the above is correct. This is because you were connecting the VID3 hole to a Vss hole, leading to VID3=1 all the time.

VID3 + Vss = 0 (i.e. the error) ----> rises the vcore by 0.1v
VID3 + Vcc = 1 -----------------> decreases the vcore by 0.1v

Big difference compared to the the T7300. When we tested it in my friend's ACER, the maximum Vcore in RM clock was around 1.2v.

I checked intel's website the maximum for the T7100 is between 1.075V-1.175V (link here)

The T7*** used with socket P support SuperFLA . This why their lowest vcore is 0.8500v. That said all intel's Core Duo, and Core 2 Duo can in theory go as low as 0.7125v. With ViD4 (-0.2v) my T5300 was stable with 0.7500v (not far off the limit of 0.7125v).
--------------------------------

Forget about my PDF for now. Are you saying with the corrected VID3 mod:
0.8500V-0.9000V and 1.0125V-1.1000V ---> temperature lower than without the mod.
0.9125V and 1.0000V ----------------------> temperature same as without the mod.
?

Either VID3, VID4 or else, the minimum and maximum vcores you can use depend on the range of voltages that your cpu have access to. The wider the range is before mod, the wider it is after mod.

For instance the minimum I could use with my T2060 with VID4 was 0.8250v, because the minimum without the mod was 1.0250v.
-----------------------------------------------------------------

I was going to tell you to test the VID4 with your L7700 a while ago. The reason I didn't was because with the mod your CPU would have booted with multi. x9 with a vcore = 0.8250v (1.0250 - 0.2), and with those settings it would have crashed when loading Windows.

For the T7100. Did you reset RMclock before doing the VID4 mod? I mean did you increase the Vcore to 1.1125v for all the multipliers before attempting the mod?

Technically speaking your laptop should at least have booted into the BIOS with this setting. This is because VID4 doesn't modify you maximum Vcore of 1.1125v. The voltage mapping in the case of the T7100 + VID4 (-0.2v) is:

1.1125v ----------------> 1.1125v (same)
0.9125v to 1.1000v -----> 0.7125v to 0.9000v
0.8500v to 0.9000v -----> 0.8500v to 0.9000v (same)

Not sure why it didn't start at all. In my case it booted to the BIOS screen, and crashed while trying to load windows.
The only thing that I can thing of is that intel has added a feature to their CPUs to stopped them from working when the voltage is too low. They probably did this when they made the move from socket M to socket P.

You might be right. It could be IDA that is making impossible for the L7100 to work with VID5.
From what I have read is that IDA doesn't kick in automatically. By this I mean that it works only when a certain set of condition is meat... in other words never.

THANKS MATE for the tests, as it is helping everyone reading this thread to figure out what they can and what they can't do with their laptops and CPUs.

QUESTION:
What was the different in temperature before and after VID3?
Can you remove the photos from this post so readers don't get confused... thanks

 

Has any one tried the BSEL mod on a laptop with an nVidia chipset?

Guys the experiments I did with socket P were in my friend's laptop. I'm planning to upgrade my laptop soon. I'm planning to buy something cheap/second hand, upgrade it with a T7300, and force the T7300 to run at 2.66Ghz via BSel mod.

So, for that I'm looking at buying a laptop with a Socket P that can use an FSB of 266mhz, an ATI chipset and a 14.1" (1280x800) screen. Any suggestions?

 

Isn't it supposed to be (at least before, i.e. in the PDF):
VID + Vcc = 1
VID + Vss = 0

The pinmod results would also tell this. This would explain why you think my incorrect (Vss) pinmod gave strange results and my correct pinmod (Vcc) gave "opposite" (not like in the table for L7700) results.

Actually the maximum in my T7100 is 1.1750V and higher for IDA. With this Vcore range (0.8500V-1.1125V) I was just referring to the Vcores in this post (since I didn't post temperatures for higher Vcores).

Yes exactly.

Yes I set them 1.1125V. But doesn't RMClock Vcore settings work until Windows loads it after the login? It could have used 0.9500V Vcore before this.

No problem, thank yourself. I have spent a lot of time trying to figure out how to make my notebook silent/cool as possible. I even bought another temporary notebook to test low voltage CPU's and now, mods like this because I can't test with my current notebook (for irrelevant reasons). Before reading and testing this I really didn't believe nobody would find out how to get lower voltages than the Vcores Intel had decided, because C2D brand had been out for so long (Pentium M had much lower minimum Vcores).

I will have a lot of use for these mods, especially if I'm able to get lower than 0.1V, since a little newer CPU's like T9X00 and T8X00 have higher 0.90V-0.95V minimum Vcore.

Not a huge temperature drop for this LV CPU but still very meaningful for me in making silent/cool notebook. I can notice a big difference in temperature at keyboard under palms between 45C and 50C.

L7700 VID3: Vcore(V)-Temperature(C)-Temperature without mod(C)

1.0125-49-52

1.0000-52-52
0.9875
0.9750
0.9625
0.9500
0.9375
0.9250
0.9125-49-49

0.9000
0.8875
0.8750
0.8500
0.8625
0.8500-45-48
------------------------------------------------------------

Is it possible to use multiple different VID mods at a time to get lowered voltages at different Vcores and in bigger decreases?

 

Correct. I corrected the typo in the previous post. I also added a new photo to show the change in voltage for both cases.

Correct.

I actually did the same. My main laptop is a Gateway laptop (socket M). I bought an Eisystem laptop (socket M), just to test the voltmod.
I coudn't test the FSB mod on the Eisystem because it had an intel chipset. So I sold it.
I didn't do any voltmod in my gateway not because I don't want to but because I can't. The reason is that I pushed the FSB of my T2500 from 166 to 200Mhz. So now my T2500 is running at 2.4 Ghz instead of 2.0 Ghz.
Before the mod my T2500 @ 2Ghz was stable with 1.05v. At 2.4Ghz this CPU needs 1.225v to run stable so no volt mod. That said I'm lucky because the range of voltage available for this CPU is from 0.95v to 1.25v, otherwise I would have had to do a overvolt to get it stable.
I also tried the FSB mod with my T5300. I modded the FSB from 133 to 200Mhz (i.e. 1.73Ghz to 2.6Ghz). The problems I run into were:
1- Even with VID1 (+0.025v) the T5300 was not stable under Orthos. The voltage range of the T5300 is small compared to the T2500. Without voltmod the maximum is 1.1750v. With VID1 (+0.0250v) the maximum go up to 1.2v. To make the T5300 I needed at least VID2+ (+0.05v).
2- The temperatures were amazingly high under load (@ 2.6Ghz). I don't know the maximum, because the CPU was not stable, but I saw 82-83c before it crashed.
3 -With the T2500 @ 2.4Ghz I hit 76c under load. I think that this is OK because most of the time I keep the T2500 locked @ 1.2Ghz, with temperatures around 35 to 40c.

From my experience with the FSB mod on the T2500 and the T5300, changes in temperature are more noticable when reducing the voltage for the highest multipliers.

Yes you can, but the problem is you'll be reducing the number of vcores accessible to your CPU even further.
For intance, if you add VID2 + Vcc to you VID3 + Vcc, you'll be undervolting by 0.15v. With this mod, your L7700 will have access to only the following voltages:
0.9125v to 0.9500v and
0.7125v to 0.7500v

 

you said that the maximum vcore for you T7100 is 1.1750v. Both VID3 (-0.1v) and VID4 (-0.2v) have no effect on this value. So your laptop should start and load windows normally with VID4.

What's voltage for IDA? is it above 1.3v? What's the T7100 frequency when IDA is activated? 2Ghz?
Voltages between 1.3125v and 1.5000v + VID4 (-0.2v) ------> between 1.1125v and 1.3000v.
Surely your T7100 shoudn't have any problems running stable with a minimum of 1.1125v @ 2Ghz!