Normally 1.4 can only go up to about 1.7 in stability.

Use NHC (http://www.pbus-167.com/)and try with lower multiplier - for ex. 13 x 133.

to horacelau

You are unlucky with the CPU. My 1.6 Banias can easily take 1.96Mhz.

IMO, you have three options:
- Use NHC or RMClock to lowering multiplier to 13x or 12x.
- Voltmod the chip, I guess you need only 0.064V to get it stable. Search Google for the words "Dothan voltmod" for instruction.
- Buy a banias 1.6GHz or, better, Dothan 1.7Ghz and overclock it to 2.26Ghz. I think this option is the most interesting since these CPU are cheap now (about 20-30$ for Dothan 735 1.7Ghz).

Thanks all for your replies. I was able to reduce max multiplier and can now run at 133x13. I didn't think 1.86 would be a problem but I guess 1.4 is low grade siliicon and probably more voltage is necessary.

Anyone who has done the "pencil" method, could you please take some pics? I've tested this on my 1.6. fs0 - fs1, the thing power ups but but screen is blank no beeps. It will work if I remove the pencil marks.

It could be that the cpu wont make it to 2.1, this use to be the symptom on desktop when overcloking beyond the cpu's limit but that was zeons ago.

You should try this:
- BIOS -> Power -> Power Mode for AC/battery : Customized -> Customize : Processor speed : fixed slow


Thinkpad boots up with lower multiplier, but with FSB 133.

If i'm not mistaken, that's got to do with the speedstep tech. If i recall correctly the multipler is locked in the cpu and you can't change them in pentium-m.

I just grabed a 750 will try again when it arrives. Thanks for your help.

I agree with xob7, you should set the CPU on BIOS to fixed low to force it work at lowest multiplier (i.e. 6 ).
Also, did you reprogram the SPD to DDR266 ? I have a feeling that your problem is likely in memory.

I did try setting the processor speed to lower and it was the same the, powers on but with a blank screen. I flashed the ram to 266 but maybe I missed a step there. After flashing cpu-z was reporting my memory as pc2100 which should be correct. I also tried with a another pc2700 stick and it still no go.

If else fails, I'll try shorting fs0 - vref. But it looks like it would be more difficult. Any easy way to do this?

winboxes:

What is the revision of your motherboard (chipset) ? You can see it in the "mainboard" page of CPU-Z. Mine is A3. I have heard that earlier revisions of 855PM chipset have lower chance of working at 133MHz FSB.

Also, you said that CPU-Z shows that the memory work at 266Mhz. It is right after re-flash SPD or after restarting the computer ? And you read this information in page "Memory" or page "SPD" ?

binh:

I have the i855pm chipset (A3), same ppl as xob7. It showed pc266 after I flashed the memory on the SPD page.

One more funny thing is I cannot upgrade my bios above version 1RETCDWW(3.06f) and Embedded Controller Program 3.03.

http://www-307.ibm.com/pc/support/site. ... 50273#bios

My Thinkpad had bios version 1RETC4WW (3.04) and Embedded Controller Program 3.03.

I managed to upgrade the bios to version 3.06f and when I attempt to flash from version 3.06f to 3.08a using non diskette the pc reboots as it should but does not load the "Bios utility program", it goes back straight to Windows. The same thing happens when I attempt to upgrade the Embedded Controller Program too.

If you look at the bios available at Ibm it's a jump from version 3.06f to 3.08a, shouldn't there be a 3.07 in between?

Anyone have any idea from which bios version did they start supporting 133fsb? Not sure if 3.06f does.

Please, restart the computer and check the memory speed in CPU-Z's Memory page. It should show 133Mhz and 3:4 divider.

I do not think that the version of BIOS has a meaning. Though my two successful attemps were made on the Thinkpads that have lasted BIOS version (3.23) and Embedded Controler (3.04).

please continue reporting your process.
I am scouting the market for old pentium M-cpus which I can afford to burst while overclocking.I am getting only 1.3 GHZ bucketfull and 1.6 Ghz much less.All scrap dealers say that intel or their accomplices never released the faster versions in
this part of the world.(INDIA).
what happens to the peripharlas (i.e. pcmcia cards,usb, wireless a/b/c..

Recieved the 750 and with some spare time manage to test it today. It fires up with 133fsb but the multiplier is detected as 6x. So the speed is 800mhz instead of 1.86 Not sure why and have ran out of ideas. Maybe it's the steping

Do you set the CPU speed in BIOS to Fixed low (see the image supplied by xob7 in this thread) ? What is power management scheme in Windows ? And which version of BIOS do you have ?

In this thread, something is not clear. Perhaps, someone could comment on the following: after connecting pin 55 (fs1) to pin 54 (fs0) on ICS950810 or CY28346 with the pencil method or a wire, in booting up, does the 1.6 processor automatically run at 2.13 GHz at 533 MHz FSB or does it boot up at 1.6 GHz at 400 MHz until someone does a pin mod of the dothan processor?

It will automatically run at 2.13 GHz at 533 MHz FSB.

Binh,

Actually, I have a Dell Latitude D 600, an 855PM chipset version A3, the latest bios, a dothan 1.6 GHz 400 MHz FSB processor, and the PLL is an ICS950810. I read your post about the PLL mod involving pin 55 to pin 54 with the pencil method and thought it could work for the D 600 like the T42 or the T41.

I connected pin 55 to pin 54 with the pencil method. I booted and it booted up at 1.6 GHz at 400 MHz FSB.

I read in your post about the pencil method and something about not "de soldering the 55 pin." I am not clear about what the "de soldering of 55 pin" means or involves. Perhaps, you could explain what the procedure is and how to do it and whether that could be the reason why the D 600 is not booting up at 533 MHz despite the pencil method.

I have queried before in this subject lot only.
what about the rest of the peripharals. Do they work after this pencil patch on the ICS?.

deleted

The pencil method works on T4x. But it doesn't mean that it can work on all other 855PM laptops. My guess that it depends on the design of motherboard. Here is my hypothesis:
- Before overclocking, the pin 54 has high voltage (3.3V, level 1), the pin 55 has low (0.0 V ?) voltage (level 0).
- In T4x, the 55 pin is grounded through a resistor (perhaps with a high resistance). When you connect it to pin 54 using a pencil, a voltage is passed from pin 54 to pin 55 that is enough for considering it the level 1 (note that the voltage at pin 55 is less than that of pin 54, what is true level 1).
- In your Dell D600 the pin 55 might be connected directly to the ground (or through a resistor with low resistance), and the voltage passed from pin 54 is not enough for considering it a level 1.
- In your case, I would like to de-solder the pin 55 and connect it to pin 54 as it was recommended in many sources on internet. I do not recommend you to connect pin 55 directly to pin 54 without de-soldering since it is dangerous if pin 55 is connected directly to the ground and pin 54 - to 3.3V line.

Since setting high voltage on pin 55 raises only FSB from 100Mhz to 133Mhz, other frequencies (PCI and AGP) are not altered, so all the peripherals will work on their normal mode. Check out the pdf file of your PLL.

thank you, Now i will try the pencil method on my 1.6 Ghz T41.

"In your case, I would like to de-solder the pin 55 and connect it to pin 54...I do not recommend you to connect pin 55 directly to pin 54 without de-soldering since pin 55 is connected directly to the ground and pin 54 - to 3.3V line"

Binh,

I have researched some on the internet about "lifting" "leg 55 off ground" and "jumpering" pin 55 to pin 54. But, some of the language of what this means is vague or some of it is translated from a different language into English and so what is meant is not clear. Perhaps, you could clarify on the matter.

1. By "lifting leg 55," is it about literally de soldering the outer half of pin 55 and bending the de soldered half up and bending it over to touch pin 54? Or does it mean to destroy the outer half of pin 55 by de soldering and, for example, working with the pencil method to connect pin 55 to pin 54? And in either case, could someone work with something like a knife to "lift" "leg 55" rather than de soldering?

2. There are some on the internet who mention the "lifting" of pin 55 "off ground" and connecting it to pin 54, and another step of pin modding the Dothan processor to overclock.

If pin 55 is "ungrounded" and "connected" to pin 54, this apparently transforms pin 55 from value "0" to value "1" and the result is that pin 55 is at "1" and pin 54 is at "1." It seems in theory that should be enough to have the processor from 1.6 GHz at 400 MHz FSB to 2.13 GHz at 533 MHz FSB. But it is not clear why some on the internet mention that there should be, at this point, a step of pin modding of the processor.

Since de-soldering is not easy, people usually cut off the pin 55.

I have found an tutorial on pin modding the Dell D600. Though the CY28346 PLL is shown in this tutorial, but the same procedure can be applied to ICS 950810.

I've checked that and it was at Fixed Max. It could and does appear to most likely be the bios. I could only upgrade to version 3.06 using the non-diskette method. I don't have a floppy so I can't test if it will upgrade with a floppy. It could be my operating system on why the non diskette will not work above 3.06. I'm using win2003.

AFAIK, there is a method of upgrading BIOS using CD-ROM. Please, try to search on the internet.

Updates about pinmodding of my T42p:

Originally, the dothan 1.8 in my T42 is able to overclock to 2.4Ghz but it is stable only at 1.436v. This makes the CPU temperature to reach a dangerous value of 82-85C when doing stress test Prime95 in hot days (35-37C, what is normal for summer in Hanoi).

So I decided to buy a P-M 735 1.7Ghz CPU for $16. The CPU was randomly chosen and is a very good one, it can do 2.26Ghz at 1.276v (Prime95 stable). The CPU temperature under stress test drops to a quite low value of 67-69C.

Moreover, a new version of SetFSB utility starts to support the CY28346ZI-2 PLL, what is used frequently in thinkpad T4x and X3x. Using SetFSB, I bump the FSB up to 141.3Mhz, i.e. CPU Freq. = 2.403 Ghz, and it is stable at default voltage 1.340v ! However, the SetFSB sometimes crashes if I try to change the FSB (lower or higher) after hibernation - resume with the battery attached in. I think that it is possibly due to a conflict in SMBus between SetFSB and Battery management utility, or it is due to the non-standard (i.e. pinmoded) mode of my PLL chip.

Can some kind soul post more information on how to use setfsb on my 1.6 Ghz T41 .I bought the
thinkpad from the scrap market and can afford to play with the motherboard as the warranty is already kaput...

+1
I tried to use 133fsb method described above, but every time I short 54 and 55 pins I get:
1) Normal windows boot, 100fsb.
2) Windows doesn't boot.
3) Windows blue screen error, different every time like:SOMETHING_BAD_HAS_HAPPENED.

I tried using setfsb, but if I change frequency sound shuts down. RAM works at 133Mhz(Binh, thanks for your guide ).Mb I should underclock other devices before using setfsb?