WIP: Karaktu's (and Zap's!) Unofficial Guide to mobile Celeron overclocking success!

[buzzly]

In my Dell 4500, with a 2.4g mobile celeron, I managed to do both 133fsb and 1.6v pin mod and it is working fine.

Now I am getting brave and do the 200fsb on a 1.6g mobile celeron by breaking off both 133fsb and 200fsb pins and put in the Dell 4500, but the computer only recongizes it as 133fsb mobile Celeron.

Did I do anything wrong?

 

[Zap]

Two possibilities. First is that the maximum the Dell motherboard supported was 133MHz FSB. If the board doesn't support the next up speed, it just won't do it. The second thing is that breaking BOTH pins off is "reserved" according to the socket 478 specs. Both pins with voltage would be 100MHz, both w/o voltage is labeled "reserved." BSEL0 w/o is 133MHz and BSEL1 w/o is 200MHz. If you aren't sure about the Dell motherboard, if you can tell us the chipset then we can tell you if it "officially" supported the higher FSB.

BTW, you can simulate the BSEL0 pin still being in place by duing the "wire trick" to connect BSEL0 to the nearest VCC. BSEL0 is pin AD6 and there are two adjacent VCC pins that can be used, AD7 or AE6.

 

[buzzly]

Thanks Zap.

The Dell 4500 has Intel 845 chipsets which only supports 533fsb.

I also tried to simulate the BSEL0 pin still being in place by connecting AD6 and AD7. Still, the computer only recognizes it as 133fsb. I guess I will put the 2.2 ghz Mobile Celeron back in and run it @ 2.91ghz. I am still happy with it.

 

[jjmIII]

Originally posted by: Zap

Originally posted by: jjmIII
What should I expect from this Mobile Cel 2.0 in this system ??

At least the same as your 1.6@2.66GHz. Likely you'd get 3GHz or more but YMMV. I have a similar chip (don't remember SPEC) in my HTPC, but I'm running it near silent at 1.2v default MHz, not a good candidate for 3GHz+ testing.

Well...

This guy named "Joe" seems to think it should do 166bus (3320mhz) since it is a D1 core. Was just curious if others had that success??

 

[Zap]

I had a D1 Celeron 2.4GHz that did 3.2GHz fine. YMMV, try it and let us know!

buzzly
Yup, no dice on that chipset with a non-overclockable Dell board.

 

[nickdigger]

Zap: Which is it?

Potayto. (vid4-0) I think the mixup was me assuming default 1.2v for lower-speed Mobile Cel's. Maybe that only applies to Speedstep variants.



BTW, you can simulate the BSEL0 pin still being in place by duing the "wire trick" to connect BSEL0 to the nearest VCC. BSEL0 is pin AD6 and there are two adjacent VCC pins that can be used, AD7 or AE6.

I believe the pin positions need to be grounded to Vss, not Vcc, in order to make them read "0" again. Buzzly, you'd need BSEL1 = 1 (open connect) and BSEL0 = 0 (tie to Vss) for 200mhz... if your board supports it.

 

[Zap]

There are no mobile Celerons with Speedstep. The P4 CPUs with Speedstep also default to 1.3v. This is all AFAIK.

Also, from how I read it, 1 = lack of signal while 0 = signal. AFAIK signal means voltage. I've done this before with connecting various VID pins to VCC (and not VSS). Of course I could be wrong (that's happened before )... does anyone have a third opinion?

 

[marill]

Thought I'd add my experiences to the list

Soyo SY-P4RC350
Wouldn't post until I bent the AE1 and then it posts at 1.25v w/ a 100mhz fsb. The overclocking options are nonexistent (only 15mhz above the default fsb) but breaking the BSEL0 pin will give a default fsb of 133mhz. Ran a mobile celeron 1.5ghz at 2.0ghz (15x133mhz) and 1.8ghz at 2.4ghz (18x133mhz). Not an overclocking board but an excellent htpc option since it has onboard s-video and composite outputs in a matx form factor. Good choice if you have a half-height case and don't need directx 9 support.

Abit BD7II-RAID
I tried a mobile celeron 1.8ghz w/ the BSEL0 and AE1 broken/bent and it boots w/ a default fsb of 100mhz and 1.2V. Note that the default voltage is lower than the rated 1.25V so I experienced random restarts w/ my 1.8. By raising the voltage 5% (bios allows increases of 5, 10 and 15%), the stability issues went away. I only overclocked the fsb to 133mhz BUT BE SURE YOU SELECT "FIXED 33MHZ" FOR THE PCI DIVIDER IN THE BIOS or random reboots will occur (at least for me). Other than that, it's happily running my 1.8 at 2.4ghz (18x133mhz). Also, I found that this board does not like my xfx geforce mx4000 pci or agp card; the resolution is locked at 640x480 w/ 8-bit color and any changes cause resets. However, my matrox g450 (agp) and old pci mystique run fine.

* EDIT *
It turns out that mobile celerons will work fine in my Abit BD7II-RAID without any modifications. The cpu defaults to 1.60v and has fully working agp/pci locks (according to the clockgen utility).

Thanks to everyone who contributed to this thread esp Karaktu and Zap for their wonderful list of motherboard compatibility. Their posts helped me build a nice, quiet video encoding box built with an Asus P4P800-VM running a mobile celeron 2.0ghz at 2.66ghz (20x133mhz) at 1.25v

 

[nickdigger]

Zap: 1 = lack of signal while 0 = signal. AFAIK signal means voltage.

Or, signal could be represented by current, ie, if pin is grounded, current flows through it. It's been a while since i had to tie pins together, but i'm pretty sure 0 = Vss.

 

[Zap]

Originally posted by: rogue1979
Well, I had a string of bad luck lately. My Asus P4S800-MX up and died.

It's my turn for bad luck. My last remaining Asus P4S800-MX also died... well, I killed it on accident. Was trying to get the stuff shoehorned into a tiny mATX case and I had to remove the HSF to install drives. Trying to reinstall the HSF... couldn't get it to clip in so I used a screwdriver. SLIP - and the screwdriver tip jabbed into a tiny component next to the RAM slots, smashing it a bit. Now board won't POST with RAM installed, but without RAM installed will happily give the no-RAM beep. I've verified parts are okay on another board (Asus P4SP-MX).

BAH HUMBUG!

Anyone know how to fix these motherboards? You can have this board for the cost of shipping.

Now I have to decide whether to get another mATX board for this CPU, or to get a mATX socket 754 board and use one of my spare socket 754 CPUs.

 

[nickdigger]

Trying to reinstall the HSF... couldn't get it to clip in so I used a screwdriver. SLIP - and the screwdriver tip jabbed into a tiny component next to the RAM slots

ouch. i'm shocked this hasnt happened to me at least 2 dozen times with those gee-damn Athlon HSF clips.

 

[Zap]

Originally posted by: Zap
Hmmm, did a quick search and indeed looks like the supply of the P4S800-MX boards dried up. There's the P4S8X-MX boardx for $53 that may do the trick.

Pro: Cheap price, updated with SATA.

Cons: Unknown compatibility with mobile CPUs (though historically it should be good), unknown overclocking functions in BIOS, the chipset "officially" supports only 533MHz FSB but Asus "guarantees" it at 800Mhz so may mean that it may not hit much more than that.

Would you like to be the guinea pig?

I guess since I killed my P4S800-MX, I'll be the guinea pig. I just ordered one from ZipZoomFly for $52 shipped 2nd day (need it quick). Buy.com has it for $51 shipped Ground, but I know stuff from them takes up to a week on ground, but it could be a good choice for people not in a hurry plus I've heard that there is no restocking fee on returns and they cover return postage. Newegg was about $55 shipped UPS 3 day, so they're out of the running. All three places charge tax in CA, but not for me in WI.

After I receive it and have had a chance to play with it, I'll post a review of the board on the Motherboards category and post a link to it in this thread.

 

[osage]

OK, question/problem.

Asus P4S800 Rev. 2 mobo bios chip showes ver 1005 bios, I got this from a trader who upgraded to a PCI-E mobo, said to be working when pulled and I have no reason to doubt that.
1.8 Mobile Celeron with all pins intact and a 1.8 Mobile Celeron with the AE1 pin removed.

board will not post with either cpu, tried know working PCI and AGP vid cards, pc2700 and pc3200 from Crucial,Kingston,Samsung, and Infinieon all know to work.
400W PS from Allied and Sparkle, 350W PS from Enlight, and Antec also know to work.

testing outside the case on cardboard box as always, no drives connected only cpu/ram/vid/keyboard, and switch.

any ideas at all ? ? ? bios update perhaps, tho I don't really know what version is on it.

 

[Zap]

Do you have a desktop CPU that you can "borrow" from another machine to test with the board?

Grrrr! Yesterday my P4S800-MX was killed in a screwdriver accident. Today (just about 10 min ago) my HTPC with a mobile Celeron (not even overclocked) and Asus P4SP-MX just died. Machine was on, then suddenly was off. Turned back on, no POST. :| I don't even want to deal with it tonight.

 

[osage]

no, I don't even know anyone who owns a P4 system, everybody is A64, or socket A.

got one ? I'll send you the mobo to test it. or visa versa

 

[Zap]

I've got a verified working CPU, a verified dead motherboard... and unverified stuff in my HTPC after last night... gonna take it apart today to see what's up with it. I've ordered a P4S8X-MX and that should be in by Wednesday (if ZipZoomFly ships out Monday). Unknown how that board fares with the mobile Celerons but I have hopes. If you want to mail me the stuff (and pay for return postage) I can test it out. Just need CPU and mobo, no RAM, HSF, video, PSU, etc. At most, would cost $8.10 each way (Priority flat rate). Alternately, postage and DC for the CPU itself would be pretty cheap, probably no more than $2-3 each way. If you live near enough to me perhaps we can hook up somewhere? PM me.

 

[osage]

PM Zap, let's do it.

 

[Zap]

An update and some good news. My HTPC is working fine now. The culprit was a loose CPU fan plug - it was plugged in but not enough. I hooked it up with the cover off and the machine booted right up! The CPU fan wasn't spinning (Zalman 7000AlCu) and i verified that nothing was stuck in the blades and that the fan still spun freely, plus that it was still plugged in. I moved the wires a bit and it came out! Plugged it back in and now it's happy... and I'm happy. Well, still bummed that the other board is definately dead, but ah well.

 

[Zap]

Some updates...

osage's CPUs are both dead, but the motherboard works fine... indeed better than the P4S800 that I used to own (no warm boot problem on his Rev 2.00 board).

I received my shiney new Asus P4S8X-MX only to find out that it does not work with the mobile Celeron, with or without the bent pin vcore mod. It uses the AMI BIOS that Karaktu warned us about on the P4S800-MX boards that were included in the Asus Vintage barebones. Look for my upcoming (soon as I finish writing it and uploading some pics) mini review on the Motherboards category. It isn't really a bad board, just doesn't work with the mobile Celeron.

 

[osage]

well thats some good news at least, I do really appreciate you testing the parts for me. Now I just need to decide if I want to take another crack at it, or just stay with my AMD systems.

:beer: thanks again.

 

[Zap]

Fiddled with the P4S800 some more today and I have more results.

Mobile Celeron CPU on Asus P4S800 Rev 2.00 board with BIOS 1007. Rev 2.00 motherboard does not seem to have the "warm boot" problem of the originals.

With bent pin vcore mod, motherboard sees default 1.2250v.

Without bent pin vcore mod, motherboard sees default 0.8500v (IIRC) and won't POST. Well, how did I find that out you may ask? I would put the modded CPU in the board, boot and go into BIOS and set vcore to manual, then save/exit. Then shut off, swap CPUs with unmodded chip, then it will POST one time. After it POSTs I can do whatever I want with it, boot off UBCD, go into BIOS, whatever, and that's how I saw the board is assigning it the low vcore (when I move vcore option to Auto it tells the detected setting). Once I either shut off, reboot or exit BIOS, then it no longer turns on - fans will turn a bit and then it stays off. Same thing if I turn off the PSU (the "fix" for the warm boot problem of older P4S800 boards). Same thing if I clear CMOS.

Basically this board only works properly (almost at all) with a voltmodded CPU.

OTOH my Asus P4S8X-MX board does not POST with a voltmodded CPU, but does with an unmodded CPU (my mini review has been updated). However, the board always says invalide microcode and so everytime it POSTs, you have to hit F1 to continue, after which it works fine.

Just more fodder for the mobile fire.

 

[Zap]

More results on the Asus P4S8X-MX. I'm gonna post here instead of my mini review because this information is of more use to those using mobile Celerons.

My previous assessment with the voltage modded CPUs may or may not be correct because that CPU also had a BSEL mod for FSB. I had broken of the BSEL0 pin so that the CPU detected at 533MHz FSB and bent the pin for 1.2v. Well, after finding out the CPU won't POST in that board, I bent the VID pin back. Voila, the CPU POSTed in the board. Elated, I disassembled my HTPC (Coolermaster ATC-620 case) and installed this board. Once in the case I POST tested again and it worked, so I bolted everything down and plugged everything in... only to find out that it no longer POSTed.

Had to rip everything apart again, and testing the board outside of the case with just CPU/HSF/RAM/PSU, still no POST. Cleared BIOS, no POST. Swapped CPUs to my other mobile Celeron that had never been modded and it POSTed. Swap back and no POST. I thought I had a dead CPU on my hands, but decided to test it in the old Asus P4SP-MX board, and it POSTed!!! Back in the P4S8X-MX and no POST. Swap in the never-modded CPU and POST. Now what? Well, I manually set the BIOS CPU settings and swapped to the other CPU and now it POSTs!!! Turned it off and on a number of times and it was fine. Hmmm... Clear CMOS, no POST!!!

With the VID pin straightened again, the only mod was the BSEL0 pin that I broke off. It should work because the motherboard (and chipset) officially supports 533MHz FSB CPUs, but... this board has a problem with mobile CPUs (see pic, I have to hit F1 everytime it POSTs). A quick look at the Intel developer PDF files confirmed my memory... I then used the "wire trick" and jumpered BSEL0 and BSEL1 together in the socket and installed the CPU. Now it POSTs everytime, even after clearing CMOS with the jumper. Problem solved!

Well, there's another problem now involving RAM speeds and overclocking. Because the memory divider is based on the base FSB, a base FSB of 400 has 1/1, 5/6 and 2/3 dividers on this board. Well, at the lowest 2/3 divider, you get a perfect DDR400 speed with the FSB set to 150MHz, any higher and your RAM is out of spec. That's okay if you have good RAM or if the board supported vDIMM boosting, but the board doesn't and I don't (at least the RAM I intend to use in the board). So, at this time I'm stuck at 150MHz FSB (2.4GHz) even though I've Prime tested this exact CPU just yesterday in osage's P4S800 board at 180MHz FSB (almost 2.9GHz). Raising FSB to 160 and no POST, probably because of my el-cheapo RAM. I've had this board at 180MHz FSB with my P4 2.53GHz CPU to see it POST at just over 3.4GHz on default vcore so I know it isn't the motherboard, and because the 2.53GHz desktop P4 works fine in the board, I know the board does what it supposed to do which is support 533MHz FSB CPUs. Just doesn't like mobile CPUs, that's all.

Oh yeah, I've also deduced why the system magically stopped POSTing once in the case and all hooked up (everything but case cover). I had swapped in the CPU after manually setting stuff using the unmodded chip, so it worked on manual settings. Well, the last time it worked, I probably turned it off before it finished POSTing. Anyone who has fiddled with Asus motherboards know that the Asus C.P.R. feature will reset CPU speed if it doesn't finish POSTing because it considers that a failed overclock. Well, with the BSEL0 mod, the board won't POST on autodetected settings!!! The moment I shut it off before it finished POSTing, the board reset CPU from manual settings to the auto settings which doesn't work with the mod.

I now know this board a little bit better, and I've also wasted about 6 hours of my life. Hopefully someone was entertained by this story besides my wife listening to my curses.

 

[allenpan]

problem. my gigabyte 8IE won't detect my P4M 1.6G as 12 mulitplier, but it detect as 16, i wanna lower to 12 and oc to 200fsb..how do i do that?

 

[allenpan]

so far, i test on my mobo it works with my Celeron Mobile 2.4 and 2.2, P4M 1.6,

ASUS P4T-E
ASUS P4T533-C
ASUS P4B
ASUS P4B-LS
GIGABYTE 8IE
GIGABYTE 8IE533

they works like wonder just the gigabye refuse to lower to 12 mutilplier

 

[Zap]

That's strange... most people want the higher multiplier.