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

[TStep]

Zap, I've had mixed results with these. For the most part they 1.6 and 1.7 P4-Ms. Just a guess, but I probably went through 5-6 over the last year or so. B0 and C1 (one bummer cpu), but mostly B0 (don't remember the sSpec off hand).

High's were fsb ~ 230 and lows ~ 190. I'd say 2.4 is a reasonable expectation achievable most of the time, but I'm usually never the lucky recipient of the golden chips anyway. From my experience, they'll boot in any board w/out pin mods, unlike the mobile celeries. More challenging than a celery because you are usually playing above fsb=200 while trying maintaining memory performance without additional expense.

Alot easier now than a year ago, but fsb between 220-240 was a tough spot with affordable memory. Running 1:1 was hard for most value memory, and 5:4 dropped you down in the 180s; while having to buy premium memory sort of defeated the purpose of the low-cost experiment. One of our field computers (in a dusty construction setting) is chugging away with a P4-M 1.6 on a P4S800 flawlessly. I set it to either 12x166 or 12x200, but I can't recall which.

 

[ammarcoux]

Hello,

I got each of these boards to POST with a Mobile Celeron 1.8GHz model SL6J4. I did not test their overclocking abilities too much.


- ASUS motherboard

P4G533-LA REV1.02 (Comes froms a HP Pavilion 533W starting with MX serial #)
- Recognizes CPU as a Intel Pentium 4 1.8GHz (works without bent pin)
- Defaults to an unknown voltage
- Tested with BIOS v3.10
- No overclocking options what so ever that I know of.
- Intel 845GL Northbridge
- ICH4 Southbridge


- ECS motherboards

PT800CE-A REV1.0A (HT Ready)
- Recognizes CPU as Mobile Intel Celeron 1.80Ghz
- Defaults to 1.55v (bend pin for 1.20v)
- Tested with BIOS v1.1A and v1.1E

- VIA VT8753 Northbridge
- VIA VT8233 Southbridge
- 1Mhz increments in the BIOS and onboard jumpers to switch in between 100-133-200MHz bus
- Can boost voltage up to 6% higher

P4VXMD REV1.0
- Recognizes CPU as Intel Celeron 1.8AGhz
- Defaults to 1.60v (bend pin for 1.20v)
- Tested with BIOS v1.2D
- VIA PT800CE Northbridge
- VIA VT8237 Southbridge
- Can switch bus speed in the BIOS in jumps
- Can boost voltage up to 6% higher



I was quite supprized that the PT800CE-A motherboard actually recognized the CPU as a Mobile Intel Celeron 1.8GHz processor. Even though they bothered putting the correct name, the voltage is still set to 1.55v instead of 1.3v. I'll send and email to them and see if they'll change it.

Almost forgot to say this, but it seems that some or all of the ECS P4VXMD motherboards have faulty capacitors. I say this because 9 (all) of the OST 6.3v 3300uF capacitors on the board I got are either domed or already leaking a bit on the top. So be aware that even if the board may not cost a lot, you need to keep a close eye on the capacitors for leakage and then replace them.

Later

 

[cornholio2]

I'm using a P4S800-MX right now and using a 1.7Ghz Intel Celeron. I have 2 256mb PC2100 memory sticks in there which is a cheapo. Looking at the PC Probe right now, mine is at 1.758v (this is the Vcore right?). I set the CPU multiplier at 17x. I didn't adjust any from the memory or tha AGP because I don't know if these are locked or not. I didn't use CPU-Z to verify it, I'm just looking at the ASUS PC Probe. Using BIOS 1010 also.

 

[Zap]

Cornholio2, your desktop Celeron 1.7 official vcore is 1.75v. The motherboard is running it almost spot-on.

Ammarcoux, I have a PT800CE-A board also. Seemed to work pretty nicely for a disposable motherboard, until it died. Some black piece of plastic broke off from underneath the ZIF socket and the board stopped POSTing.

 

[ammarcoux]

Originally posted by: Zap
Ammarcoux, I have a PT800CE-A board also. Seemed to work pretty nicely for a disposable motherboard, until it died. Some black piece of plastic broke off from underneath the ZIF socket and the board stopped POSTing.

Arrrr that kinda sucks. The only black plastic I see underneath the board's zip socket are the retention bracket clips? I also looked at my ECS PT800CE-A and notice that is has 9 OST 3300uF 6.3v capacitors just like the ECS P4VXMD. OST is known to make faulty capacitors so I'm going to keep an eye on those. Just wondering but would you still happen to have the board. I would be interested if that piece of plastic came from a capacitor.

Here's a link to www.badcaps.net. It's a forum that is dedicated to the topic of bad caps on computer equipement. I highly recommend everyone to check out this site since bad caps are responsible for a lot of weird computer behavior and failure.

 

[Zap]

Originally posted by: ammarcoux
Just wondering but would you still happen to have the board. I would be interested if that piece of plastic came from a capacitor.

Not from a capacitor. I think I threw away the plastic, but remember exactly where it came off. When you lift the lever the top section of the socket slides towards the side where the lever comes out of the "cam" area. Behind and underneath where the lever makes a 90º angle into the socket is some black plastic. I think it may act as a stop to the lever. Anyways, a chunk of that plastic broke off. Don't know if that's what's causing the thing to not POST, but also the ZIF socket no longer holds the CPU in as tight anymore so it is likely. Yes, I still have the board. I'll take another look at it tonight after work.

 

[TStep]

Maybe a chard of plastic worked it's way into one of the 478 socket widgets (whatever the things in the socket itself that recieve the pins are called) and is insulating the pin-to-wigdet connection.

 

[ammarcoux]

Yeah yeah I see the black plastic your talking about. Its the bottom part of the socket. Thats a pretty weird defect.

 

[Zap]

Tstep, I'm pretty sure the socket is broken. A shame because the board had only a few minutes of use. If any of you guys want it to examine/attempt to repair, just pay me shipping and I'll mail it to you (absolute maximum would be $4.30 for Priority Mail with DC). After it broke I removed the battery and the Northbridge heatsink for other uses, and I swapped out the P4 retension bracket for a broken one from another board. Otherwise it is in the exact same shape as before it broke. Also have the box, manual and driver CD.

 

[lapierrem]

What about the celeron mobile. - the northwood ones with the silver top?
I have one of these on my laptop, it's apparently different voltage. Overclockable?

 

[TStep]

Zap, thanks for the board, but the better half would eat me for breakfast if another "project" comes through our front door.

lapierrem: as far as I've seen, the celerons with the IHS are probably desktop Celerons w/ 128k L2 being used in a laptop. So they should work flawlessly in a desktop mobo. Check the sSpec code on CPU World to see what it actually is.

 

[Zap]

TStep is probably right about desktop chips used in notebooks. Many of the cheaper laptops (Dell 1100 comes to mind) used normal desktop chips instead of lower voltage "mobile specific" models.

Well, got my mobile P4 chips in and have done some preliminary testing to find out how high they can POST (stability testing later when I have more time). The 1.4GHz P4 is a B0 stepping and it will POST at up to around 12x230=2760MHz. Does not POST at 240MHz FSB. The 1.6GHz P4 is a C1 stepping and it will kinda POST at up to around 12x250=3000MHz. I'm thinking this chip might be limited by the motherboard since it does some funky stuff at that high a FSB. I'll have to see if my friend with an Asus P4P800 Deluxe (formerly mine) is willing to do some testing for me.

My next round of testing will be with Shuttle XPC mini systems. I'm getting an SS51G (SiS 651) and an SS56G (SiS 661). A friend already uses a mobile Celeron in an SS51G so I'm pretty sure I'm good to go on this one. The SS56G will be breaking new ground with mobile chips (unless someone's holding out on us).

 

[Jon187]

Hahah, great thread Joe.

Cheers from Down Under.

Jon....

 

[Zap]

Originally posted by: Zap
My next round of testing will be with Shuttle XPC mini systems. I'm getting an SS51G (SiS 651) and an SS56G (SiS 661). A friend already uses a mobile Celeron in an SS51G so I'm pretty sure I'm good to go on this one. The SS56G will be breaking new ground with mobile chips (unless someone's holding out on us).

Well, picked up the two Shuttle mini systems as refurbs from Newegg. The SS56G is a version 2 and it POSTs. Tested with onboard video as well as an AGP card. The mobile CPUs need the pin mod or it won't POST. The SS51G seems to be DOA. Tried with same stick of RAM and even pulled my P4 2.53GHz from my other system, no POST and fan speed stays high (another indication of no POST on these besides no picture on screen). Bleh. Gonna fiddle with the SS51G a few more minutes before I give up, then concentrate on the SS56GV2. It has BIOS settings for voltage and such, though I'm leery of vmem because changing memory voltage has been known to cause these Shuttle systems to instantly die.

Grrrr, ticked that the SS51G was DOA because I had a good use for it.

 

[ammarcoux]

HHahahhaah finally got the MSI 845E MAX2-LR mortherboard to work! It POST no problem with the Mobile Celeron 1.8GHz I have.

As I said in a previous post, it was a really weird problem where the POST code would keep on looping between the chipset and mempry check. I check all the caps but they looked fine so I figured it must be a short or broken trace on the board. After looking at it for a bit I did find a broken trace which just so happened to go to the DDR slots. Coincidence I think not! I removed the varnish of the ends of the trace and bridged it with solder and it worked. Here's the picture of the broken trace with the solder. As you can the see in the green circle, there is a diagonal slash made from who knows what and the solder that bridges the gap. All in all my $15 eBay investment turned out ok


The board with the fixed trace

 

[TStep]

These little dogs are great. Just picked up another 1.5 that overclocks 1.0 Gig at least. Very little testing done so far though. Now, what to do with this one??

 

[sukasem]

There have mobile P4 (prescott I think 1M/533) that doesn't have speedstep feature.
So, we can get multiplier 21 from Mobile P4 2.8/1M/533

Any1 try it yet.

 

[ender11122]

Well i just got a celeron mobile 1.6 and a P4S800-MX and mine wont go over 145 FSB no matter what I do. Any suggestions. I have some Corsair 3200LL in with it right now, but still no luck

 

[ender11122]

When I go up to FSB of 180, it runs the mem test at the post screen and says "memory test failed" every time. I have it running in sync with the FSB with timings run by SPD, and the chipset Clock mode running on "async". Anyone have any ideas??

What is the second number in the CPU Frequency selection, the x's (180/xx)?

Are there any other locks or anything that should be enabled in order to OC these things?

 

[ender11122]

ANOTHER interesting point. When running 145 fsb I can get into windows, but when I start Prime 95, it runs fine but it pulls the volts way down. I run idle with 1.68 volts but right when I start prime it dips to 1.60. When I stop prime, back up to 1.68. Should there be that much variance in my volts? I thought these things did not pull much.

EDIT: I tried my antec PSU and same results, so the PSU is not the problem.

 

[Karaktu]

1) You may have a chip that might just not overclock;

2) Make sure that your voltage is 1.6ish;

3) Make sure your memory is set to 133 or 166; at AUTO or 200, you are giving it quite an overclock;

The second number "180/xx" is your PCI bus speed. 133/33, 166/33, 200/33.

 

[ender11122]

If I had to guess, I would say #1. Bummer . Thanks for the guide though, most helpful!!!

 

[imported_Kensai]

Moo.

 

[Zap]

Just discovered something pretty cool for mobile socket 478...

Asus Vintage barebones $102+6

This is a compact black mATX tower that uses the Asus P4S800-MX motherboard and looks to have a standard ATX power supply. $102 plus $6 shipping makes for a relatively budget setup. The motherboard usually goes for around $60 by itself, so the case is around $48 shipped. Not too bad, plus it's already put together. Anyone ever use one of these?

 

[ender11122]

sweet find. I may just use that for another try at one of these systems.