A troublesome P1.6A?

[esmo]

My first post here, hoping to get some insight from the friendly AnandTech community

So, I replaced my old trusty Coppermine 800MHz system with a P4-1.6A, a P4S533 mobo and 2x256MB Kingmax PC2700 RAM. Everything works pretty smoothly except that the CPU doesn't seem to overclock to the uber-high speeds some (seems like most) of the 1.6As do.

More specifically, it doesn't seem to go over 2.2GHz (138MHz FSB). That's on stock voltage. System seems pretty stable on 140 too but fails on Prime95 (rounding errors). Anyway, even going to 145 FSB just doesn't work. Funny thing is, I can try even 1.8V+, but the system is still unstable even on the desktop, never mind any stress tests. So basically seems like there's a wall for the CPU, under which it runs comfortably with no real tweaking at all, but which it won't go over even with silly voltages.

I don't think memory is the problem, I'm using it at 4:5 with the 138MHz FSB, and 1:1 with much higher FSBs doesn't help. I'm using the stock HSF with AS3, and temperatures seems to be well under control, around 50°C max under load.

So do I just have a not-so-stellar overclocker here or have I missed something, or maybe there's something else I could try? I'm not really disappointed if 2.2G is as far as it will go (especially since the current vid card is holding it back by silly amounts), but I haven't really overclocked anything since my Celeron 300A so I'm a bit out of the loop here

Here's the relevant system specs again:

P4 1.6A (SL 668, Costa Rica, can't remember the date) w/stock HSF & AS3
ASUS P4S533 (BIOS 1.007C (IIRC))
2x256MB Kingmax PC2700
Creative GeForce 2 GTS (yes, I'm replacing it within a few weeks)
350W PSU

 

[THUGSROOK]

your problem is in your bios settings, or your motherboard just plain cant run any higher.

its not the cpu.
its not the mem. (unless its a wrong bios setting etc)
pci speed arent into play yet.

you are using a real P4 psu right?

 

[esmo]

I was wondering about the BIOS settings myself, but really couldn't find anything that could be amiss. Guess I'll try some manual memory settings and maybe rip the other bank out to see if anything happens.

And yeah, I bought a new PSU since my first thought was that my old 230W one just wasn't cutting it. I'm using all 3 power connectors on the P4S533 if that matters any.

 

[Duvie]

generally the p4s533 mobos have been doing well up to 145-150...at 138 -140 and having problems is loweven for the sis mobo...

It is really critical to know the date of the chip...early on 1.6's weren't as good as today and 2.133ghz wasn't even assured back then...maybe you have a 001 chip or packing date of late 2001....Rare but there have been a few...


The wall so to speak has always seemed to be mobo or hardware related....however in my experience rounding errors are usually related to cpu, power or heat....a hardware in prime95 would have been a ram timing or mobo memory chipset issue...

what brand power supply??? Are you using ezplug (4 pin standard molex connector) or square P4 plug??? What is the amps on the 12v rail???

 

[vfdfs2]

Are you using a USB mouse?

I could not clock my 1.6A beyond 2.2GHz on my Asus P4B266-C, too, but eventually figured it out. It was the Microsoft Intellieye Optical USB mouse preventing me from overclocking. Used a PS/2 to USB adapter, system is @ 2.7GHz (can go as high as 2.8, but not stable).

Original thought it was the memory, then the motherboard, bios, changed everything from videocard to hard drive to case to the goddamn power supply. Nope, nothing work, finally went to change the keyboard and mouse. I have a Master in electrical engineer, i dont have a clue why a USB can do anything with the ability to prevent it from overclocking.

 

[Astro33]

seems like i have similar problem.
My system doesn't run stable enough about 140 FSB, 1.55 Vcore, everything else default. It could be the USB intellimouse. But having USB is so much more convenient!

P4 1.6a (5/22/02), stock hsf, Asus P4B 533 w/audio, 2x256 MB Samsung PC2700, bios rev 1004, XP Pro.
Tried up to 1.6 V for Vcore, it would not run stable at 145 FSB.
I don't know what the normal temperatures are but at 2.2 Ghz it is 35C/46C and 37C/49C at 2.3 Ghz. Seems a bit high. Do you people suggest those fans that mount right across from cpu fan and sucks the air away from cpu directly out of the case through a duct? not sure if this setup of cpu fan blowing air on cpu while an exhaust fan competes with the cpu fan for air.

 

[esmo]

Thanks for the replies and suggestions, everyone.

Yes, the power supply is ATX. The mobo has 3 connectors, the "old" standard type, the 4-pin P4 and the 4-pin that HDDs also have. Like I mentioned, all of them are used, and voltages seem regular enough, not like it seems like that PSU has any kind of problems.

I have a Logitech MouseMan Dual, but I'm pretty sure that I'm using the USB-PS2 -connector with it. I'll try unplugging my joystick from the USB port too, though

I'm more or less guessing that in the end the CPU is the root of all evil however. I'll be happy to discover otherwise though It should be a pretty new date P4, I'll check the date and see if I can find anything out by trying the things that I mentioned and were suggested, and report back.

 

[galtsfan]

I'm not sure but I don't believe you need the "ezplug" connected. That is for folks who don't have a PS with the ATX 12v (4pin) plug. I also have a board with both available (P4B533-V), but I'm only using the ATX 12v plug.

 

[vfdfs2]

I have my vcore voltage @ 1.625V, BTW.

 

[Duvie]

Good point forgot about the usb thing...I never had a mobo with a usb problem..."knock on wood"...I guess it appears epox is better then asus at a few things.....

166fsb and runing usb printer, scanner, and dig camera in system....My MSI had no problems with it as well as my sis645 mobo from asus...


Disable in windows then disable in bios...unplug all usb devices...then try it...

If it is the usb...dump the board...usb devices are convenient and there are plenty of boards who got their act together...This was an early aus mobo so they have fuixed it I believe...

 

[esmo]

Well, tried out a few things. Didn't help much, but here's the rundown anyway.

The mouse did use the PS2-USB -connector, but unplugged my USB joystick anyway. And I know you probably don't need to use the standard 4-pin power connector, but I just plugged it in just to be safe, knowing that some people have done so. For the below tests I did remove it just in case.

So, the initial setting was at 138MHz FSB (2208MHz), Vcore at 1.5V, mem at 4:5 and 2.7V (I'm sure the memory would run at 2.5V, but just forgot to set the jumper back). Everything ran smoothly, no errors in Prime95 (at least after about 4-5 hours of running it). Temperatures under load were about 47°C/116°F for CPU and 41°C/105°F for mobo (MB temp seemed high to me compared to the CPU, but not sure what it should be at)

Tried 141MHz at 1.5V. Prime95 failed at less than a minute. Upped Vcore to 1.575. Same thing. Also with 1.65V. Figured that higher voltage wasn't going to fix it.

Took off the case, set the computer on it's side and propped an extra case fan over the CPU and vid card so that it blew air over the SiS chipset. Also removed all PCI cards and the 2nd stick of RAM. This time it actually ran for a while (didn't have a lot of time, so ran Prime95 for about 45 mins or so), temperatures were about the same as with the initial setup, slightly lower for MB. Then tried 145, but WinXP barely managed to start. At 150 the system doesn't even POST.

So, backed up to the start settings, which still run OK. I'm still a bit perplexed though, guess it would help to try a verified 2.4GHz 1.6A on the 'board to see if it would work, but I don't happen to have any of those at hand. The CPU's packdate is 6/11/02, so it's certainly pretty new.

So, could the culprit still be the motherboard? What should the MB temps be at anyway, and where are they measured?