Unable to post at higher than 151 fsb

scalf

Member
Nov 29, 1999
37
0
0

Hi,


I am currently running a tbird 1.4 on a soltek sl-75drv5 mobo at 1533/133. The proc is unlocked and I am using samsung pc2700(from newegg). I have not been able to get past 151 fsb no matter what. My goal is to run at 1500/166 so only the proc is overclocked. Has anyone had this problem?

Thanks
 

Buz2b

Diamond Member
Jun 2, 2001
4,619
0
0
I don't have that combo but you might want to mention what you have tried and what hardware is installed in the system so that others can help out. Possibly mention what HSF you are using, what temps, what errors have shown, # of fans or how good the airflow is, what settings you have tried, etc, etc. You get the idea. And good luck! :cool:
 

scalf

Member
Nov 29, 1999
37
0
0

Current state

Tbird 1.4 @1533/133fsb 1.8v
420w topower powersupply
volcano6 cu
soltek drv5
samsung pc2700
radeon 8500
tekram dc-390u3w (seagate x15, fujitsu man 10K and wd enterprise 10K)
sblive 5.1
intel 10/100 nic

I have tested with

crucial 2100
no pci cards installed
voodoo 3 in place of radeon
lowered the multiplier so proc was underclocked at 166fsb

heat should not be an issue because the machine does not even post.
 

Buz2b

Diamond Member
Jun 2, 2001
4,619
0
0
In spite of the mention of heat not being a factor, I can see three things that stick out a bit; the RAM, the multiplier and the HSF. The Crucial RAM will overclock some but you might be pushing it a bit. Oops! NM, I see you tried PC2700 also. The other thing is that HSF. Not what I would consider an overclockers unit. Yes, it might still post with an inferiour HSF, but it's not an exact science. I'm not exactly a scientist in this area either. ;) Then there is the multiplier. Is this something you adjusted after you tried just upping the FSB? Or are you trying to run this at that level all the time? The 1.4 T-bird can be a decent overclocker, but there are some that do not get any higher than you. If you look at this page about overclocking levels of cpu's you will see many at or below your level (and yes, a great deal above) and you will also see your board mentioned at levels at or around where you are at. Maybe you can find some hints there also. I wish you luck.

Fixed Link, sorry!
 

scalf

Member
Nov 29, 1999
37
0
0

I been looking at the forum at amdmb and this seems to be a common problem. Lots of people can't get the drv5 to boot at 166 fsb. Anyone here have this issue with the drv or other boards?
 

Buz2b

Diamond Member
Jun 2, 2001
4,619
0
0
Keep in mind that there are not a lot of boards that will run at 166 very well, if at all. It may not necessarily be an issue with the board either; as I mentioned. Should you keep looking and trying? Heck yes! But you may not find an answer. Good luck though.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,765
31,758
146
The reason many are posting the same issue is that it is evidently a limitation of the board itself
Overclocking could be a nightmare to the impatient. This is not your bump your fsb and tweak your voltages thingy. Bumping the FSB is not that easy as you have to do a gradual increase in mhz. Going straight, say to 150mhz FSB, will cause the system to not even POST. But doing the FSB bumping in increments of not more than 8mhz per bump, at least in my case, should be alright. The maximum FSB overclock we can achieve with this KT333 board is 156mhz FSB. However that is only with a synchronous DRAM speed of 156mhz as the asynchronous 166mhz option will will only work up to 145FSB. I tried using their overclocking software built into the BIOS, RedStorm, but it would not work perfectly and would always hang everytime. Rebooting it after using the said overclocking proggie would only push the system to a certain limit. In this case where everything was set at max and both the memory and the cpu running synchronously at 133mhz, the RedStorm overclocking utility could only take the system to 145mhz fsb while the manual overclocking works up to 156mhz fsb. One feature that I really like to mention is the DRV5's ability to automatically reboot itself to the default fsb setting when overclocking went over the limits. There was never any instance that I have to clear the CMOS after a failed overclock nor did I even have to press the insert button to get the default settings up again. This is the first board I have tested that has that particular positive behaviour. To sum it up, we were able to take the DRV5 rock solid up to 156mhz FSB at 133mhz and only 145mhz FSB at asyncronous 166mhz memory bus. Our Athlon XP 1600+ (1400mhz) running smoothly and solidly at 1712mhz (11x156mhz). That's already XP 2100+ speeds.
link It would seem based on many users experiences with your board that 166/166 just isn't going to happen :(

 

rogue1979

Diamond Member
Mar 14, 2001
3,062
0
0
I just installed an Epox 8KHA+ and was having the same problems. I also have PC2700 and was trying a Radeon 8500LE. Also tried a V3. The two things that were holding me back were dimm voltage and lack of a better video card. I heard Samsung PC2700 didn't like high voltage, but when I moved up to 3.0v I hit 166MHz stable. When I ditched the Radeon 8500 and lost the V3, a trusty Geforce 2 Ti gave me 173MHz.
 

scalf

Member
Nov 29, 1999
37
0
0

Ok!, I got the board to post by jumpering it. When I tried this before I missed the 1/5 divider dip. Now it gets to the point "starting windows 98" and stops. There are no errors and it always halts at the same spot. I lowered the multiplier to run the proc at 1240 to reduce the chance that it is my problem. Then I slowed the samsung ram to the most conservative settings. I will probably try to bump the ram voltage next. I would have expected the samsung to run at this speed without any increase in voltage. Anyone else have to bump the ram voltage?
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,765
31,758
146
You can try bumping the DRAM voltage(it's there for that reason ;) ) but don't expect much, I seriously doubt the memory is the issue. I looked around and it just seems like the board doesn't like 166/166. If you get it stable at that setting then you are ahead of the pack on results.