Can't OC 850 past 900

[btvillarin]

System Specs:
Athlon T-Bird (B) 850
Asus A7V133 rev. 1.05

My vCore is set to 1.85v, and my VIO is at 3.45v. Now, I'm trying to take small steps to overclock my T-bird as much as I can. But, after running at 900 stable, I wanted to try 950. But, I could load into Windows without a fatal exception error each time. When I went back to 900, it's fine. I closed my L1 bridges using a conductive pen and a pin.

Don't tell me I can't go any further...I checked my processor, and it's not overclocked already or anything.

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Ncogneto]

Sometimes when the l1 bridges are not correctly connected you have problems overclocking with the .5 multipliers. give it a go at 1 gig and tell me what happens.

A little bit of knowledge is a dangerous thing!

 

[btvillarin]

At stock voltage or 1.85v?

BTW, why wouldn't they be correctly connected? I used a pushpin and conductive ink... :frown: If I did choke it, and I really hope I didn't, what do you suggest I do?

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[btvillarin]

Okay, I'm at 950 right now. (Weird...it didn't work b4... ) I couldn't POST completely, as in it wouldn't get to the part of detecting my IDE devices, at 1000.

Current Jumper Settings:
VIO: 3.45v
vCore: 1.85v
FSB: 100
Multiplier: 9.5
JEN: 1-2

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[btvillarin]

Okay, I was just transferring some mp3s from one partiton to another, and my system just restarted on me. (My temp was at 46°C and climbing.) Luckily, I could recover them from ScanDisk. But, that means it's not stable. I bumped back to 900 and a vCore of 1.80v.

Argh...

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[peteb]

are you upping your multiplier or your FSB?

What memory and memory settings are you using?

If you have good RAM, try increasing the FSB rather than the multiplier. Above 120FSB the PCI multiplier jumps to 4:1 bringing your bus under spec again.

If you can you could try bringing that chip up as a 133/266 FSB by lowering the multiplier?

-* This Space For Rent *-
email for application details

 

[btvillarin]

I hope you or someone can explain that concept about the bus under spec, that ratio, etc.

Anyways, I finally sucked it up and took out the T-bird to put more conductive ink on those bridges. I also cleaned and reapplied the AS2 because my temps weren't the way they should.

I set my vCore for 1.75v, my multiplier at 9.5, and FSB for 100. I posted. Plus, it ran through 3DMark 2000 a couple times. I'm gonna guess my ambient temp right now is 21°C. It's pretty cool that my temp is currently only 39.0°C and sinking 0.1° ever so slowly. I'm not sure when it'll stop.

I have a question about the Compunurse and setting it up. How close can you get it to the core without getting in the HSF's way?

Thanks for your replies! Man I'm pumped that my vCore doesn't have to be at 1.85v...

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Guest]

Ok guys, hate to piss you off, but I have my T-BIRD 850 at 1.09Ghz, and it WILL POST at 1.1Ghz, but it complains of 'Check sum errors'.... 1.85VCore, and the RAM is at 110Mhz....

System Specs.
Abit KT7-RAID
Athlon T-BIRD (B) 850Mhz -> 1.09Ghz

Yes, I'm the BOY who got Win98 to load in under 10 sec... ah....

 

[peteb]

right - the PCI bus runs at a multiplier of the FSB.

On Asus, there is a divider ratio for FSB to PCI. For a 100Mhz system, the divider is 3, for a 133Mhz system the divider is 4. The PCI bus is supposed to run at roughly 33.3Mhz, so this keeps the bus in spec.

Now, because we overclock, the PCI bus wanders out of range. At 100FSB, with a divider of 3 your PCI bus is at 33.3Mhz. We take this up to 110FSB and now it is 1/3 x 110 = 36.66Mhz. If this continued all the way to 132Mhz your PCI bus would end up at 44Mhz - oh dear. What Asus do is swich the multiplier at a mid point. 120Mhz and below, the divider is 3, so at 120FSB the PCI bus is at 40Mhz - probably still too fast, but then a 121FSB it drops to a divider of 4 (working back up to 133FSB) so now the PCI bus is underclocked at 30.25Mhz... We then get back to 33.25Mhz@133FSB and upwards....

The compunurse - well, I laid it out on the unmounted chip and decided it would touch the core without unsettling anything...

-* This Space For Rent *-
email for application details<P ID="edit"><FONT SIZE=-1><EM>Edited by peteb on 05/22/01 02:32 AM.</EM></FONT></P>

 

[peteb]

I don't think you are going to piss anyone off - it's good that you can get your system up at a good speed.

btvillarin is still using 1.75V so I think he still has options, but his cpu is nice and cool so no worries.

-* This Space For Rent *-
email for application details

 

[btvillarin]

Aaron: Guess what ...I'm pissed off... Just kidding! Hopefully, I can get that high as well.

Alright Pete, so between 100MHz and 132MHz, divide by 3 to get the PCI bus. After 133MHz, divide by 4. Correct? How can I find out the bus for each of my cards? I have a Soundblaster Live! Value, 3Com NIC 3C905B-TX-NM something, and an ATI Radeon 32MB SDRAM.

How did you get the Compunurse to touch the chip without letting the thicker cord get in the way of the HSF? I'm just going to keep the probe on, because I'll monitor my CPU the whole time. The distance from the edge of the ceramic to the edge of the core is a little farther compared to the length of the "flat" part of the probe.

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Ncogneto]

Random restarts can be a tell tale sign of inadequate power. Glad to see you are making progress however. One thing you might try is to remove your CPU fan from the motherboard header and get a three to 4 pin adapter to run it straight of a power supply plug. Also, I find running 3dmark while playing a looping mp3 a good test to check stability.

A little bit of knowledge is a dangerous thing!

 

[btvillarin]

I'm not sure if you read more than what I think you got to, but I fixed that problem.

System Specs:
Athlon T-Bird 850
Asus A7V133
Swiftech MC370-0A
300W Power Supply
ATI Radeon 32MB SDRAM
SB Live! Value
3Com NIC
HP CD-Writer 9150i
IBM 75GXP Deskstar 30GB 7200RPM ATA100
Teac Floppy
2x80mm Sunon High Speed case fans

BTW, the cooler is a 4 pin already. Thanks for that tip on looping mp3s while running 3DMark. I'll try that in my testing.


<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[peteb]

noooo up to 120FSB, divide by3, 121 onwards divide by 4!!!

Youy can only findthe limit of your PCI cards by old fashioned trail and error. What I do is remove anything not needed - ie, anything but the video card and go from there.

I don't use a compunurse, my probe can just about reach. If yours doesn't then you obviously cannot set it against the die. I don't think there is an accurate way you'll judge the surface temp of your cpu using it.

-* This Space For Rent *-
email for application details

 

[Ncogneto]

I just found it odd a 850 T-bird not doing 1 gig. I have so far gotten 4 of them and all did 1 gig @1.75 volts no problem. In addition to this they can all do an additional increase using FSB overclocking. I do however choose to do the hard mod I find it much more relieable in my own expeirances.

A little bit of knowledge is a dangerous thing!

 

[btvillarin]

Oops...my bad....I understand now, thanks.

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[btvillarin]

So they all reached 1GHz @ 1.75v...so what do you think is wrong with mine? I don't know what to do. Did you set it at 133x7.5 or just 10.0x100? Plus, what's a hard mod? If you mean changing the settings with the jumpers, that's exactly what I do, too.

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Ncogneto]

A hard mod is a little more involved And yes I ran them at 10 x 100 then upped the FSB to around 105 to 107, memory settings at 133 ( which netted a 137 to 140 memory bus). I have not yet jumped into the 133 FSB just as of yet as I have not had a need to, and I am wiating for the chipsets to sort themselves out and mature a bit. I did have the pleasure of setting up one epox DDR Board for someone else but really didn't have it around long enough to play with it ( he was a little impatient). A good place on how to do a hard mod can be found here.

<A HREF="http://www.ocinside.de/index_e.html" target="_new">http://www.ocinside.de/index_e.html</A>

This is especially usefull for those people who wish to Oc there processor but have no such functions available to them via there Mobo. Oh, you need to follow the link to the article that reads "paint your t-bird"

Something you may wish to try is to pencil your l7 bridges all closed. This will give you 1.85 volts at boot, which then should be overriden by bios if Using the l1 trick. But then booting does not seem to be your issue.

From one of your previous threads I still think that it is at least a possibility of your power supply being inadequate. Overclocked processors do tend to draw more power then there simularly clocked stock parts (ie say a stock 1 gig compared to a Oc 850@ 1 gig).

A little bit of knowledge is a dangerous thing!

 

[btvillarin]

Why do I need to pencil my L7 bridges closed to get me 1.85v at bootup? So I don't have to worry about not POSTing due to a low vCore? Then what about the jumpers; after POST, will it go to the voltage I intend it for?

The Asus A7V133 allows me to change the multiplier, FSB, VIO, and vCore through jumpers OR the BIOS in jumperfree mode. Plus, how can my CPU draw more power if 1.75v, for example, is set for the voltage? Are you saying I might have to get 350-400W power supply instead of using the 300W I have now?

"I have not yet jumped into the 133 FSB just as of yet as I have not had a need to, and I am wiating for the chipsets to sort themselves out and mature a bit." What do you mean by the chipsets to sort themselves out and mature?

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[btvillarin]

I just wanted to quote something from the article you referred me to: "If your mainboard allows all multiplier settings it's only necessary to connect the four L1 bridges."

My motherboard allows all these changes through DIP switches and/or jumpers.

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Guest]

Hey, btvillarin! Try 1.85v with 110Mhz FSB, and a multiplier of 9x. This is what I currently have my system at (1.09Ghz seems to be harder and harder to keep stable. My cooling system worked better in the winter. If really wanna know why, I'll post it later). Thats 990Mhz, and its stable. Give it a shot.

Yes, I'm the BOY who got Win98 to load in under 10 sec... ah....

 

[Croaker]

I have an A7V133 with tbird 800@933(7*133) At 1.85v. With a global win fob32-1. Artice silver is ordered but not in. Some one stated the divider for the PCI was 1/3 up to 132. This is not true on the A7V133. In the BIOS is states at 120 the divider is 1/4. You should try 7.5*133(1000)or 7*133(933)

 

[btvillarin]

Right now, Aaron, I'm at 1.85v at 1000 (133x7.5). So far, after running 3DMark for about 2 hours on loop, it hasn't crashed. I should probably run SETI@Home also for awhile too. But as of now, I don't think I'll go any higher. It's been cloudy lately, and I think when it gets hotter, then heat will be way more of a problem. (Since you live in California also, I live in Arcadia near Pasadena. Weird weather, though...) Personally, what did you do to test your stability. For me, I run a 20-minute 3DMark2000 benchmark with an mp3 repeating in the background. Then, I run SETI@Home in the background [all the time] while surfing the Internet (mostly reading around the forums and stuff), running AOL Instant Messeger and MSN Messenger in the background.

What was your cooling system, by the way?

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>

 

[Guest]

Well, to check stability, I run WinAmp looping MP3 (Duel of the Fates, of course! Irony!), and Sandra 2001 'Burn-In' on continuous CPU benchmarks for about an hour or so. It works great. Bad speed, it crashes, if its good, it keeps on 'fighting' away.

Ah... the cooling system. Um... a Chrome Orb on my KT7-RAID (It had to be modified in order to fit.) And 3 intake fans that are fitted on the side of my case. I cut holes out for them. And one exaust fan out the front (some 7000RPM no name fan. I know its 7000 because its hooked into the system fan header.) It sits inside of my case between the CD-RW and the floppy drive. There is so much air flow, you could blow dry your hair with my case. (Although the air is cool, not hot.) I said that it works better in cooler weather because in the winter I could get system temps of about 15C and CPU temps of about 29C (CPU= OC 1.09 Ghz 1.85v)

I'm a cooling nut. Next is a water block system. I'm starting on it soon...

How the hell u get an Athlon (B) at 133Mhz FSB?!?!?! Good stuff? What is ur cooling system!!!

And one last thing. You divide by *3* until 120Mhz and by 4 at 121 and greater on KT133 chipsets while KT133A chipsets go and divide by 3 the whole time.

Yes, I'm the BOY who got Win98 to load in under 10 sec... ah....<P ID="edit"><FONT SIZE=-1><EM>Edited by MegaAaron on 05/22/01 03:05 PM.</EM></FONT></P>

 

[btvillarin]

How the hell u get an Athlon (B) at 133Mhz FSB?!?!?! Good stuff? What is ur cooling system!!!

I forgot if it was an Athlon (B) or what. I'll have to read the core sometime, write it down then post it. But right now, it's at 133FSB. Is that bad? LoL...And, are you being sarcastic or something? It's hard to tell when you said "Good stuff?".

I'm using a Swiftech MC370-0A, and my two case fans are 80mm High Speed Sunon ball bearing fans pushing 42.5cfm, I believe. I would mod my case, but I can't spend anymore $$$. :frown:

BTW, I have an Asus A7V133 rev 1.05, and it's a KT133A. So, I guess I always divide by 3...peteb said something else, so I don't know now...

<i>OC...unless your computer's cheezy (is that a good rhyme?)</i>