Johnny 5 is alive! (sort of)

[jaybird]

Hi people!

I'm back!

However, things are not the way I would like them.

UPDATE:

PIII conversion on 0.41 board did not work, went back to 0.40 board and the board nolonger posted and I fried one of my beloved PIII 1100E's

So, what to do, just for grins, I cut the zif socket wires, removed NEO370's and 1100E's, installed an "unknown" lapped pair of 366's and went for broke!

BAM! went straight to 550, then to bp6fsb and right to 561 (102 fsb) with out a hicup!

Right now I am on a BP6, dual 366's @ 561, 768 meg PC133 SDRAM, dual hard drives, 4 optical drives, etc, no problems!

However, my +5 volt side is very low (4.72 volts), I have tried 2 power supplies and each is low with everything installed, so, I am going to go to the dual PS setup, a 400 watt unit to power up everything except the MOBO and a 300 watt (high +5 volts) unit for just the MOBO. If this works, my next step is to lap 2 433's that I have and try them, then 2 500's, I will keep you all posted.

If this works, it means we can maybe all over clock to at least 100 mhz fsb with only a $15.00 USD investment! (basic parts, not including PS).

Stay tuned for "the rest of the story".

Regards,

jaybird

[Derek]

He's BLEEDING!

[Compulsivephreek]

so you think my 466 could go 100 fsb ? one has a Intel cooler and the other has a huge aluminum block with a 60 mm fan and it actually runs cooler than the other cpu ! hey, is the CPU 1 towards the rear of the board while cpu 2 is near the front of the case ? the voltages are almost the same, but one is maybe 2.05v while the other is right at 2, is that acceptable ? the power supply is an Allied 400w, and I only have one hard drive, one optical, and the stuff in the sig. Case runs 42C, but the cpus read 35 and 38c, go figure. I know that case temp is high, but this is very generic case, no front or rear fan so I use the power supply and a PCI slot cooler to draw hot air out, computer is stable. I have a feeling that case temp is not accurate, how could the cpu run cooler that the air around them. I may test that temp and see what it really is.

[davd_bob]

Hi Jaybird. I didn't even know you were missing.

PIII conversion on 0.41 board did not work, went back to 0.40 board and the board nolonger posted and I fried one of my beloved PIII 1100E's

When it rains, it pours.
About that 1100...OUCH.

...installed an "unknown" lapped pair of 366's and went for broke! BAM! went straight to 550, then to bp6fsb and right to 561...

Never any doubt a pair of 366s go to 550 in hands as capable as yours.

However, my +5 volt side is very low (4.72 volts),

Is that really out of spec. I though a +-10% was allowed, although exact voltage should always give better results.

my next step is to lap 2 433's that I have and try them, then 2 500's, I will keep you all posted. If this works, it means we can maybe all over clock to at least 100 mhz fsb with only a $15.00 USD investment! (basic parts, not including PS).

He's BLEEDING!

I gotta go with Derek on this one. Im sure its only a Pipe Dream. (But there are rumors...)

...Stay tuned for "the rest of the story".
Regards,
jaybird

We will be waiting to hear what sort of cooling solutions you use.

Compulsivephreek,
In case no one else said it...welcome aboard.

...so you think my 466 could go 100 fsb...

Dual 466@700 Not a chance unless you use peiltiers, and/or water cooling and then not likely.
The Mendocino core hits a wall about 550-600MHz and it can only be passed with a goodly amount of "creativity" in cooling. Most of us think thats why Intel abandond the chip after releasing the 533.

...Case runs 42C, but the cpus read 35 and 38c, go figure...I have a feeling that case temp is not accurate, how could the cpu run cooler that the air around them...

It is common for a BP6 system temp to register higher then the CPUs. Also diff temps on CPUs is normal. In fact you will rarely see them the same temp.

...one is maybe 2.05v while the other is right at 2, is that acceptable ?...

Totally acceptable, as long as they are set to 2Volts in the BIOS. If you are running at 66FSB you should try LOWER the voltage. It will lower the temps on the CPUs, you can do this in Soft Menu II. Check out my sig for clean low voltage on OC'ed 366s. ALSO...make sure you have grease under the "Greene" and if you can score a fan to put on it thats even better. It will help the system temp.

...I use the power supply and a PCI slot cooler to draw hot air out
...

Excelent. Thats more cooling then lots'a folks have.

...computer is stable.

Now THATS what I'm talk'n 'bout !

[kuun]

ok... im not gonna do what david did.. but

to contradtict him.. both my533's run at the same temp... according to MBM5

a pipe dream indeed but it still may succeed props for trying tho..

yes the mendocino cores are really only stable around 561 and lower... some can be around 60 with some really ncie aircooling.. but that also reuires good cooling on the greenie heatsink on the bx chipset as well.. which.. i need to do one day...

dude those PCI slot exhaust fans are great.. im going to get one for all my machines..and maybe 2 for the bp6 SCSI server (dual cases with dual ATX poewrsupplies)

hearing another bp6 being stable.. thats fantastic! score 432857842750943875 for bp6

[kuun]

ok... im not gonna do what david did.. but

to contradtict him.. both my533's run at the same temp... according to MBM5

a pipe dream indeed but it still may succeed props for trying tho..

yes the mendocino cores are really only stable around 561 and lower... some can be around 60 with some really ncie aircooling.. but that also reuires good cooling on the greenie heatsink on the bx chipset as well.. which.. i need to do one day...

dude those PCI slot exhaust fans are great.. im going to get one for all my machines..and maybe 2 for the bp6 SCSI server (dual cases with dual ATX poewrsupplies)

hearing another bp6 being stable.. thats fantastic! score 432857842750943875 for bp6

[jaybird]

I am running GlobalWIN FEP-32's, lapped, as are the 366's and of course Arctic Silver and a "mini-peltier" on the BX with a lapped "greenie" and Arctic Silver on both the BX and on the top of the "Greenie" where the Pelt is mounted. On startup I see an initial drop of over 15 Deg. F. on the BX before it begins to heat up.

Temps are 96 to 98 Deg. F on the cpu's and 85 Deg. F on the BX.

Regards,

jaybird

[davd_bob]

geez,
Dual 366@605 commong right up !

Lap the 433s and see if you can get 96FSB !!

[jaybird]

I couldn't get past the 100-102 thing so just for "grinns" I stuck in the 2 original 366's that I thought I had "cooked" when I had left my 72W Peltiers on and no fans (other board). Guess what? They went straight to 104 FSB! However, not stable.

So I said to my self, self, try one old and one newly lapped 366 (the 2 that were "old" wouldn't take the same voltage).

I did that and 106 there we were (sort of) but not stable so I dropped the cpu voltage from 2.20 to 2.10 on both and now all is stable at 583.

cpu's at 89 Deg F on #1 and 93 Deg F on #2, BX at 96 Deg F with Mini-Peltier. CPU's @ 10% and 99meg of ram out of 768 as we write this.

Will "burn in" and go for 594 in the morning!

Regards,

jaybird

[jaybird]

No, I don't think all celerons can go to 100 fsb.

However, that said, I do believe we can ALL achive better overclocking with a few (under $20 USD) mods, thats all I am trying to prove.

To date, I have taken a board that was not stable @ 100 fsb and made it stable with 366's @ 100.

I have by lapping the cpu's and using the biggest heat sinks available for the bp6 and using just a few simple mods required for possible PIII use (regulators and caps) made a version 0.41 board stable at 100, 104 and 106 fsb with 366's, I can boot at 108 and go into windows and 110 but not get into windows.

I DO NOT belive that the entire limitation is due to the cpu's, I have noticed that once the BX gets above 34 Deg C (registered at the "greenie", not the actual chipset itself) that my system becomes unstable, I have noticed that by actually REDUCING cpu core voltage that I was able to get to another "step".

Therefore (skuz spelink ) I belive with proper cpu selection (come-on, cerlies are SO CHEAP today on eBay!), cooling and attention to the BX, anybody with a STABLE bp6 and $20 USD or less, can get to AT LEAST the 100 to 104 fsb range (we will see about other than 366's this week ).

Regards and "flames accepted".

jaybird

[purrkur]

However, that said, I do believe we can ALL achive better overclocking with a few (under $20 USD) mods, thats all I am trying to prove.
[...]
Regards and "flames accepted".

Hehehe! Good to see you are putting on your asbestos underwear and getting ready for those flames

I think overclocking is good fun and when pushing the limits I always learn something new. I would just like to point out one thing: When your hardware is at the limit, data can be damaged and apps can become unstable because of the overclock even before the computer itself along with operating system becomes unstable or crashes.

I once read a weblog from a Microsoft developer who got a job of going through crash reports (I know you all have seen the annoying "sorry this app has crashed, do you want to send a report to MS??" type of message). What was interesting to read was that a tremendous amount of crashes he looked at came from overclocked computers!

My .02c worth on the subject is as follows: I love overclocking and testing the limits just as much as any other. However, once I have found the limit, I will back down a notch, maybe two. I value a stable system much more than a fast one that can turn unstable.

Just to take your example: You got two 366's that run at 100MHz FSB. Great stuff in my opinion especially since we all know that the ones that are stable at that frequency usually don't give any issues whatsoever. You have been pushing it though. Lets say you end up running your system at 104MHz FSB. That would give you 572MHz on the CPU's which is only a 4% increase. That, my friend is nothing you will be able to feel. However, you might definitely be running at the level where your data is being distroyed, bit here and bit there until ultimately your system goes down hard.

If you check out my testing where I had two Celerons running at 600MHz (among other things, over here, you will see that at 600Mhz the Celerons were way out of breath and it was almost identical in performance to the 576MHz Celeron setup in the CPU intensive tests.

So no, this is not a flame , but just my .02c worth on overclocking in general, including on the BP6. But please, let us know how your overclocking experiments go and keep us posted on your progress!

[Billl]

However, that said, I do believe we can ALL achive better overclocking with a few (under $20 USD) mods, thats all I am trying to prove.
[...]
Regards and "flames accepted".

Hehehe! Good to see you are putting on your asbestos underwear and getting ready for those flames

I think overclocking is good fun and when pushing the limits I always learn something new. I would just like to point out one thing: When your hardware is at the limit, data can be damaged and apps can become unstable because of the overclock even before the computer itself along with operating system becomes unstable or crashes.

I once read a weblog from a Microsoft developer who got a job of going through crash reports (I know you all have seen the annoying "sorry this app has crashed, do you want to send a report to MS??" type of message). What was interesting to read was that a tremendous amount of crashes he looked at came from overclocked computers!

My .02c worth on the subject is as follows: I love overclocking and testing the limits just as much as any other. However, once I have found the limit, I will back down a notch, maybe two. I value a stable system much more than a fast one that can turn unstable.

Just to take your example: You got two 366's that run at 100MHz FSB. Great stuff in my opinion especially since we all know that the ones that are stable at that frequency usually don't give any issues whatsoever. You have been pushing it though. Lets say you end up running your system at 104MHz FSB. That would give you 572MHz on the CPU's which is only a 4% increase. That, my friend is nothing you will be able to feel. However, you might definitely be running at the level where your data is being distroyed, bit here and bit there until ultimately your system goes down hard.

If you check out my testing where I had two Celerons running at 600MHz (among other things, over here, you will see that at 600Mhz the Celerons were way out of breath and it was almost identical in performance to the 576MHz Celeron setup in the CPU intensive tests.

So no, this is not a flame , but just my .02c worth on overclocking in general, including on the BP6. But please, let us know how your overclocking experiments go and keep us posted on your progress!

Purrkur makes some good points. Just like to add my .02c worth also. It's very difficult to tell exactly how far you can OC and not cause any damage. I usually try to be on the conservative side and watch for odd behavior. Window's apps die all the time for seemingly no reason. One of the things I like to do is run something very processor intensive and see how the system reacts over time. Like Purrkur said you can very well be slowly destroying your data over time. One of the problems is it creeps up on you. You don't know your hosed until it's way to late.


Billl

[jaybird]

I agree totally!

That is why, even though I can boot and do some work @ 583, or even 594, I have chosen to back down to 572 @ 104 FSB. This seems to be the "sweet spot" for this system.

572 @ 104 FSB and 2.10 volts is "max" for this board and CPU's, period!

Still, not to bad for what I've got to work with and what it has cost me !

I am still going to try dual power supplies. I believe that various voltage levels and current available to either MOBO or aux units is critical, not just total "WATTS" available.

I am going to use a 450W unit for "aux" delivery and a 300W unit for the MOBO alone (this unit has shown a very high and constant +5 volt level that the 450W unit has not shown).

Lets see what happens!

Regards,

jaybird

[jaybird]

I agree totally!

That is why, even though I can boot and do some work @ 583, or even 594, I have chosen to back down to 572 @ 104 FSB. This seems to be the "sweet spot" for this system.

572 @ 104 FSB and 2.10 volts is "max" for this board and CPU's, period!

Still, not to bad for what I've got to work with and what it has cost me !

I am still going to try dual power supplies. I believe that various voltage levels and current available to either MOBO or aux units is critical, not just total "WATTS" available.

I am going to use a 450W unit for "aux" delivery and a 300W unit for the MOBO alone (this unit has shown a very high and constant +5 volt level that the 450W unit has not shown).

Lets see what happens!

Regards,

jaybird

[davd_bob]

For what its worth I think 300A and 333s might get over the 100FSB hump. I don't have any nor do I plan to get some to try it.

From perusing this forum I think cooling the BX is the key to excessive FSB and getting a low enough multiplier on the Celly.

[jaybird]

Well, I guess the 366's @ 572 (104 FSB) are still tops if FSB is your goal.

However, I now have 2 433's @ 604 but I don't like the 92 FSB.

They will not go any faster

Next will be a lapped pair of 500's ( I don't expect much because of the multiplier).

I am still going to try dual PS's, a 300 for the mobo and a 450 (just because I already have them) for everything else. My +5.0V seems to be getting dragged down for some reason.

daved_bob, I believe your right about the BX, everytime I try to go above 104 my BX temp goes above 94 Deg. F (at the heatsink) and I get lockups, maybe I need a stronger Pelt on it, I may even try cooling the SouthBridge to see if that helps.

So far I have about $15 USD invested in the "upgrades", had most everything I needed in the scrap bin

Regards,

jaybird

[davd_bob]

I think the bench purrkur did is the best ever done on a BP6 and possably the best comparison presentation on any mobo. It ought to be "required reading" for anyone thinking about OC'ing anything.

purrkur mentioned a pair of 400s. I kinda wish he included that pair running at 600 on 100FSB. With his skills Im pretty sure he could have gotten them to work. I believe the 533@600 he used had "issues" because of the OC on the rest of the system and issues were not CPUs generated.

BTW, other then my 366@550 I never OC a mission crittical system. In fact, mine are usually underclocked because no-one I know can tell the difference when they are inputting data on a spread sheet.

[purrkur]

I think the bench purrkur did is the best ever done on a BP6 and possably the best comparison presentation on any mobo. It ought to be "required reading" for anyone thinking about OC'ing anything.

Thanks for those kind words

purrkur mentioned a pair of 400s. I kinda wish he included that pair running at 600 on 100FSB. With his skills Im pretty sure he could have gotten them to work. I believe the 533@600 he used had "issues" because of the OC on the rest of the system and issues were not CPUs generated.

Hmm. Actually the board did just fine at 75MHz which is what you need to get a 533 Celeron to run at 600MHz. Both 400's and the 366 ran just fine when clocked at 75MHz fsb. It was just the 533's that were unhappy at that fsb Interestingly enough, I never tried my 400's at anything over 75MHz fsb. I guess the reason was because I had never read of anybody doing anything great with those plus the fact that I value a stable system that I never figured it would worth my time to play around with them.

BTW, other then my 366@550 I never OC a mission crittical system. In fact, mine are usually underclocked because no-one I know can tell the difference when they are inputting data on a spread sheet.

Hehehe! "underclocking" is actually worth it on lazy and warm summer days or when you just want to make your system quiet. My main system (Athlon 2GHz) has a cpu fan that can run at three switch-selected speeds. The only overclock I have done on this system is running the fsb up to 167MHz (333MHz effectively) while lowering the multiplier (Athlons are only multiplier locked when going up, not down). I have also reduced my multiplier to be able to run the cpu fan at its lowest speeds to reduce noise and heat generation in the summer. But I'll admit that I do require cpu power and I rarely run any type of office programs except for spreadsheets every now and then.

Btw, I noticed the machines you are running. I did the same benchmarking I did on the BP6 on three (older) different machines which turned out to be very interesting. The machines were:

1. Celeron 533MHz (66MHz fsb), 256MB memory, Intel ZX chipset
2. K6-2 500MHz (100MHz fsb), 256MB memory, VIA MVP3 chipset
3. Sun Ultra 5, 360MHz UltraSPARC IIi, 384MB Memory, unknown chipset.

Machine 2 & 3 had the same type of disk drive (40GB Seagate) while the Celeron had a Maxtor 80GB drive. I have tested both on a single machine and I found that the maxtor drive is slightly faster overall than the Seagate. All machines were running exactly the same version of Debian (stable).

The results are found in a spreadsheet . If only I could find the time to do a proper writeup and cough it up on the web. The results were actually quite interesting even when comparing the single Celeron 533MHz machine to my BP6 tests.

[jaybird]

I went back to the 366's @ 104, quite stable, really!

I wanted to do a complete system burn-in so I tried something new (at least for me).

I ran CPUBURNIN and BURNIN by Passmark at the same time. Both CPU's went to 100% and it used 600 meg of ram while also running hard drives, floppy drive, and all 4 of my optical drives as well as maxing out my 128 meg dual head vid card!

A funny thing happened after aboout 8 hours of this, my cpu temps had DROPPED about 10 Deg F (MBM 5) under full load and my BX had dropped from 105 Deg F under full load to just over 92 Deg F (based on a thermocouple in-bedded in the heat sink between the BX and the chipset).

Now, the most interesting thing, I ran the entire battery of tests again, only this time at 106 FSB (583 Mhz)- 10 hours without a single glitch!

However, while on-line I did get an explorer.exe failure once during a 2 hour session (did not have to re-boot, only reopen my ISP and continue on with out loosing anything).

Was this over-clocking related? Don't know but will keep all informed.

Regards,

jaybird