overclocking from 366 to 550MHz

purrkur · Post by **purrkur** » Wed Jan 05, 2005 5:40 am

I have completed the first modifications to my BP6 motherboard which was to replace 9 caps that were leaking or bulging. I have hooked up the board, installed Debian and done some testruns on it with a pair of 366MHz Celerons.

The board booted but only one CPU was reported. I noticed that the BIOS was ancient (from 1999 with HTP BIOS version 1.21) so I upgraded to the latest and greatest and rebooted. Sure enough, both CPU's were detected and things started rolling. The board was stable during installation and configuration so I decided to move the fsb up to 100 to see what all the fuss is about

After doing that, the board refused to boot properly, getting stuck somewhere during the OS startup. I remembered reading here that cooling the BX was very important at 100MHz fsb so I replaced the heatsink with a large passive chipset cooler (with the necessary gunk inbetween). I also installed heavy duty coolers on the CPU's.

I rebooted and the BIOS reported two 550MHz CPU's but the OS only saw a single one for some reason. Back to BIOS, I changed the 550(100) CPU setting to user defined and set the parameters there instead. I also upped the CPU voltage to 2.1 volts. Reboot and Bob's yer uncle. The machine ran at 2x550/100 without any problems even during some CPU stress testing that I did. However, I checked with "sensors" and found that my Vtt voltage would stay at 1.46-1.47 volts. I have not changed the caps around the ZIF sockets (none of them were bulging).

So, questions to those who have 366MHz Celerons overclocked to 550MHz:

1. Have you also experienced the strange problem with BIOS reporting two CPU's while the OS only sees one?
2. Have you had to increase the CPU voltage in order to get 550MHz to work?
3. What is your Vtt voltage (especially if you haven't done any Q6 mods)?

hugoc · Post by **hugoc** » Wed Jan 05, 2005 8:50 am

1. Have you also experienced the strange problem with BIOS reporting two CPU's while the OS only sees one?

No - only where the OS doesn't support SMP. I've gotten SMP in WinNT, 2K, XP, BeOS, and Linux.

Because the problem went away when you bumped the voltage up, I suspect your CPUs are not equal. One CPU probably wasn't working. You might find that you can run one CPU at 2.0v and the other at 2.1v and still get SMP, of course, which CPU is which you'll have to establish by trial and error.

2. Have you had to increase the CPU voltage in order to get 550MHz to work?

No - see .sig. But I have a good pair of chips. When C366 was first out I read stories of people bumping their voltage up to 2.3-2.5v and using peltiers to reach 550MHz, so if you have an earlier chip, count yourself lucky!

3. What is your Vtt voltage (especially if you haven't done any Q6 mods)?

Hmmm... I don't seem to see that. Here's my sensors output:

eeprom-i2c-0-52
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 128

eeprom-i2c-0-50
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 256

w83782d-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 5000
VCore 1: +1.89 V (min = +1.81 V, max = +1.98 V)
VCore 2: +1.39 V (min = +1.81 V, max = +1.98 V)
+3.3V: +3.36 V (min = +3.14 V, max = +3.46 V)
+5V: +4.97 V (min = +4.73 V, max = +5.24 V)
+12V: +12.16 V (min = +10.82 V, max = +13.19 V)
-12V: -11.95 V (min = -13.18 V, max = -10.88 V)
-5V: -5.20 V (min = -5.25 V, max = -4.75 V) ALARM
V5SB: +4.97 V (min = +4.73 V, max = +5.24 V)
VBat: +1.87 V (min = +2.40 V, max = +3.60 V)
fan1: 0 RPM (min = 2657 RPM, div = 2)
fan2: 0 RPM (min = 2657 RPM, div = 2)
fan3: 0 RPM (min = 2657 RPM, div = 2)
temp1: +28°C (high = +60°C, hyst = +55°C) sensor = thermistor
temp2: +28.5°C (high = +80°C, hyst = +75°C) sensor = thermistor
temp3: +34.0°C (high = +80°C, hyst = +75°C) sensor = thermistor
vid: +1.900 V
alarms:
beep_enable:
Sound alarm disabled

Hmmm, my vbat seems to be low. Maybe time I got a new CMOS battery.

Note also that core2 always reads .5v under what it should. I have to adjust my sensors program to compensate.

Anyway, from what people say, later Abit boards in general are notorious for not being able to keep vtt stable. It should be linked to memory voltage but a lot of people find that it doesn't increase or decrease properly as they change the memory voltage. So it might be nothing to worry about.

hyperspace · Post by **hyperspace** » Wed Jan 05, 2005 11:08 am

purrkur...

Is capacitor EC10 the 100 µF variety?

Billl · Post by **Billl** » Sat Jan 08, 2005 5:57 am

purrkur wrote:2. Have you had to increase the CPU voltage in order to get 550MHz to work?

The 366's in general need a bit of a voltage tweek. It's pretty rare to find a pair that will run at 550 without doing so. I'd say what you have is pretty normal. Your lucky you even got one to run stable at default voltage. I wouldn't run above 2.2 tops, if you need more then that you need a new chip

You would have to have very good cooling at anything above that. But usually they are rock stable with just a small tweek(2.1).

Billl

purrkur · Post by **purrkur** » Sat Jan 08, 2005 4:43 pm

Thanks all of you for your answers. Some feedback:

Hugoc: You mentioned that when you got your sensors working that you were using the standard file that came along with the sensors package. I should have mentioned why I made a BP6 specific file. The first reason being that the original file is very large and very confusing. It is much better to have a streamlined config file with just what you need in there. The second (and largest reason) is that there are failures in the standard sensors.conf when it comes to the BP6. This isn't really unusual. The standard config file doesn't work correctly with my ASUS A7N8X-X motherboard either even though sensors work just fine. The reason for this is simple: The sensor chips are the same but their implemention varies greatly between motherboard makers.

The issues that are wrong for the BP6 are VCore2 and VBat are switched (and your vBat is actually your Vtt IIRC), There is no fan3 tacho output and I have also done other modifications like adjusted fan1-2 output so that they read speed output down to 664rpm (if I remember correctly) and I have also put in sane min/max values so you can really use sensors to alert you to problems with your motherboard. I have done other mods as well that I can't remember on top of my head (Calculation mods if I remember correctly).

Hyperspace: I have actually changed the EC10 cap to a 1000microF 10V job.

Billl: Thanks for the tip. I am going to install some software that can give me detailed info on the chips I got. It took me forever to get the heatsinks on so I am reluctant to move them in order to check the chips!

hugoc · Post by **hugoc** » Sun Jan 09, 2005 9:33 am

OK, I added your custom sensors.conf, and these are my readings now for comparison:

eeprom-i2c-0-52
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 128

eeprom-i2c-0-50
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 256

w83782d-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 5000
VCore 1: +1.89 V (min = +1.81 V, max = +1.98 V)
Vtt: +1.50 V (min = +1.42 V, max = +1.57 V) ALARM
+3.3V: +3.34 V (min = +3.14 V, max = +3.46 V)
+5V: +4.87 V (min = +4.73 V, max = +5.24 V)
+12V: +12.52 V (min = +10.82 V, max = +13.19 V)
-12V: -12.28 V (min = -13.18 V, max = -10.88 V)
-5V: -5.35 V (min = -5.25 V, max = -4.75 V) ALARM
V5SB: +4.92 V (min = +4.73 V, max = +5.24 V)
VCore2: +1.89 V (min = +1.81 V, max = +1.98 V) ALARM
CPU2 fan: 168750 RPM (min = 664 RPM, div = 8)
CPU1 fan: -1 RPM (min = 664 RPM, div = 8)
CPU1 temp: +30°C (high = +60°C, hyst = +55°C) sensor = thermistor
CPU2 temp: +32.0°C (high = +60°C, hyst = +55°C) sensor = thermistor
MB temp: +37.0°C (high = +60°C, hyst = +55°C) sensor = thermistor
vid: +1.900 V
alarms:
beep_enable:
Sound alarm enabled

purrkur · Post by **purrkur** » Sun Jan 09, 2005 12:50 pm

Hugoc: Looks better, although I was surprised to see that you had two alarms out of three that were false. Strange stuff. Did you restart lm_sensors after changing the config file or did you just run "sensors -s"?

hugoc · Post by **hugoc** » Tue Jan 11, 2005 12:25 pm

Yes, I don't know what's with the alarms. My output this morning reads:

eeprom-i2c-0-52
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 128

eeprom-i2c-0-50
Adapter: SMBus PIIX4 adapter at 5000
Memory type: SDR SDRAM DIMM
Memory size (MB): 256

w83782d-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 5000
VCore 1: +1.87 V (min = +1.81 V, max = +1.98 V)
Vtt: +1.49 V (min = +1.42 V, max = +1.57 V)
+3.3V: +3.33 V (min = +3.14 V, max = +3.46 V)
+5V: +4.89 V (min = +4.73 V, max = +5.24 V)
+12V: +12.16 V (min = +10.82 V, max = +13.19 V)
-12V: -11.95 V (min = -13.18 V, max = -10.88 V)
-5V: -5.20 V (min = -5.25 V, max = -4.75 V) ALARM
V5SB: +4.95 V (min = +4.73 V, max = +5.24 V)
VCore2: +1.90 V (min = +1.81 V, max = +1.98 V)
CPU2 fan: 0 RPM (min = 664 RPM, div = 8)
CPU1 fan: 0 RPM (min = 664 RPM, div = 8)
CPU1 temp: +30°C (high = +60°C, hyst = +55°C) sensor = thermistor
CPU2 temp: +29.5°C (high = +60°C, hyst = +55°C) sensor = thermistor
MB temp: +34.5°C (high = +60°C, hyst = +55°C) sensor = thermistor
vid: +1.900 V
alarms:
beep_enable:
Sound alarm enabled

I don't know why the -5V line is a problem. It's within 5% of where it should be. I did restart lm_sensors after I changed the .conf file, but this reading is after it had been turned off all night and then booted up this morning, so it's about as restarted as it'll ever be. GKrellm won't read vtt either, which is strange. The option for that sensor just isn't there.

Dave Rave · Post by **Dave Rave** » Tue Jan 11, 2005 1:00 pm

you say the cpu v is at 2.10 yet your sensors are showing 1.89 ??
if you are upping the speed, without the compensating overclock of the voltage, the cpus prolly won't show

Dave Rave · Post by **Dave Rave** » Tue Jan 11, 2005 1:03 pm

also, the vtt at 1.46 is close to 1.50
the range of the fisrt you showed at
1.42 to 1.57 is not bad but less than good.
leave it running stable and see how much the vtt changes, though if the ec10 has been done it should be fine

davd_bob · Post by **davd_bob** » Thu Jan 13, 2005 5:33 am

purrkur,
My board and yours have a simular cap history.

I replaced 8 Jacksons that looked bad but didn't repalce the other ones.(d'oh) The caps at the zif's were Tahey and looked ok so I didn't replace them. The teltale sign was the system temp was always well higher then the cpus, sometimes up to 10 deg over the cpus before replacing the 4-packs and below the cpus after the repair.

For some reason I stored the board about Aug and built it back into a system about Oct. The current PSU is a 350 no-name brand. Sometimes everything works 100FSB with voltage at 1.9 others I have to set it at 2.1 and drop back to 66FSB to get both cpus to show up. I haven't tracked vtt.

I am short on PSUs atm but Im sure my issue is power starvation/weak caps. I will do another teardown and replace the rest of the Jackson caps and rebuild with a 550watt.

Billl,
The BP6 is a unique mobo. Before I started using my current weak PSU I had good luck running below 2.0 volts on both cpus no-matter if OC'ed.(2 pair of 500s, 1pair of 366@550) Kuun runs a pair of 366@550 below 1.9 volts.

hyperspace · Post by **hyperspace** » Thu Jan 13, 2005 8:31 pm

Here's mine before using purrkur's sensor.conf file.

Code: Select all

[root@multiverse /]# sensors
eeprom-i2c-0-52
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       128

eeprom-i2c-0-51
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       256

eeprom-i2c-0-50
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       256

w83782d-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 5000
VCore 1:   +1.94 V  (min =  +1.86 V, max =  +2.05 V)
VCore 2:   +1.50 V  (min =  +1.86 V, max =  +2.05 V)
+3.3V:     +3.30 V  (min =  +3.14 V, max =  +3.46 V)
+5V:       +5.00 V  (min =  +4.73 V, max =  +5.24 V)
+12V:     +11.86 V  (min = +10.82 V, max = +13.19 V)
-12V:     -12.20 V  (min = -13.18 V, max = -10.88 V)
-5V:       -5.10 V  (min =  -5.25 V, max =  -4.75 V)
V5SB:      +5.11 V  (min =  +4.73 V, max =  +5.24 V)
VBat:      +1.94 V  (min =  +2.40 V, max =  +3.60 V)
fan1:     4687 RPM  (min = 3139 RPM, div = 2)
fan2:     4560 RPM  (min = 6683 RPM, div = 2)
fan3:        0 RPM  (min = 10714 RPM, div = 2)
temp1:       +25°C  (high =   -65°C, hyst =  +107°C)   sensor = thermistor   ALARM
temp2:     +25.5°C  (high =   +80°C, hyst =   +75°C)   sensor = thermistor
temp3:     +28.5°C  (high =   +80°C, hyst =   +75°C)   sensor = thermistor
vid:      +1.950 V
alarms:
beep_enable:
          Sound alarm disabled

And now using purrkur's sensor.conf file.

Code: Select all

[root@multiverse /]# sensors
eeprom-i2c-0-52
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       128

eeprom-i2c-0-51
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       256

eeprom-i2c-0-50
Adapter: SMBus PIIX4 adapter at 5000
Memory type:            SDR SDRAM DIMM
Memory size (MB):       256

w83782d-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 5000
VCore 1:   +1.97 V  (min =  +1.86 V, max =  +2.05 V)
Vtt:       +1.52 V  (min =  +1.42 V, max =  +1.57 V)
+3.3V:     +3.33 V  (min =  +3.14 V, max =  +3.46 V)
+5V:       +5.03 V  (min =  +4.73 V, max =  +5.24 V)
+12V:     +11.92 V  (min = +10.82 V, max = +13.19 V)
-12V:     -12.11 V  (min = -13.18 V, max = -10.88 V)
-5V:       -5.05 V  (min =  -5.25 V, max =  -4.75 V)
V5SB:      +5.13 V  (min =  +4.73 V, max =  +5.24 V)
VCore2:    +1.94 V  (min =  +1.86 V, max =  +2.05 V)
CPU2 fan: 4687 RPM  (min =  664 RPM, div = 8)
CPU1 fan: 4560 RPM  (min =  664 RPM, div = 8)
CPU1 temp:   +25°C  (high =   +60°C, hyst =   +55°C)   sensor = thermistor
CPU2 temp: +25.0°C  (high =   +60°C, hyst =   +55°C)   sensor = thermistor
MB temp:   +28.5°C  (high =   +60°C, hyst =   +55°C)   sensor = thermistor
vid:      +1.950 V
alarms:
beep_enable:
          Sound alarm enabled

Great work, purrkur! Thanx!

purrkur · Post by **purrkur** » Fri Jan 14, 2005 5:12 pm

davd_bob: Thanks for the info. I think I will replace the other caps as well. Just a thought though, do you really think that your BP6 needs more than 350W?

Hyperspace: You are welcome

davd_bob · Post by **davd_bob** » Sat Jan 15, 2005 5:37 am

I doubt it needs more then 350watts because I never run more then 2 hds, but I only have spare 100 and 250s...if I have to buy one I will get at least 500.
I estimate the power consumption as follows.(just a guess)
2x25w 2 CPUs
125w mobo and cards(prob. overestimated)
125w 2 HDs
50w 2-CPUfans, 1-case fan, 1-video fan, 1-chipsetfan.

purrkur · Post by **purrkur** » Sat Jan 15, 2005 4:00 pm

I am thinking about getting one of those wattmeters to be able to see how much my computers (and other stuff) is really pulling.

Wolfram · Post by **Wolfram** » Mon Jan 17, 2005 6:05 am

Some numbers I got from my PCs some time ago:

PC 1:
Shuttle AK32A (KT266A), Duron 1600@2288 with full L2-cache enabled (so running as an Athlon XP Thoroughbred B) @1,85V, Thermalright TR2-M3, 2x512 MB DDR-RAM, Sapphire Radeon 9800 @412/324, Soundblaster Audigy 2 ZS, Hitachi Deskstar 7K250 160 GB, Pioneer DVD 214, 3,5-Zoll Floppy,2x8cm-fnsr (1x@7V, 1x @12V), Enermax 350W, Windows XP Professional SP1. CRT: Iiyama Vision Master 450(19´´)

Windows desktop: 164 W
Windows desktop full load: 233 W
CRT: 89 W
full load incl. display: 324 W

Looks like the overclocked and overvolted Duron is really hungy...

PC 2:
MSI K7D Master-L, 2x Athlon XP (Thoroughbred A + B, at 1,6V) 1800+, 2x256 MB registered ECC-RAM, 2x Arctic Cooling Copper Silent 2 Rev.2, Powercolor Geforce3 @235/258, Arctic Cooling VGA Silencer (low), Matrox G200 PCI, ST Audio DSP24 Value, Seagate Barracuda IV 80GB, Seagate Barracuda 5400.1 (40GB), AOpen CDRW 121032, 3,5-Zoll Floppy, No-Name "LC Power 550W" PSU (Lüfter auf 7V), two 8cm-fans @7V. Windows XP Professional SP1. CRT: Iiyama Vision Master 450(19´´)

Windows desktop: 141 W
Windows desktop full load: 168 W
CRT: 89 W
full load incl. display: 257 W

Windows desktop with 2x Folding@Home: 143W

This PC pulls almost as much power when it's idling as it does under full load. Only Toast makes it draw a little more power.

PC 3:
Abit BP6, 2x Celeron 366@523, 2x 8cm-CPU-fans @7V, 2x 256 SDRAM, Matrox G400, Adaptec 2940 SCSI, Soundblaster AWE64, AVM Fritz! PCI, Lifetec TV-Karte 9415, RTL8139D + 3Com 3C905B NICs, Samsung SV1203N (120GB), LiteOn CDRW 31240, 3,5-Zoll Floppy, No-Name 250W PSU (Lüfter @5V). Windows 2000 SP4. CRT: Highscreen 17’’ aka Hansol 720E.

Windows Desktop: 72 W
Windows desktop full load: 108 W
CRT: 53 W
full load incl. display: 161 W

standby incl. peripherals (scanner Mustek 12000SP, Longshine 5-port-switch, speakers): 25 W

PC 4:
PcChips 758 (SiS620), Celeron 566 (1,5V-version, passive cooling), 1x 256MB SDRAM, Realtek 8139D NIC, Seagate U7 60GB, Samsung 112W PSU with 8cm-fan @7V. Windows XP Professional SP1. No display.

Windows desktop: 31 W
Windows desktop full load: 38 W

Full load was produced in 2D-Mode with the "Toast" utility. "Realworld" applications like 3DMark 2003, Far Cry and UT2004 produced lower loads, even with stressing the video card in 3D mode. I don't know how exactly Toast stresses the CPU, but it seems to do that very well...

purrkur · Post by **purrkur** » Mon Jan 17, 2005 6:40 am

Whoa! That was very interesting!

The issue that surprised me the most was that your dual Athlon isn't pulling more djuice! Tell me something, are you just using regular XP's and not MP's in that motherboard?

I never figured out the difference between XP and MP processors (other than the fact that AMD tests the MP processors in a dual configuration setup).

Wolfram · Post by **Wolfram** » Mon Jan 17, 2005 7:05 am

purrkur wrote: The issue that surprised me the most was that your dual Athlon isn't pulling more djuice! Tell me something, are you just using regular XP's and not MP's in that motherboard?

Regular XPs. Now two T-Bred A's, and this mixed TBred A/B setup before.
I also would have bet the dual Athlon was the power-hungriest PC. But the overlocked and overvolted Duron seemed to draw 90 watts or even more. Possibly the voltage regulators also eat much power. Though the Shuttle AK32A is by no means set up as an overclocking board (missing BIOS options), it was perfectly stable with that Duron and that overclocked Radeon 9800, same with a Geforce 6800 GT

If I had been running PC1 24/7 it would have eaten the average power consumption of a one-person-household here in Germany (~ 2000 kWh), costing more than 300 Euros a year(!). My personal budget is ~2800 kWh a year, but I spend the entire weekend at my girlfriend's place. Guess what is eating up all this wattage

I never figured out the difference between XP and MP processors (other than the fact that AMD tests the MP processors in a dual configuration setup).

Yes, AFAIK the MPs are just pre-selected (and with different L5- etc. bridges of course). I read that sometimes TBreds may not work. But never read why exactly that should happen.

Billl · Post by **Billl** » Thu Jan 20, 2005 8:42 pm

davd_bob wrote:Billl,
The BP6 is a unique mobo. Before I started using my current weak PSU I had good luck running below 2.0 volts on both cpus no-matter if OC'ed.(2 pair of 500s, 1pair of 366@550) Kuun runs a pair of 366@550 below 1.9 volts.

It must be unique I've never had any luck OC'ing them below 2.0. But that was mostly single cpu systems. No system I've ever OC'ed has worked with lower voltages, and most require a tweak up to get stability. I wonder how the BP6 manages it? Anyone have any ideas here?

Billl

BP6.Com

overclocking from 366 to 550MHz

overclocking from 366 to 550MHz

Re: overclocking from 366 to 550MHz