OC of my athlon 1200C.

[lhgpoobaa]

Results are in for overclocking my system.
today i got the pesky thermal pad removed and two 80mm casefans installed.

note: indicated core voltage is some 0.05V higher than what i set in the bios.

testing done case closed, SETI@home for 100% processor usage. playing mp3's for easy ID of lockup. like playing musical chairs when the music stops its gone!

athlon 1200 C (i.d. unknown, i forgot to look when getting the HS removed, ah heck)
a7v133.
PC150 kingmax ram.
winfop 32-1 cooler.

8.0x150=1200 @ 1.65V (1.70V indicated) - very minorly unstable, locked up with overnite testing. will be useful for daytime hot summer day running.

8.0x150=1200 @ 1.70V (1.75V) - stable, 4 days @ 100%. 45C idle 50C full 28C mobo.

7.5x160=1200 @ 1.70V (1.75V) - blue screen after 20mins. reboots failed to load windows. tried CAS2 & 3, same result. HDD most likely at fault. doesnt like 40mhz PCI.

9.0x150=1350 @ 1.75V (1.80V) - random lockups 1 to 2 hours running time.

9.0x150=1350 @ 1.80V (1.85V) - stable for > 3 hours, no lockup. must test more. 50C idle 55C full 29C mobo.

10.0x150=1500 @ 1.85V (1.90V) - no bootup. tuned off real quick!

9.5x150=1425 @ 1.85V (1.90V) - lockup within 20mins.

9.5x148=1406 @ 1.85V (1.90V) - approx 45min-1hour till lockup. 52C idle 59C load 29C mobo.

10.0x140=1400 @ 1.85V (1.90V) - what its running at while i type this (with frequent saves!), opps. lockup after 1 hour. bummer.

CONCLUSIONS / QUESTIONS.
1. seems that 1400Mhz is my magical limit. looks like i dont have one of those special AXIA's then.
2. has anyone else run their system beyond 1.85V? mines set to 1.85 but it shows up as 1.90V.
3. i think on the whole 9.0x150=1350 is best for me. next ill try to push the FSB up 1 mhz at a time, find the true system bus limit.
4. anyone want to guess what stepping my core is?
cheers. this has been fun.


AMD chips run hot. The world is getting hotter. Therefore, AMD is causing Global Warming!

 

[Bud]

Good post. Thanks for all the data. My 1200 runs pretty much like yours, which tells me I probably don't need to get pc150 ram to see if I can get further.

I'm not in touch with my feeings, and I like it that way!

 

[HighCv2]

Great info man! Thanks!

I am running mine @ 10 * 133

Although what I tried was 12 * 133 and that would not post... bummer LOL

"He who laughs last doesn't get the joke"

 

[lhgpoobaa]

true.
but having pc150 ram is nice for the increase in bandwidth .

just left my PC on overnite. (about 8 hours running at 100%)
9.0x150=1350 @ 1.85V indicated. 53C lovely and stable. mmmm
had to reboot though.
gotta love windows and its "you have to reboot for settings to tak effect" crap.

"i love the smell of Overclocking in the morning!" Says my Hamster.

 

[Guest]

Great info dude ! It would be good to know if it is an AXIA or not for sure.Has any one done any benchmarks on thermal paste and or tape?

 

[lhgpoobaa]

Fact that i cant quite reach 1400 in any form probably means it AINT an axia, and more likely is the earlier AVIA. but thats cool

well i know for a fact that the thermal pad supplied with the winfop32-1 is next to useless!
for example
8.0x150=1200 @ 1.75V = 52-53C idle 60C full & 36 Mobo
introduce 2 generic 80mm casefans, strip the thermal pad for some grease and:
8.0x150=1200 @ 1.75V = 45C idle 50C full 28C Mobo
even at the moment overclocked
9.0x150=1350 @ 1.80V = 50C idle 55C full 28-29C mobo
so even overclocked 150Mhz and 0.05V my temp is still below what it was with the thermal pad.


"i love the smell of Overclocking in the morning!" Says my Hamster.

 

[Guest]

You and that hampster! Think ya need to get out more often.lol

Just new @ everything and dont take me to seriously!

 

[Guest]

cool! that is a big difference!

 

[FatBurger]

I've got an Avia, and I can't even get into windows above 10x133 (1.33gig), even though I have the voltage all the way up. Oh well.

I picked out an Axia for a friend of mine, maybe I'll trade him (he's not interested in overclocking). We both have 1.2/266.

-----------------
Whoever thinks up a good sig for me gets a prize

 

[noko]

My 1.2ghz T-Bird has no problems doing 1.46 at 100% cpu for hours. Here is proof:

<A HREF="http://home.cfl.rr.com/noko/146bench.jpg" target="_new">Kicking 1.6ghz P4, easily</A>

I've operated at 1.52ghz with this setup but it wasn't stable for long periods of time. Check out the cpu voltage, speed, and cpu temp. Not bad for a closed case with only one low flow 80mm case fan, eeh? Playing QuakeIII in a 800x600x32 maxed out, I was also watching the news in which overlay won't be captured here in the image above.

 

[lhgpoobaa]

update here:
slowly increasing the system bus speed beyond 150.
so far it looks stable at
154x9.0=1386, but this requires a voltage jump to 1.85 set. think im really getting to the limit of my PCI cards here.


"i love the smell of Overclocking in the morning!" Says my Hamster.

 

[noko]

What power supply are you using, your limitation of 1.4ghz maybe a weak power supply. 350w plus is really needed when you are upping the voltage and going past 1.4ghz.

 

[lhgpoobaa]

just a good quality 300W one here.
u think it will really make a diff?

"i love the smell of Overclocking in the morning!" Says my Hamster.

 

[Guest]

yes. The psu is sooooo important when overclocking. Run the cpu at default. Fire up via HWM and run a program like seti or something that you know is running your cpu at 100%. Take note of your 5v when your system is idle, and then look at it when your cpu is 100%. Now overclock your system to what your max seems to be and then watch your 5v idle and full load. Using a 1.9vcore I overclocked my 650 to 890. It would boot but if I put a load on the system it would restart randomly. I watched the 5v line when I put the system under load and it would drop to 4.6 and then restart. After I did the 5v mod I was able to have my system run at that speed at 100% for as long as I wanted. It would idle at 5.02v and under load it would drop to 4.92. My cpu is now sucking 2.11 volts, my 5v idles at 4.81 and drops to 4.66. This is even after I moded my mobo to have better 5v range. But it won't restart when it drops that slow because it is a more stable voltage. Let me know what your 5v looks like.

 

[Guest]

noko, just to note. That image you showed proves that you are a fruad, a fake and a lair. The radeon image quality is horrible!

 

[lhgpoobaa]

alrighty...FYI.
running at my "standard" O.C.ed config here...
9.0x150=1350 @ 1.80 set (1.86V displayed on MBM)
50C idle 54-55C load 28-29C mobo

IDLE VOLTAGES:
vcore: 1.86
0: 0.11
+3.3: 3.55
+5: 4.81
+12: 12.34

FULL VOLTAGES:
vcore: 1.86
0: 0.11
+3.3: 3.55
+5: 4.76
12: 12.40

hmms. not much of a fall. 0.05V for the 5V and a gain for the 12V.

must remember that i can run 1350 very stable at 1.80 set. if i wish to go any higher i must immediately move to 1.85 set, and anything beyond around 1390 it just doesnt remain stable.
i think 9.0x154=1386 is my max limit.
and 1.85V is the max i can go on my mobo anyways.
and its not just a FSB being too high lockup. i did try 10x140 and that too died in short order.




"i love the smell of Overclocking in the morning!" Says my Hamster.

 

[noko]

Opps, that was on my MX400, sorry. LOL :lol:

 

[noko]

When you overclock and raise the voltage up you are significantly raising the power the cpu is taking. A 1.4ghz T-Bird at 1.75v takes around 74 watts. When you raise that voltage up even at the same speed the wattage will go up to 100w plus. My 400w LeadMan PowMax LP6100D power supply is rated 30a (pretty high) for the 5v supply which supplies the cpu and other components in the pc. Right now I am operating my 1.2ghz at 1.46ghz with 1.91v to the cpu. I am sure I am exceeding 100w of power to the cpu. So what does this mean:

100w/5v = 20a (20 amps is being fed to my cpu)
This means my 400w power supply has a surplus of 10a on the 5v line which is being used for other things as well. So everything is working well

Now lets take a LeadMan PowMax LP6100C power supply that is rated at 300w, its 5v line is rated for 20a:

100w/5v = 20a (once again 20a is being fed to my cpu)
Cpu alone is taking all that the power supply is rated for on the 5v line so in this case my power supply 5v line is being exceeded in which I doubt I would be able to operate my system at 1.46 and even 1.4 maybe to high of overclock with a 300w power supply.

Note the LeadMan PowMax 6100B, a 250w power supply is rated only 15a for the 5v line. What do you think would happen if I tried this setup - - It wouldn't work, at best it might run my cpu at rated cpu speed and voltage.

So if you are serious about overclocking then a 350w P/S is really starting to push it. A 400w P/S plus is probably your best bet.

Does this answer your question?

 

[Guest]

Tbird 1.4ghz @ 1.75 takes 72watts
tbird 1.46ghz @ 1.91 takes 89.7watts

Looks like your actually taking 17.94 amps away so you have even more head room than you thought. Voltage running through the cpu takes more wattage than a higher mhz cpu (duh) so if you ran the 1.46 @ 2v you'd get 98.1 which is almost 100watts.

 

[noko]

72w is thermal power which is represented by I^2R losses or heat generated by the current flow through the cpu but energy converted to heat due to resistance. Since the cpu is not doing any kind of work that is pretty much the electrical energy being used as well but is probably a little bit higher due to resistance of the traces, power transistors, inductors etc. from the power supply. So you can actually calculate a resistance since you know the electrical power in a round about way. Earlier phsstpok found the spec for the AGP slot which has a 2a rating for 5v for the DDR ram on the card. Interesting that overclocking the CPU can also affect the power to the video card memory causing a possible problem there if the voltage drops to low. Not sure what else the 5v supplies besides the 5v line on USB. Also realize if the voltage drops on the 5v line current flow goes up, 89.7w/4.7v = 19.1amps now I was operating at 1.52ghz with a voltage of 1.94, I am sure I was exceeding 100w there which would be 20a plus 2a on AGP (that would be max) so in short a 300w P/S is not sufficient for very fast T-Birds. Now do you know what else the 5v line supplies. <P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/15/01 02:55 AM.</EM></FONT></P>

 

[noko]

Thanks for the clarification.
Also on second thought a underated 300w P/S may be able to hack the power requirements. Plus there will be another way to calculate the electrical power the cpu is using and that would be using simple heat transfer equations using the delta T (temp) of the temperture across the heat sink. The delta T change would represent a given heat transfer rate and a change of that delta T would give you the change of thermal power and thus electrical energy cpu is using. I am sure the cpu is more complex for a energy system then a simple DC circuit since you do have capacitance and inductive reactance with the resistance so impedence maybe needed to be calculated. Someone more with a electrical engineering back ground can enlighten us or me at least on this.
<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 07/15/01 03:05 AM.</EM></FONT></P>

 

[Guest]

My attention has been brought to this app called radiate which has all the calculations in it to figure out what the watt of a cpu is (at full load) and how well your heatsink is performing. I tried a few variations of making a vortex today but I couldn't get it to work well enough, just too much cfm being pushed into a narrow passage.

 

[Guest]

oh yeah, here is a link
http://www.overclockers.com/articles430/Radiate_Installer.zip

 

[noko]

It is a cool program, except it only goes up to 1200mhz. At 1200 mhz when I plug in 1.85v it tells me the cpu is at 72w, at 1.94v it says 79.1w, but I am running at 1.46gh at the moment and the program doesn't go up that high. Still I will be able to extract data to manipulate. At 1.94v it is roughly 6.5w per 100mhz increase in cpu speed. So from 1200mhz to 1460mhz, would be about 17w more, 79.1 + 17 = 96.1w. Hey that looks close to your number. 96.1w/4.70v = 20.44a from my power supply, so if I jack my cpu voltage up to 1.94v at 1.46ghz I would be exceeding 20a in which a 300w power supply wouldn't be rated for in addition the agp can consume additional 2a. Highly unlikely you would have a stable system at 1.46ghz and a 300w P/S. Cool.

 

[Guest]

make sure you select Athlon (thunderbird), that way the max is 1700mhz. Also, the wattage changes from the different type of cpu's. Some 300watt (like mine) are rated for 30a also.