There's nothing better then SLK 800 (also unavailable here in Slovenia), so I guess I have to build it myself. I just love a good challenge :D
Printable View
There's nothing better then SLK 800 (also unavailable here in Slovenia), so I guess I have to build it myself. I just love a good challenge :D
Now you just need a SLK-800 for comparison :p
Maybe I'll have to get 1 to you someday ;)
:D very good :D
d'oh.. you should have made a slim fin copper version of the Thermalright AX7 or GlobalWin TAK68 with 80mm fan on top instead of sides.. hehe 8-)
I'm sure the air cooled rig will be interesting also for a comparison.
But what does your O/C stand at now with the latest block revision? Also water and CPU temps?
Is the cold weather holding up for you where you are?
Morphling1 : Going where no one has gone before on water.
:toast: :cool: :slobber: :D
I took I little breather from overclocking, I need to make some blocks now :D In the mean time I'm folding at 2.55 GHz 2V 20°C water (new block) 36°C load internal diode (8k3a corrected).
I will be back, just finish new top for my GF4 water block so I'll put it in tomorrow than some real overclocking begins in w2k and cold air blowing on my heatercore :D :D
morphling1
That a hell of performance for a waterblock. I will be glad to see a test against cathars block. I think yours is beter. Wouldn't your c/w be around 0.10°C/W? Keep up the good work on your blocks =)
BTW this is my first post at xtremesystems =)
Thanks, and Welcome to xtreme.
I reply you PM and I'm tomorrow I'll rerun my diode calibration calculations like I describe to you. With tbredB I would have much more accurate temps with any other board, but if that calculations work then I'll trust my 8k3a again, for now I rely on my stable OC: :)
welcome :D
Well, you could always set the processor up at the same settings in a diff puter and compare temps using the same cooling.. as long as your close it should be withing a degree or so..? 8-)
wow, just noticed this thread ;)
AWSOME OC!!!!!! a 1422mhz chip doing 2700mhz+ incredible!
and i love you waterblock design too...any chance of more pics of it?
Someone drop a vapochill/prommy on the thing already, or some LN2..! 8-)
I couldn't wait until I was finnished, but my cutting tool became dull so I couldn continue but anyway here it is almost finished.
http://www2.arnes.si/~mlivak/Air%20c...0air%20003.jpg
http://www2.arnes.si/~mlivak/Air%20c...0air%20004.jpg
http://www2.arnes.si/~mlivak/Air%20c...0air%20005.jpg
Still need to make shroud and finish cross cutts
I'm just testing that other 1700+ I'm currently at 2250 MHz 1,85V air cooled prime95 stable :D pretty good temps are 33°C with 20°C room in socket diode on my old 8kha+ I will try higher overclocking soon with higher voltage too, but I need to unlock that 13 and up multy because I can't realy get stable past 180 on this mobo.
goood temps.. :D
send one to me, ill test here in brazil with room temps at 40ºC :D
Ouch 40°in the room, I would be melting, I feel hot if my room is 25°C :D
Here it is after one hour of prime 95 temp very good overclock stable. I couldn't take screen shot because I'm not connected with that computer and no floppy either so here the real screen shot :)
http://www2.arnes.si/~mlivak/Air%20cooler/testing.JPG
Lookin good, lookin good! You definitely need a SLK-800 to toy w/...I want to see a morphling1 air-cooler vs SLK-800 :D
:D
damn nice dude, you just made me order a tbred b.
Do you change your fsb in windows ?
I have the same motherboard what program are you using?
Nice oc btw ;)
Have you done any vdd mod on the board? if sow what is it on ?
R
SetFSB is the program to have if you have Epox mobo.
I did vcore mod on the board is that what you asked?
It's pretty simple, I can take pictures later if you want.
ohh
I thought you had a Ep-8RDA+ :) nevermind
Why is the SetFSB so good ? To lazzy to oc do it in windows insted :>
R
Set the voltage, prime, if it fails, drop it back. It is the most simple method of overclocking/testing, and cuts out shutting down/loading the bios/booting to windows.
softFSB is available for most Motherboards that have common chipsets.. KT266/AMD761/KT266A/KT333 etc etc
search these forums for more info
search in google for "EPOX setfsb"
SetFSB homepage:
http://www.page.sannet.ne.jp/k-hazama/software-E.htm
morphing1, one question
i have a shuttle ak31 and an Tbread B1700+.. shuttle is holding me at 133mhz fsb and tbreadb at 12.5x,
Do you think ABIT KX7-333 cant take me to 2ghz ? i´ll use 166FSB and 12.5x multiplier
NoOne,
"might" require a vcore mod dependin on ur chip, but that board will definitely do 166FSB w/o a problem :)
Has anyone figured out how to modify the bridges/pins of TBreds with a default multiplier lower than 13x so that they are unlocked and the upper intervall is available for overclocking?
The solution to this would be to buy an XP2100+ at least, as it has a 13x multiplier. But I don't feel like spending the money when I know I can oc an XP1700+ to the same frequencies.
I have memory that is good for 200+ MHz at really aggressive timings and I'd like to use this performance and 12.5x to achieve 2.5 GHz, this is is just enough to push a good sample XP1700+ Tbred-B to the limit I guess. But I'm borderline .. I'd like to be able to experiment with higher multipliers if possible.
My tbread runs fine at 2Ghz with default vcore :D.. but 200 fsb must rocks :D
Whel, lets spend more money :(
i don't know how good there painting guide is to this specific topic
www.ocinside.de
they have a tbred painting guide. if u know the limit u can paint the dots to get a 13 x or 14 x as wanted.
I don't know how the following problem stand in general, but I just talk to Spaceboy in he got major problem with his KX7 and tbred B, only 180 MHz fsb and random lockups in some apps. No problem with tbred A though, just something to think about.Quote:
Originally posted by NoOne
morphing1, one question
i have a shuttle ak31 and an Tbread B1700+.. shuttle is holding me at 133mhz fsb and tbreadb at 12.5x,
Do you think ABIT KX7-333 cant take me to 2ghz ? i´ll use 166FSB and 12.5x multiplier
That painting guide shoud be good, but for unlocking 13 and up on lower mp tbreds you need to cut the last L3 bridge or even better solder 100 Ohm resistor between AJ27 socket A pin on back of the mobo and ground. That was a tip by Nicozeg I'll try it today or tommorow I need to get soldering iron from a friend.
Ok!
how can i cut one bridge?
morphling1,
I've got a similar toy now ;)
If this thing clocks anywhere near my 2400+ I'll be extremely happy...don't really expect that tho. First testing will air-cooling w/ a Thermalright SLK-800 & a 48cfm fan. I'll report back :)
P.S. If I'm hijackin ur thread, I'll make another...but there are plenty of threads on these Tbred B's now so I figured I'd poke in here, lol
Hey no problem, this thread shoud be for all 1700+ nuts out there :D
Report soon :)
doing 3D @ 2.2GHz already no problem, not bad at all...lets see if it chokes soon. :)
what voltage ? prime95 also stable ?
off topic: where did my folding banner go, I didn' change a thing.
I put it to 1.9v right away, 12.5x166...just clockin w/ SetFSB right now, up to 180FSB so far seems stable...doing 3DMark loops. No Prime yet, I'll get to that yet...I never even Prime'd my 2400+ yet, I just know what it does 3D stable at.
regarding ur sig, change the "www.liquidninjas.com" in the link to "www.liquidninjas.net". There's a post in the folding forum about it...they can't have it on .com anymore & they're beta testing the .net one that's why it looks a little screwy right now. :)
Be back w/ more......like I said, just doin 3D for now. I'll know very soon whether it clocks like my 2400 or not ;)
Thanks for exaplanation about banner.
So how much did your 2400+ go, is that what you have in sig max 3d benchable speed and at what voltage, same 1.9 V ?
I must say, I am totally amazed. 12.5x190(2437MHz) was a piece of cake for this thing. I'm hittin a bit of a wall here @ 2500MHz...takin alot more voltage...unstable so far. I have to go to dinner w/ my wife, then I will try to see if I can do 2467MHz stable @ a decent voltage on air for a 1GHz OC on air. Surely can't wait to get this chip under my Spir@l...or another block, hint hint, hehehe, j/k :)
ya 2400+ in sig is max 3D-stable speed. 2v to do that.
this 1700+ is def. the better OC'r...not by much, but hey, its a $49 CPU :D
I cant wait to see what the 03## 1700+'s do.. if they are going to make any..?
So exactly the same feeling as I have. Its' just incredible, what this chips can do and how quickly multiplier become bottleneck :D
I'm getting the hints, yours will be in the mail tomorrow ;)
I know you made some changes to get the highest performance on tbreds, will you need to do any further changes for it to work with a Barton when they come out?
I wanna see 3GHz! :D:D:D:D:D:D:D:D
Well I think 1GHz OC will be prime stable on air. Was having some stability problems when I realized my 5v rail dipping to 4.7v? I was really baffled by that 1 as I had it set to 5.1v & it barely ever moved from that. Well, I shut down & tweaked it back up to 5.1 & now it just shuts right down upon load. I think it has taken a dump. Darn shame, I haven't killed any hardware in a looooooooong time. ;)
Oh well, off to diagnose :)
Well, my Antec 400W PP-412X is dead as could be. System running fine again w/ a 250W Enlight, lol, but voltage rails dipping too low to get much stability. Now I will hafta order a PS & wait :)
Reckon Antec won't take this back w/ the "void if removed" label torn? ;)
Ouch... Hope you get a new PS soon.. ;)
Vantec 470 anyone???
Antec 400 R.I.P.
I'd guess you're more unhappy with the delay than the Antec's demise.
Not for a while yet.. 8-(
I have to test a T-Bird 850, 1200, and a Duron 800 to see if they are dead first.. they all have slight chips oh the cores.. and im nearly broke.. 8-/
Well after doing the numbers a 1700+@ 2.2V @ 2700 MHz puts out about 186W. Your runnin about 45C with 20C water, so your about 20C off. Still, 45C on die with a 186W load beats the hell outta a White Water Rapids,:D:D:D:D:D:D:D:D:D:D:DQuote:
Originally posted by morphling1
No idea how far with prometeia, but I tell you one thing, my waterblock is a lot more effiient with heat transfer then prometeia evaporator, so If I can cool my water in the range of 0°C, I would say I'm about 10°C off the prometeia maybe even less.
hehe, the CPU have over 200W :flame:Quote:
Originally posted by Penguin4x4
Well after doing the numbers a 1700+@ 2.2V @ 2700 MHz puts out about 186W. Your runnin about 45C with 20C water, so your about 20C off. Still, 45C on die with a 186W load beats the hell outta a White Water Rapids,:D:D:D:D:D:D:D:D:D:D:D
Is the stock voltage 1.5 or 1.65?
1700+ "B" = 1.6V ;)
D'oh!:D
Well, thats
[(2700/1466.666...) x ((2.2/1.6)^2)] x 49.4 = 172W
Unless I'm mistaken about the Maximum Power Dissipation?
--------------------------EDIT-------------------------------------------------
Well, I finally was able to find AMD's white papers, :rolleyes: and its
[(2700/1466.666...) x ((2.2/1.6)^2)] x 59.7 = 208W
And roughly
[X(208 - 30C)] = 50
Which equates to at least a .03 C/W Waterblock!!
:banana: :banana: :banana: :stick: :stick: :stick: :banana: :banana: :banana:
You guys need to drop all those radiate heat calculators or similar and take a look at AMD specs sheets.
Heat output is directly dependable on voltage increase, that is correct but that does not hold any more for frequency increase, heat output just barely increase with the same voltage.
Ok lets do a little calculations, based on AMD specs for tbred B
1700+ 1.6V 1466 MHz max heat output 59.8W
2200+ 1.6V 1800 MHz max heat output 62.8W
so 333 MHz increase in frequency produce 3W of heat. So 1 MHz of frequency add 0.009W
Overclocking to 2700 MHz at 1.6V add another 11.10W + original 59.8W = 70.9W
Overvolting to 2.2V increases heat output linear by 37.5%
So my cpu at 2700 MHz and 2.2 V produce 97.48W and this is tha ball park we're talking about.
Craig, judging by the picture of Barton, it's only a little longer but the same width as tbred, so I won't have to change the design for it
Somehow I'm always the last to know. Thanks for the heads up:)Quote:
Originally posted by morphling1
Heat output is directly dependable on voltage increase, that is correct but that does not hold any more for frequency increase, heat output just barely increase with the same voltage.
That's not exactly right..... if you look at document 25175 from AMD datasheets you'll see that 2000+/1.667GHz TbredB pulls out 61.3W of heat and 2400+/2.000GHz TbredB puls out 65.3W.... now that's 4W of heat, not 3..... but it's not as simple as that....Quote:
Originally posted by morphling1
You guys need to drop all those radiate heat calculators or similar and take a look at AMD specs sheets.
Heat output is directly dependable on voltage increase, that is correct but that does not hold any more for frequency increase, heat output just barely increase with the same voltage.
Ok lets do a little calculations, based on AMD specs for tbred B
1700+ 1.6V 1466 MHz max heat output 59.8W
2200+ 1.6V 1800 MHz max heat output 62.8W
so 333 MHz increase in frequency produce 3W of heat. So 1 MHz of frequency add 0.009W
Those CPUs are NOT exactly equal.... another example - in the samo datasheet there are two types of 2000+ TbredAs, they run with different Vcore - 1.6V vs 1.65V but they are labeled to pull out the exact same amount of heat....
And believe me - one 1467MHz CPU won't "add" just 11W to its power consumption when overclocked 1300MHz more to 2700MHz.... it's not possible...
This is a semi-conductor technology, not maths.....
You're right it's not math, but it it holds much better then linear increase in speed and heat output
skip 2400+
Lets take a look at 2600+
2133 MHz 1,65V 68.3W
Compared to 1800 MHz 2200+ that's 333MHz increase in frequency which like I said adds 3W
So 62.8+3=65.8 now here we also have increase in voltage so 65.8*1.65/1.6 = 67.85W so only 0.5 W off the correct mark
So if I'm of by lets say 5W at 2700MHz 2.2W that's nothing compared to 186W calculation right :)
Hey, live and learn, right?Quote:
Originally posted by morphling1
that's nothing compared to 186W calculation right :)
But of course :toast:
o.o
I think yours and cathars' blocks are doing pretty close.. but mines gonna make them wimper 8-)
Just gotta procure the materials.. stupid hobby shops in alberta.. 8-/
:eek:Quote:
Originally posted by STEvil
o.o
I think yours and cathars' blocks are doing pretty close.. but mines gonna make them wimper 8-)
Just gotta procure the materials.. stupid hobby shops in alberta.. 8-/
Can't wait STEvil :toast:
I should mention that my pump is likely going to have a slight affect on the outcome.. being able to put forward some serious flow rate and psi for under $100 CDN.
I will have stuff up as soon as I possibly can, but my time to work on this stuff is a bit limited right now.. 8-/
I've got a 450W Allied coming. Have 1 runnin my Duallie & its sweet. Adjustable pots & only $46 shipped :)
I don't think I'll buy another Antec PS...that thing died in only about 6-8 months :(
Now I'm getting anxious ;)
Somebody posted a pic explaining the different steppings, but i cann ot seem to find it ? anyway if anyone knows what the manufacturing code for a t-bred 11700 or has that diagram can they please let me know!
ta
A 1700+ Tbred B is AXDA1700DUT3C :)
Ordered 2 of these bad boys from New Egg last night.
Should be here early next week. Hope I see some similiar reesults with em :)
Peace
WannaGoFast
I'm gonna say it just one more:Quote:
Originally posted by morphling1
You're right it's not math, but it it holds much better then linear increase in speed and heat output
skip 2400+
Lets take a look at 2600+
2133 MHz 1,65V 68.3W
Compared to 1800 MHz 2200+ that's 333MHz increase in frequency which like I said adds 3W
So 62.8+3=65.8 now here we also have increase in voltage so 65.8*1.65/1.6 = 67.85W so only 0.5 W off the correct mark
So if I'm of by lets say 5W at 2700MHz 2.2W that's nothing compared to 186W calculation right :)
"one 1467MHz CPU won't "add" just 11W to its power consumption when overclocked 1300MHz more to 2700MHz.... it's not possible..." :rolleyes:
Athlon XP 1700+ 1.47Ghz SoAQuote:
Originally posted by N8
A 1700+ Tbred B is AXDA1700DUT3C :)
Mfr code: AX1700DMT3C
Athlon XP 1700+ 1.47Ghz SoA
Mfr code: AXDA1700DLT3C
Any idea if either of these could be a t-bred?
AX1700DMT3C = Pallie
AXDA1700DLT3C = Tbred A
AXDA1700DUT3C = Tbred B
Mind you AXDA1700DLT3C is what Newegg advertised, yet they shipped a AXDA1700DUT3C :)
Thanks N8 :D whats the diff between and a&b?Quote:
Originally posted by N8
AX1700DMT3C = Pallie
AXDA1700DLT3C = Tbred A
AXDA1700DUT3C = Tbred B
Mind you AXDA1700DLT3C is what Newegg advertised, yet they shipped a AXDA1700DUT3C :)
For me? About 400MHz of OC ;)
The B being the better choice of course :)
geee that is some difference!Quote:
Originally posted by N8
For me? About 400MHz of OC ;)
The B being the better choice of course :)
I can only hope they all clock so well...got 4 more comin :)
what the hell are you doing with all them?:confused:Quote:
Originally posted by N8
I can only hope they all clock so well...got 4 more comin :)
Replacing all the CPU's in my folders :)
COOL
N8
You say you ordered 4 more from Newegg? When did you get the JIUHB 52 , or when did you order it?
Ordered it last week, got it wednesday. Ordered the other 4 as soon as I saw how this OC'd...no idea whether they'll come from the same batch or not :)
Tod 2600+ 2133MHz cpu add just 8.5 W from 1466MHz, and that with running 0.05V higher at 1.6V it would add 8.25W that in 666MHz overclock times 2 that more than my overclock, and it would hypoteticaly add 16.5W you can increase that to 20W if you want but I'm still only 10W off at the most.Quote:
Originally posted by TodB
I'm gonna say it just one more:
"one 1467MHz CPU won't "add" just 11W to its power consumption when overclocked 1300MHz more to 2700MHz.... it's not possible..." :rolleyes:
2700 MHz cpu at 1.6V based on same design would not produce more then 80W.
N8 I'm just scared how your folding numbers will go up with so many machines running 2.3 or more GHz folding non-stop, and not eve mention that you wouldn't get one high end cpu for that money :D :D
If that little cpu isn't a direct gift for overclockers from AMD, I don't know what is.
Voltage doesnt add linear so 186W may very well be correct. Nice oc morphling1, just ordered me a 1700+ JIUGB 0251. Hope I just get that prometeia soon and we might see 3GHz+ :banana:
When me and a friend LN2-cooled a duron 600 to 1521MHz we used two powersupplies in paralell to get sufficient current, one 2x300W server redundant powersupply and one enermax 431W. We were above 3V with it and the cpu radiated 248W. We couldnt get over 2,5Vcore with one powersupply so i think its safe to say that the cpu consumes more wattage than u think morphling ;)
I think got that a little wrong, of course voltage add power linear. And as for total cpu output forget that 186W how do you think that some air coolers would handle that kind of load ? Because 2,2V and 2.5G was seen on air.
But forget that I don't wanna argue about that, I just know that this cpu just don't output more then 100W of heat even at 2.2V 2700 MHz, because my heatercore don't have any more problem cooling the water as it was with old palomino maybe even less. And that tell everything. Do you even have the idea what kind of heat flux would that be if cpu from 85mm2 would outputs 186W, no way in the world my block (or any other) is that good to keep cpu temps ~20°C above ambient with one 1000 l/h aquarium pump.
As for PSU I'm just testing them for our local hw site. 300W Enhance PSU didn't complain one bit when I tested at 2.1V 2500MHz together with 2 cdroms 4x120mm full speed fans (each 5W) fans one 80GB hdd, with 2hours of Tost loading
To bad BillA doesn't visit this forums :) I'm sure he would laugh at those numbers.
Fantastic! Very cool oc! ;)
Nice indeed.
Anybody know how the JIUGB chip clocks?
Cant get my hands on a HB in sweden jet :/
I didnt mean that the duron we overclocked radiated 248W, no way near, but the heatoutput DOESN'T add upp linearly (spellcheck plz?). When u increase the vcore the current also increases, thats why the heatoutput isnt linear against the vcore increase.
Xizten: www.overclockers.se has xp1700+ JIUGB, in case ur interested :)
i know, but i dont want to but it and found out that it sucks :) And i havn't got money for a 8rda+ jet :/Quote:
Originally posted by HMB
I didnt mean that the duron we overclocked radiated 248W, no way near, but the heatoutput DOESN'T add upp linearly (spellcheck plz?). When u increase the vcore the current also increases, thats why the heatoutput isnt linear against the vcore increase.
Xizten: www.overclockers.se has xp1700+ JIUGB, in case ur interested :)
Yes of course the current is in the equation too, but who knows how it changes, it doesn't necesary mean it increases too. I just looked at amd tech docs. and go from there and everything pretty much fits together the way I calculate.
Some computer engineer here, we need help :)
It should be 59.8*(2700/1467)*(2.2/1.6)² = ~208Watts
But that’s impossible because it would mean that an IRC friend would aircool an 117W processor with a old cooler master heatsink. That’s an all aluminium version.
So I'm really wondering how to calculate this.
Edit: morph cant u send a message på BillA? we really need to find out the real way to calculat how much the cpu outputs.
Theoretically the CPU should act as a resistive impedance , so you can calculate the power with the Power = (Voltage) squared divided by Impedance . Of course the impedance vill vary with the temperature , and it's not perfectly linear with the voltage , but it's a good approximation .
The other thing is that the impedance will also vary with the frequency (higher frequency = lower impedance) , but this isn't a linear variation either .
The 248W figure seems pretty belivable to me , the voltage was almost 2 times higher (wich would mean 4 times more power) and the CPU clock was almost 3 times the original one , so 3-4 times more power dissipated would look belivable .
Stole the Allied 450W outta my duallie for now(don't worry folders, replaced it w/ a temp pos PS) & I'm priming right now @ 2437MHz, 1.95v :)
Wife away for the weekend, couldn't stand to let it sit all dern weekend! ;)
Hey guys, I don't upgrade much, but I decided to join the 1700+ t-bred "b" fun. Nothing compared to what the guys here get, but I'm proud of it.
Lookin good! What stepping/week is that 1? :)
& what are your system specs? ;)
Nice.....only 1.7v & crucial PC2100 memory :)
Very impressive :D
That's why I'm so surprised. I need an 8RDA+ with the PCI/AGP fsb lock
Why does it suck? The stepping or? JIUCB is the worst of the J-breds i've heard, and JIUHB the best. But i dont think the difference between JIUGB and JIUHB is very large.Quote:
Originally posted by Xizten
i know, but i dont want to but it and found out that it sucks :) And i havn't got money for a 8rda+ jet :/
I got my 1700+ Jiucb yesterday, up and running today, now @ 1900. My mobo isn't the best one, Msi Kt3-Ultra-Aru, but I'm gonna clock more ;)
It can differ alot between one chip to another, you can get a really good Jiucb and a really bad Jiuhb too, even though Jiuhb's are the best.Quote:
Originally posted by HMB
Why does it suck? The stepping or? JIUCB is the worst of the J-breds i've heard, and JIUHB the best. But i dont think the difference between JIUGB and JIUHB is very large.
Quote:
Originally posted by HMB
Why does it suck? The stepping or? JIUCB is the worst of the J-breds i've heard, and JIUHB the best. But i dont think the difference between JIUGB and JIUHB is very large.
Mine is good, I am on 2,2ghz with the JIUCB stepping but with 1,950Vcore (210x10,5). Im happy with allmost a 800mhz oc
R