Printable View
I've built turbopll myself. Yes it is kinda hard to do but it is doable.
here is picture of it:
http://www2.arnes.si/~mbabic/BH6/TurboPLL.JPG
The cooling was some high RMP 8cm fan on cooper heatsink. And some -25°C air (it was winter in the time of testing). And that is it. No water no promy or vapochill...
Gotta love that chip. It works 1400MHz @ 2.3V. For 3 years now.
Celery 700@1.9GHz aircooled??
Funny if a 0.18u CPU can hit higher speeds than a 0.13u tualatin celery.. ;)
can you tell some details about that turbo of yours (min, max output MHz etc)?
for instance what are the dipswitches for? and how do you feed ref. clocks to mobo? That device in the piccy looks like its only replacing the crystal.
Yes that's right macci, that device only replaces crystal and does not create clk signal for processor. Because it is very hard to find chips that are capable working at frequencies higher than 50MHz. My turbopll creates frequencies from about 5MHz up to 32MHz with 2kHz steps. So I was able to change real CPU frequency with 120kHz steps. Good enough. DIP switches are here for seting multiplier wich then set output frequency to 2kHz*multiplier. 4 DIP switches are used to form one digit. The first group of four switches are able to be set from 0 to 15, other three groups are only able to go from 0 to 9. So max multi is 15999*. And max frequency is 15999*2000kHz=32MHz.
This TurboPLL is completely my design from components I was able to get here in Slovenia. You can probably see that it is quite messy setup. The multiplier chip I used, I was only able to get it in SMD pckage so i kinda converted it to DIL. I was also having a lot of problems with output signal and with frequency fluctuations but I was able to solve it using coax output cable and very high capacity on RC filter... but this is another story.
And yes 3.2V on .18u celery is kinda stupid but I was testing in steps. First my plan was to try 2.6V but then I saw that this chip just loves high voltages so i proceeded in testing at higher voltages. I was lucky that MOSFets on board were able to do such a high voltages. But the MOSFets were modded (another pair of them was added on board) and the choke was also replaced with some high duty one.
I also have one picture of my mobo...
http://www2.arnes.si/~mbabic/BH6/BH6.JPG
This is picture with no TurboPLL connected but you can probably see the connector placed where crystal should be placed.
Other mods on this motherboard are fixed PCI frequency (to 40MHz)(on picture just under the northbridge) and fixed realtime clock frequency(14.318MHz wich is the same as original crystal frequency) (needed it because my turboPLL replaced the crystal and then the windows clock was wrong... (on picture at side of mobo). You can probably see the additional MOSFets and high duty choke just under it.
And it was air cooled. If you haven't noticed the power consumption at that voltage and frequency was 94W wich is about as much as 2400MHz AthloXP @ 1.9V
I dont have any Sandra unbuffered benchmarks so I can't help you with that.
yeeesh!
Wish my BE6-II still worked so I could do something like that to it.. wonder how much my Celeron 400 would like 3.5v? heh..
Tribesman: How well do you think the BH6 can handle high fsb:s? If im gonna compete this year 200% overclock is what im aiming for. Celeron 266/300(non-A) is what i think the best candidate. Hmm, or i could just unlock multi? :D
Edit: Oh and tribesman, about your vcore, hey i've been there, once you start pumping it up and gain more speed you just dont wanna stop :D Too bad my duron did stop :( up from 30W to 250W (according to the online wattage calculator). I was shocked to see that i even could go past 3V.
BH6 will handle 150MHz without any problem. I did that on two diferent cpus (P2 350, Cel 700 and P3 1G) The highest was 184MHz with celeron, but that is probably because of PSU Vmods. I was running 3.3 rail at 4V. And because chipset and RAM on BH6 is powered directly from 3.3V line, raising that line helps alot.
And you are completely right about pumping up the voltage. I'm sure if my mobo supported 4V CPU voltage I'd try it.
Find the power feed line and "inject" the 5v rail into it and dumb it down to 4v.... heh.
Or better yet let it stay at 5V :cool:Quote:
Originally posted by STEvil
Find the power feed line and "inject" the 5v rail into it and dumb it down to 4v.... heh.
Edit. Anyhow, im thinking of not participating in this years competition, im selling my prometeia and the new prize is a vapochill :rolleyes: Havent got enough time for it either when i come to think of it :(
So your gonna let OC&Burn win this one without any competition
Probably, im also thinking about moving to Gävle so as i said, not alot of time :)Quote:
Originally posted by meh
So your gonna let OC&Burn win this one without any competition
Let's see. 1700+ that does 2466. 1800+ that does 2633. 1800+ that does 2533. 2400+ barton mobile (1800mhz) that does 2600.
Those are the chips I've got right now. All do this fun stuff under standard water. About a 72% overclock is my best there.
Best i've managed was either my 1.2c @ 1.75 2.25v or my 2000+ @ 2600mhz 2.05v... both were just sitting in the bios for a coupe hours, though.. 8-/
wow with only 1.6v? nice!
^^
Teh Celery 300 @ 600 was my finest hour, although it only ran like that for an hour or two, damn PC-Chips motherboard, no voltage settings.
How can i calculate how many % i've overclocked?
im no good in math :P
Start by getting rid of that av, no logical thought possible with her there ;)Quote:
Originally posted by Gef
How can i calculate how many % i've overclocked?
im no good in math :P
Divide your max O/C MHz by Stock MHz.
Multiply by 100, then subtract 100. That's your OC percent.
Ok so...if i had overclocked mine to 3.5GHz then it would be:
3500/2600*100-100=34,61%
am i right?
I'm norwegian, i think i understand that subtract is minus right? :P
Gef - you are correct, 34,61% O/C.
I'll try to watch my English (seriously), don't want to confuse anyone over language ;)