rediscovering q6600

S

sgnx

n00b
Joined
Oct 30, 2006
Messages
19
hey guys, so i've been running my g0 q6600 @ 400mhz x 8=3.2ghz for two years now, on a p35 ds3p, with a zalman 9500 hsf, which btw i don't think is sufficient, and I could never do 400x9=3.6ghz without temp+stability issue... a few days ago, the beloved @bios app killed my p35-ds3p (yes, i know, i'm an idiot, let's not talk about this...); deciding holding onto my q6600 a couple more year before jumping on the iX ship, i ordered an ep45-ud3l motherboard and a tuniq tower 120. basically i was hoping that i'd be able to push my q6600 to 3.6ghz this time, and was wondering if you guys would have some experience/tips/suggestions to share on this? much appreciated!:D

oh btw, ram is 2gb x 4 = 8gb ddr2-800 a-data, vga is 8800gts 512, in case this info is also needed.
 
I did the same thing killed my old ip35pro mobo (not with @bios or whatever that is) bought a Asus p5q pro turbo and pushed my q6600 from 3.2 to 3.6ghz. Everything is running good, and will last me for a good while till ready to upgrade (at least 1-2 years, or until I break something again on a bored day off).

I got a xigmatec S1283 cooler and not sure if its any better then my old ultima 90 though. It is quieter (92mm fan vs 120) though (main reason I bought it).

edit: To add, I needed 1.42v to get 3.6ghz and it doesn't seem to want to go higher then that. My old mobo had bad vdroop causing voltage to be way to high for me to run 3.6 comfortably (1.53v in bios, 1.52 at idle 1.42 in load). Even with 1.5v though wont go 3.7, was hoping I might get a bit more then 3.6 but wasn't able, still happy with my 2 year old q6600 though, haha.
 
Last edited:
If you have not found it already, there are two huge threads on Gigabytes in the Motherboards/Intel forum. Between the two you should find everthing you need to know to get started and plenty of help if you run into a problem.

BUT - Keep in mind the "L' is the light version, the cpu voltage regulator circuity (mainly talking about the MOSFETS) do not have coolers on them and you need to pay close attention to cooling the components around the cpu area for a high OC with a quad.
 
hey thanks, so what do your temps look like?

tissimo said:
I did the same thing killed my old ip35pro mobo (not with @bios or whatever that is) bought a Asus p5q pro turbo and pushed my q6600 from 3.2 to 3.6ghz. Everything is running good, and will last me for a good while till ready to upgrade (at least 1-2 years, or until I break something again on a bored day off).

I got a xigmatec S1283 cooler and not sure if its any better then my old ultima 90 though. It is quieter (92mm fan vs 120) though (main reason I bought it).

edit: To add, I needed 1.42v to get 3.6ghz and it doesn't seem to want to go higher then that. My old mobo had bad vdroop causing voltage to be way to high for me to run 3.6 comfortably (1.53v in bios, 1.52 at idle 1.42 in load). Even with 1.5v though wont go 3.7, was hoping I might get a bit more then 3.6 but wasn't able, still happy with my 2 year old q6600 though, haha.
Click to expand...
 
Thanks, i'll look for the threads you mentioned.
I guess I should probably have order the "R" version, but now I'm not sure if I want to go through the trouble of ordering the R and RMAing the L: not sure it'll be worth it or not (do you think it'll be worth it?). I mean if both can handle 400 fsb quite easily (the impression I got about p45 from reading around) then there's no point since I don't plan on going higher than that anyways.

I guess I'll just see how it turns out when the board+hsf arrives on monday

BillParrish said:
If you have not found it already, there are two huge threads on Gigabytes in the Motherboards/Intel forum. Between the two you should find everthing you need to know to get started and plenty of help if you run into a problem.

BUT - Keep in mind the "L' is the light version, the cpu voltage regulator circuity (mainly talking about the MOSFETS) do not have coolers on them and you need to pay close attention to cooling the components around the cpu area for a high OC with a quad.
Click to expand...
 
BillParrish said:
If you have not found it already, there are two huge threads on Gigabytes in the Motherboards/Intel forum. Between the two you should find everthing you need to know to get started and plenty of help if you run into a problem.

BUT - Keep in mind the "L' is the light version, the cpu voltage regulator circuity (mainly talking about the MOSFETS) do not have coolers on them and you need to pay close attention to cooling the components around the cpu area for a high OC with a quad.
Click to expand...

Curious. I've been running my Q6600 at 3.6 (400x9) for almost 2 years now on my P35-DS3R. There is no cooling on the mosfets or anything else related to the voltage regulation. I haven't run in to any issues with it, but at the same time I could never find any coolers for those parts either.

You think it'll kill my board eventually?
 
No. What he means it that they can get hot depending on how much voltage you are running through your chip. If your CPU doesn't need excessive volts to get to 3.6 then you should be fine.

Watch your MOSFET temps while increasing voltage and stop cranking up the voltage if they start getting too hot and back down a bit. Overclocking does shorten the lifespan of your components, but not to an extreme extent. By the time something does take a dump it would be far outdated. Just don't be a voltage fiend.
 
Forgive me for being so ignorant but it's the first time I heard about MOSFET... so how do I monitor its temp and what's the acceptable range? Thanks!
 
np, its what we are here for.

Trouble is there is no way to monitor the temps with anything but your finger, I guess you could glue a thermal probe on one. They usually are designed to run near and up to120C with 150C rated devices used in the better boards. One would have to look up the data sheet based on the part number printed on the device, if it is readable. Both board/system temp and cpu temp in the bios are not mosfet temp. This is just not with your board, with the possbile exception of one or two very high end boards (no Gigabytes I am aware of) MOSFET temps are not reported and I am not sure I would trust anything that said it was a MOSFET temp without heavy investigation. This is why I warned you.

http://www.newegg.com/Product/Image...-UD3L LGA 775 Intel P45 ATX Intel Motherboard

The CPU voltage regulator circuit is what changes the 12V from the 4pin AUX connector near the edge of the board near the socket to whatever voltage your cpu needs or you set in the bios , about 1.3v or so depending on your CPU. There are 4 main components of the circuit and 3 of them are duplicated several times to provide phases. A phase is a group of those 3 components that are only on for a small chunk of a time window supplying power while the other groups are resting, kinda like the cylinders of a car engine firing in order to provide smooth continuous power but they do not all fire at the same time. This allows more power over a time window without any one cylinder needing to be big or take major strain. Your particular board has 4 phases. Each group or phase on your board consists of:

Around the CPU socket you will see the three main components (the controller chip just looks like a regular IC and carries no power so we are not concerned about it). Capacitors, Inductors, MOSFETS and together they create a DC (12V from power supply) to DC (whatever the Vid or bios calls for) voltage converter/regulator.

7 Capacitors that look like little cans right next to the socket and one off to the side. There are a couple of others in the back corner that do something else. These 8 caps filter the DC voltage to eliminate noise (switching transients) from the MOSFETS. Each phase/group uses 2 capacitors to eliminate the need for one larger capacitor.

4 Inductors, that look like gray square boxes both between the capacitors and behind them. These are "coils" that work to smooth the current flow and act with the capacitors to form a noise filtering circuit an also in conjunction with the capacitors makes a rectification circuit that "smooths out" the chopped up DC from the MOSFETS.

12 MOSFETS, black low rectangular almost square devices, 6 in 2 groups of three behind the caps and in between the inductors and 6 in one long row. These work in 4 groups of 3 and each 3 in a group/phase work together to allow more current flow as opposed to having one large MOSFET. These are semiconductor/transistor devices and if driven hard can overheat.

All of the above can overheat with the increased current demands an OCed CPU requires and the whole area deserves air actively blowing over it. That is one reason the stock Intel Heatsinks point down, the "blow-by" washes air over all these parts and the MCH/Northbridge heatsink nearby. Even warm air is fine compared to no air that most of the big towers coolers that blow out the back fail to provide.

The caps are the new "solid" cap and are less prone to heat failure and increased aging leading to failure than the old style and just due to their nature they do not get greatly heated, its complicated but the caps while liking some air are not the critical failure point they once where.

The inductors are basically just coils of wire embedded in some-kind of plastic/ceramic and while they will get very hot, there is nothing to go wrong unless so much current is sent through them that the wire actually melts inside or wire insulation fails. Unlikely. Again some air is helpful but these are pretty tough customers.

Finally the MOSFETs. These are power transistors and are made with semiconductor junctions inside, a simple IC designed to switch power. They are controlled by another much more complex IC that switches each member of the group of 3 off an on in a precise sequence. Very basically they "chop up" the 12DC from the power supply back into a lower voltage AC that is then filtered and changed into DC by the inductors and capacitors. These devices get very hot and are designed for it under "normal" conditions. Also each phase/group has a sub-group of 3 MOSFETS to share the load. However these fail when they fail, its over. More expensive boards put these under their own little heatsink or heat-pipe. Because of their low profile they do not get air easily if the stock heatsink is replaced. I have seen examples where they got so hot with a quad and your same board that it actually scorched the board brown from the heat.

For long reliable life and stability when OCing getting air to this area and on those components will make the board very very "happy". Even if you hang a fan with string or rubber bands, get some air on those components if you replace the stock heatsink.

The above is of course a very basic generalization and the component placement description is based on your exact board.
 
Last edited:
I'd probably just point an IR thermometer at it. Not sure on acceptable range...though if you don't have 'sinks on it, I'd probably keep it in the 50's?
 
wow thanks a lot bro, i really appreciate all the great info; i'd always thought it'd just be the cpu, mem, and nb, and that's it, lol:) you sure know your way around the processor!!
 
that sure sound more appealing to me than just using my fingers:D

Mav451 said:
I'd probably just point an IR thermometer at it. Not sure on acceptable range...though if you don't have 'sinks on it, I'd probably keep it in the 50's?
Click to expand...
 
They run hot normally, I would not be at all concerned with 100C on a MOSFET but would about there and up start looking to cool it down. If it was an Intel cpu, 100C is very close to the shutdown temp and cause for much much alarm/concern. CPU circuits are small and heat effects them more. MOSFETS are clunky in comparison and designed to handle the power.

If heat where a brick wall, the cpu is a Ferrari, MOSFET is a M1-A2 tank but even some walls are too much for even a tank.

IR thermo is a great idea and a nice tool for the "fully equipped" OCer. Wish I had one.
 
You must log in or register to reply here.
Back
Top