• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Goal: 5Ghz on fx8320

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
but i read it was soldered on and it couldn't be done without ruining the chip. was i given bad info?

The answer is right in front of you. Does it look like it's ruined?
 

Attachments

  • 8320spi1m.JPG
    8320spi1m.JPG
    157.2 KB · Views: 147
lol sorry, i guess that was obvious. what i meant was, is there a special process to do this? will a carefully placed razor blade still work or will i need to heat it to melt the solder?

i'm up for a challenge, and have plenty of experience with solder, but would i need to avoid heating to a certain temperature?

sorry for all of the questions, i wasn't able to find answers in my searches. if you could simply point me to a thread that explains it that would be great!
 
If I'm not mistaken he delidded it just for looks.

Negative. Since Phenom I agena, I've been de-lidding.

Why? Lower core/cpu temp. faster removal of heat.

Why? Get's clocks on K8 over 4ghz. See example de-lidded cpu.

First place US, 11th in the world... used to be like 9th... lol.

http://hwbot.org/submission/2226105_shrimpbrime_cpu_frequency_athlon_64_6400_x2_4063.93_mhz

If you want more info on delidding an AMD chip send shrimpbrine a message ..... he can help



Check out this thread ......... http://www.overclockers.com/forums/showthread.php?t=731806&highlight=lidding

Another thread you guys can check out is my tutorial.
http://forums.extremeoverclocking.com/showthread.php?t=332705

thanks! i sent him a message :D

I'll get you a reply then :D
 
Mr Scott how much did delidding drop your temps? I've read mixed results on delidding and lapping.

Lapping is over rated. Waste of man hours.

De-lidding... Well there are ups and downs.

The ups ^

Cpu and core temps on ong haul loads, will be within 2-5c of each other and in stock or near stock OCs the exact same number.

Over all temp drops are as much as 10c or more at the cores and in turn lowers the Cpu socket temp down by 5c. Remember this is an Average number not exact to each environment.

I've hit clocks with K8 cats could only dream of. That K8 4ghz is soooo close yet so out of reach, I was able to post and run windows at 4.1ghz but no validation.... 4063, was the highest I could Cpu-z. Some day here in the US this will be beaten. But for now, that de-lidded chip and some luck made it all happen. Tried numerous times with the IHS plate on it. No luck.

The Downs :(

Mounting. Tricky it is to gently set a water block down. Straight even and slow. chipping cores is Bad!

Mounting, Cpu water block must be TIGHT. I mean to the point of bending a backing plate or mounting plate. Must be tight. tight is not always good. Had a board no post until I backed the screws off 3/4 turn. posted right up after that.

higher risk to static shock. Believe it or not, the IHS plate covers a lot of important stuff.

Possible Cpu damage and loss of life span from heat IHS removal process.

Although this is a given, Mr. Scotty has been running my exclusive 965BE that was de-lidded literally years ago now. He can be testimony that so far so good is a true saying!

As far as I know, on average, the delid would net roughly 100mhz extra OC.

From my PM box as testimony....

Hey Shrimp,
Just a quick update.... the cpu is running flawlessly, and I wanted to thank you for your help throughout that processor. I recently took the PC apart completely and re-applied coolermaster thermo grease (I was running Arctic Silver 5 prior to that). MY temps went down to lower than I've ever seen them and my overclocks are higher than ever. Keep In mind I am still running air cooling but am able to achieve 4.1ghz overclock and a 3122mhz NB overclock no problem. So all is well that ends well. I'm guessing with watercooling my results might be different but for now my temps are right where they where (approx.) before the de-lid. Just thought you'd like a little feedback on my rid and how the delid has effect it. Hope all is well,

SC

All I can say, is it works. Not always the greatest, but even 50mhz can mean the world to a guy like me :D

Processor de-lid list.

FX-6100 De-Lidded working
FX-8320 De-lidded working
970T De-lidded Working
1090T De-lidded Working
965BE De-lidded working
940BE De-lidded working
9850BE De-lidded working
9950BE De-lidded working
FX-4100 De-lidded NOT Working OOOPS :chair:
6400+ De-lidded working
 
I have quite a few also, across multiple platforms all the way back to PIII's. ;)
De-lidding works, just got to be a little more careful and resourceful on the mounting.
 
Scotty knows about the mounting for sure. You can ask him for advice any time.

@Blaylock,

thanks, it's dated and half way through the entire thread I did come up with a better way to do the de-lid by clamping the PCB in a small vice. Vice grips or pliers to grab and pull the IHS off at just the right moment.

It's actually a quick roughly 60 second procedure I do. Gotta lap the core and ready for running. In total the entire process start to finish is about 45 minutes to an hour.
 
Mr Scott, if you don't mind my asking, how far could you OC your fx8320 before the delid? I'm entertaining the idea of sending shrimp my CPU for a little topless action lol
 
Mr Scott, if you don't mind my asking, how far could you OC your fx8320 before the delid? I'm entertaining the idea of sending shrimp my CPU for a little topless action lol

He couldn't tell you because I had the cpu and de-lidded it before sending it to him.

I can't really say for certain because I did not use this set up as a guide nor did I calculate exactly how much I'd get from this particular one.

But I can safely say that it was close to 200 mhz from 4.8ghz stable to 5ghz+ stable. Although it did pass rigorous testing at 4.9ghz, it wasn't 100% because of high core temps over 62c.
 
i know there are no guarantees in this business, was just curious. i believe considering how far i've gone without the IHS removed (and how i've determined for sure temperatures are what's holding me back) i would be very pleased with the gains from IHS removal.

do you think it would be necessary to lap the bow out of my ek supreme water block?
 
Not in a good position to take some screenshots yet but I've been able to clock to 4.9ghz fairly stable, but not 100% stable. I am pretty sure I got 4.8 totally stable but pretty high temps. For my "daily driver" I'm at 4.75ghz. I think I've got my vcore too high though because it got almost to 60C while doing something once. I think it'll stay stable at this clock at a lower vcore.

I learned a few things that may or may not be specific to my processor I figured I'd share for you anyone who might be reading.

As stated many other places, 55C is definitely the sweet spot with this processor. I found if I go much at all over 55C I'll run into stability problems, so it's best to just aim for below 55. I can *definitely* tell my limiting factor for this processor is the heat. That's a great thing, because there is nothing more annoying than hitting the end of your overclocking ability because your processor just wont go faster no matter how many volts of juice you pump into it. I feel I have lucked out and gotten an exceptional example of an fx-8320 as I'm seeing similar results as people with fx8350's. I'm pretty confident if I could find a way to get 5-10C more off of my temps, I'd easily blow past 5ghz.

I learned that there is a magical sum of vcore+cpu/nb that will get my CPU to 55C under full load conditions. The beauty of this is I can balance between the two to find my optimal overclock by adding some volts to one, and taking away from the other.

For example, in my case, it seems like my magical sum is somewhere around 2.65. I can do 1.35vcore and 1.3cpu/nb for a high FSB overclock (my CPU doesn't seem to need much voltage on the NB for FSB performance). Alternately, I can do 1.5vcore and 1.15 cpu/nb. My CPU doesn't seem to like much under 1.15 for the cpu/nb though, so i avoid going any lower than that. ultimately i get the most gains out of pumping as much vcore as i can and as little cpu/nb voltage as i can because with a faster FSB i found that I need both a higher vcore and a higher cpu/nb because raising the FSB raises both the FSB and the final clock of the CPU. My sweet spot seems to be around 1.15 cpu/nb with a fsb of around 230.

I also found that when I finally got around to adding my backplate to my CPU it improved my temps by at least 5C and made a huge difference in overclockability. If you're in a position like I was with no backplate, definitely prioritize getting one. it will help your motherboard out a great deal by not flexing it, and give you better temps. The stock backplate that comes with the sabertooth rev 1.0 has metal threaded posts that can easily be hammered out and used for a backplate

Finally, make sure you have plenty of air going around the socket and voltage regs. This was another step that made a substantial impact on my stability and overall temperatures.

Anyway, I'll post screenshots and final results once I get my new ram in. Then my build will be complete.
 
^^ this guy here... heh.

You could have used an AM2+ bracket. Would have worked the same.

That's decent Ram. Easily do 1000mhz effective at 1.5v. I can't say from personal experience with G-skill, but from what I see, it's pretty decent. I'm sure they'll clock fine.

And next time before you go monster garage your builds.... come for advice! Heck if needed I could have sent you one of my backplates. Always got a spare one or two laying around.
 
And next time before you go monster garage your builds.... come for advice! Heck if needed I could have sent you one of my backplates. Always got a spare one or two laying around.

Lol! I see nothing wrong with going monster garage on my builds :p I'll keep you posted on how it goes with the ram. Considering lapping the IHS and see if that will do anything. Anyone happen to know the thickness of the IHS? Google suggests 1/8 inch but I'm not sure that's recent
 
I highly doubt it's an eigth of an inch( +3mm). More likely .6 -.7 mm definately less than 1mm. This isn't from first hand experience but ..... Seems about right for forming.
 
Back