Easy Overclocking


Well after much fiddling I seem to have it locked down... man I can not emphasize just how SENSITIVE this hack is to even the tiniest difference in the amount of lead on the resistors.

Well, I got it now where it will lock up within 20 seconds of running a game at 200mhz on DrMD, but seems to play fine at any settings lower than that.

EDIT: I also found that if you even so much as shake the system a bit too hard, what was once enough lead on the resistors to give you 196mhz now wont even support 180mhz :D VERY erratic and unstable results I"m getting.

So... okay all you doomsayers step aside for a second, and the realistic of us ponder this:

If it will run 196mhz stable but not 200, does it make reasonable sense than that the voltage has only increased enough to support 196mhz, which SHOULD mean that the voltage is still within reasonable limits and not high enough to fry anything? We all know that the reason it freezes up at higher clock speeds is because the voltage is not high enough to support those speeds, talking about a stock non penciled unit... so shouldn't the same apply to this mod?


By the way for those asking about how much performance increase, I only monitored it in one game. In Sonic 3, at 164mhz I was getting roughly 31-33fps, now at 180mhz I'm averaging 40-42 fps, so it's average 7-11fps increase in that game. Also Castlevania runs MUCH smoother though i wasn't counting frames on that one. It's a big increase no doubt about that.
 
Mine is still working today fine at 180 mhz I dont think im going to do anything else unless it stops o/c ing to 180 or dies lol. Im quite satisfied with the improvment that 180 brings.

If it dies ill let you know...

Would be nice to have 180mhz versiond of snex9x and opensnez anyone know where I can get them of if they even exist?
 
Redid it again and this time seems to be holding solidly... able to run 180mhz for 30 mins straight, I tried Osnes 200mhz and it just locks up though... so that's the new limit, somewhere between 180 and 200. DrMD soooo much smoother now with that extra 10fps, I want to try Quake and Mame later.

I honestly don't think this is gonna burn anything since the voltage is obviously not so high... if it was so high I think I could do 200mhz... the only problem right now is if it stops doing 180 even again... but I gave the system a vigorous shaking and it's still functioning fine.

If this holds up and doesn't fry the system I'd have to say this is the sweetest thing to happen to the GP community since DrMD v2 or possibly ever.

And by the way, not everyone wants to do the resistors because these contact points are VERY small and you REALLY have to have a steady hand and the proper soldering iron and solder (not just any will do) to pull that off properly. I've chipped a couple of my video game systems and can say from experience that kind of work is a pain in the ass and not too hard to mess up, and in a circuit board like the GP's with a lot of stuff spaced closely together it's not THAT hard to get solder onto other points that you didn't want it to get on and then you have even MORE trouble.
 
This reminds me of the "overclocking the GP could fry the LCD scare stories, and the GBA "If you adjust the volt/contrast POT you'll kill your screen" BS :p

If it breaks, it breaks, if it doesn't, it doesn't.
 
You crazy people!! I wouldn't dare do it, I'm happy with mine running at its current speed although it's not so bad at 184Mhz.
 
Jarska333 posted on Apr 20 2005 at 12:08 PM said:
Could adding a drop of glue on top of the pencil help it... Stick?

LOL :lol:
 
Last edited by a moderator:
noob here

just want to ask again, do I draw on the top of the resistors themselves?
or, do I make a bridge between the 2 pads?


like ( Q = pads, X = resistor, P = pencil )

QQPPP
XX P
XX P
XX P
XX P
XX P
QQPPP

or do I just draw on top of it?
thx
hmmm, can't wait for Quake at 220, might have to keep my gp32 then...

~Octavious
 
well after trying this mod i couldent acheve any higher i couldent even get to 166 , after removeing the coating of pencil which is tricky to apply it was able to do 166 again, im not sure where to go from here , a warnin to ppl the coating is somewhat tricky to remove once on with a standard rubber , i had to cut a rubber into a small shape just to fit it in there

this experment was when bridgeing R52 together only....

Steve
 
I opened my flu and found something extrange...
On top of R48 I can read 300 and on top of R52 640

I thought all GP32 had R48 of 200K and R52 of 470K ... or are there some kind of encrypted codes to assign valors to SMD components?
 
I see that this thread has generated a lot of interest, but I can also see that there are a lot of misconceptions about what this is actually doing and how these CPU actually work.

1) You are decreasing the resistance to the current and therefore increasing the voltage applied to the cpu, obvious to most of us, but why does this work?

2) A clock generator sends a fixed frequency through the CPU, expressed in mhz, 133 mhz, 166mhz, etc. Along this frequency is the binary code of 1's and 0's. 133 mhz is 133 million 1's or 0's per second.

3) The CPU "senses" whether or not this signal is a 1 or 0 by the height of the signals wave or trough, the top of the wave being a 1 and the bottom of the trough being a 0. If you increase the frequency, ie overclocking, say from 133 to 166, then the height of the waves decrease... if the wave height of the electrical signal gets too low, the CPU no longer can "sense" if it was a 1 or 0... the result is a crash. Overclocking frequency alone will not kill a CPU, this is either due to heat or overvoltage.

4) The reason we up the voltage is to increase the wave height back to its original specs. If you overclock from 166 to 184 and it stops working, ie wave height no longer within spec, then uping the voltage to 1.9 or higher may bring it back with spec. and it works fine.

5) That being said, not all silicon is fabricated equal. Different runs have different qualities and so long as it is within tolerances, CPU's get cut and sent out. It is no secret that the big companies like Intel often produce their lines of CPU's from the same silicon wafers. A small percentage are superb quality, and get marketed for the flagship CPU models. 3.4 ghz and such. They continue down the line until the crapy ones get called 1.3 ghz celerons and so forth. (This is just a rough idea of the process, the point made is that there are varying degrees of quality and this quality dictates the overclock and overvolt potential).

6) This then means that just because you can ONLY overclock to 172mhz on your GP, doesn't mean that you are within SAFE voltage tolerance. IF you have crap silicon, it may take 2.1 volts to get to 172mhz. NOT SAFE. Or if you are f@#$@ing lucky, you can get 184+ on the stock unmodded voltage. It's all luck and chance.

7) What needs to be done, REGARDLESS of what overclock you achieve, is checking the voltage after every graphite attempt. Apply graphite and then blow off the excess loose particles, THEN check the voltage to ensure that it does not exceed 2v.

8) Lastly heat dissipated in watts is a function of frequency X volts, and every CPU has its thermal limit. Increase one or both it increases the heat. Heat is caused by resistance inside the CPU, and as the heat rises, so does the resistance (one reason the wave height decreases), causing your game park to run fine for 10 minutes, but as it heats up the resistance increases and the wave height drops, thus 1 and a 0 become unkown and it crashes. If possible, maybe cool your cpu, aluminum or copper. Secondly, overvolting damages the cpu when over tolerance, the little gizmos inside only handle so much potential from ground to positive.



As for the number on the resistors DistWave, they are probably the resistor value expressed in multipliers just like colors are on barrel types.

If you do this mod, blow off the excess dust to help "stabilize" the resistance that you get from effectively running graphite in parrallel with the resistor. Don't leave a lot of loose dust in their to shack around and vary the ohms.

Enough for now, happy modding. If you are careful about the voltge, you shouldn't hurt a thing.
 
guys, dont confuse this with athlon unlocking, its completly different, on an athlon you do it to close teh jumpers to actually enable the "feture" you dont incrase any voltage.
but with this you are changing the voltage going to the µc of the gp.
 
Vimacs posted on Apr 20 2005 at 01:28 PM said:
guys, dont confuse this with athlon unlocking, its completly different, on an athlon you do it to close teh jumpers to actually enable the "feture" you dont incrase any voltage.
but with this you are changing the voltage going to the µc of the gp.


It's very different, when you unlock an Athlon what you are doing is enabling the CPU to run at a different FSB (front side bus). This bus speed times the the clock multiplier is what sets the final CPU speed in mhz. The reason Athlon originally did this has to do with the silicon quality issue. If a 2100+ is of the same quality as say a 2600+, then then you can just clock-up a 2100+ to a 2600+ for no additional cost. Therefore they lock the CPU so that you can't cheat them out of money so to speak. They got wise to the O/C community unlocking them and thus locked the Barton cores tighter. Essentially the ARM9's are unlocked already and you can overclock them via the settings.
 
Last edited by a moderator:
Ok, before apply the mod i've looking for the points where the polymeter should read vcore and I think T7 is GND and T8 vcore... at least my flu gives 1.78v in that points. Anyone knows if this is true?

Thanks.
 
reaper79 posted on Apr 20 2005 at 02:20 PM said:
8) [...] If possible, maybe cool your cpu, aluminum or copper. [...]

How do you do that? This was what my previous question was. I don't know where I heard about the penny thing. So what would the 'proper' way to add a cpu cooler be?
 
Last edited by a moderator:
Tried it today and... my 166mhz blu+ is now 152mhz stable blu+ and the heat was OK even after 3 hours of nonstop playing :D Thanks for suggestion!
 
LOOM posted on Apr 20 2005 at 10:13 PM said:
Tried it today and... my 166mhz blu+ is now 152mhz stable blu+ and the heat was OK even after 3 hours of nonstop playing :D Thanks for suggestion!

Uh? :unsure: Uhm, yay?
 
Last edited by a moderator:
LOOM posted on Apr 20 2005 at 08:13 PM said:
Tried it today and... my 166mhz blu+ is now 152mhz stable blu+ and the heat was OK even after 3 hours of nonstop playing :D Thanks for suggestion!

Well... er... I sure hope that that's a typo.

Anyway, I tried it, and it did nigh on nothing. Gonna try again and use some more tasty graphite as I barely used any last time.
 
Last edited by a moderator:
Back
Top