5300 Overclock

Success!
I've managed to overclock a 5300cs to 5300ce speeds! That is, I changed the bus multiplier to achieve a 117MHz CPU speed from a standard 100MHz 5300cs. As far as I know, this is the ONLY website that documents this sort of modification to the PowerBook 5300 series.

One thing I won't mention on this page is where to get tools or parts. For the record, I don't recommend any modifications you might perform to your hardware. I take no responsibility for any damange you do to yourself or your hardware. And, frankly, if you don't know where to get parts, or what parts to use, then you shouldn't be doing this mod anyway. The learning curve can be pretty steep too, so PLEASE, if this is your first mod and you are doing it to your only PowerBook, RETHINK IT!
There are 2 ways to achieve a speed up of your stock 5300 Series machine. The first is probably the route you should go if you have little experience with board-level, surface mount soldering. Not that the components are much bigger, but there is a little more room for your soldering iron when dealing w/ the 33.3333MHz oscillator.1) The easiest way by far to speed up the 5300 series is to replace the 33.3333MHz oscillator on the mobo with a faster part. To figure out the speed of your new CPU, multiply the frequency of the new part by 3 (ie 36MHz * 3 = 108MHz - if you install a 36MHz oscillator, you'll get a 108MHz CPU). If you have a 5300ce, multiply by 3.5 (36MHz * 3.5 = 126MHz).

     We quickly reach a limit - chips can only run so far over their rated clock speed (even under the best of circumstances, it's hardly likely anything more than a 20 - 25% boost will be stable). So the next problem is to figure out how to change that multiplication factor from 3 (or 3.5 in the case of a 'CE) to, well, anything else (a lower multiplier will allow us a faster bus while keeping the CPU in a reasonable range)....

2) Doing this encompasses the second method for overclocking a PowerBook 5300 Series machine. The multiplier on PowerBook 5300, like many other Macs (see Marc Schrier's wonderful clock chipping website), is set by a series of resistors on the motherboard. These resistors are not documented by Apple, and it often takes some doing to figure it out. To my knowledge, no one has yet posted information about changing the multiplier on a 5300... So here goes! This is some preliminary info I have. Having a CS and CE mobo side by side, I've noticed these differences. Following a hunch, I took a look at these resistors...

Resistor Position:	R43	R46	R47	R48	R50	R127	R128	R129
100MHz (3x)	X	-	X	X	-	-	-	X
117MHz (3.5x)	-	X	X	-	-	X	-	X
In this case, an "X" indicates that a resistor _is_ installed at that location. That would mean that the resistors on Mad Dog's site match up w/ the 5300 series in a fashion something like this: (R9,R2)=(R50,R128) (R6,R5)=(R47,R129) (R8,R7)=(R43,R48) (R4,R3)=(R46,R127)

I downloaded some technical info on the PPC 603e from Motorola's website. Page 14 of that document shows a diagram w/ the pinout of the 603e. It goes something like this:
PPC603e pin 213: PLL_CFG0
pin 212: SYSCLCK (bus frequency)
211: PLL_CFG1
210: PLL_CFG2
209: AVDD (this is connected to pin 4 of the oscillator - Vdd)
208: PLL_CFG3
Page 24 (or section 1.8.1) outlines how to set the bus multiplier. The chart there seems to imply that PLL_CFG[0-2] set the bus multiplier... PLL_CFG3 sets the core to VCO multiplier. Would that be noticeable when you clock chip? -- If not, that might explain why "no effect" is observed when certain resistors were moved about in the 1400 chipping...
Hmmm...

In any case, I kinda skipped this step. I traced out PLL_CFG2, saw it connected to the resistors I was concerned with, and jumped the gun. I was pretty confident that I had the correct settings for clocking up from 100MHz to 117MHz. In the future, I plan on fully tracing out the paths to assemble a table for the 2x, 2.5x, and 4x multiplier settings.

In the meantime, though, I managed to change the multiplier on my 5300cs by moving R48 to R46 and R43 to R127. Due to problems associated w/ using a CE display w/ 512K of VRAM, however, I haven't had much of an opportunity to run diagnostics on the upclocked board. But at first glance, 116.7MHz seems stable and Newer's Gauge Pro's memory test ran flawlessly for several minutes. When I learn more, I'll post more.
Update 9/11/02: I have run a few benchmark tests in MacBench 5.0 with the 5300ce and the overclocked 5300cs (called 5300cse). The results are just what one would expect: the systems perform about equally as well as the 1400/117 in processor tests. In the disk tests, the 5300s lagged a bit behind. suspect this is because the stock IBM 1.1GB disk shipped by Apple in the 5300 only runs at 4,000 RPM - If I recall correctly, the 1400 has a faster stock disk.

MacBench 5.0 benchmark results (5300cse refers to a 5300cs overclocked to 117MHz w/ a 'ce display)

Thanks to Dave Ip, I have a handful of 5300c mobo pics for the appropriate area. Click on the thumbnails below to view an image compressed for use on the web. Click on download original to view the image in it's full, unadulterated 1024x768 glory.

P.S.: (My lawyer made me say this) These modifications are not recommended. I can take no responsibility for any bone-headed stunts you try to pull with your own hardware! ;-) Modifying your Mac in this way will absolutely void any warranty you might have left on that ancient machine. And don't try doing this if you aren't comfy with the guts of your PowerBook and don't know your way around a soldering iron... ...er, not that I really do...

	Clock Generator ~56KB A closeup of the 33.3333MHz oscillator that sets the bus. Replace this with something faster to upclock your 5300. This is the easiest way to speed up the stock machine... [ Download original ~100KB]
	Resistors close up ~76KB I'm pretty sure these are the resistors responsible for setting the PLL for the 603e. I circled them for you to make them easy to spot. (Yes, I know I spelled clcok wrong... I'm too lazy to fix it) [ Download original ~140KB]
	Overview/Perspective ~64KB This pic gives you an idea of the location of these components with respect to the CPU. [ Download original ~168KB]
	Resistors, again ~64KB Another image of the resistors, without the hiderance of my text additions to the picture. This might help you get a clearer, more unobstructed view. [ Download original ~144KB]
	Resistors, once more ~68KB A slightly different view of the area of interest. [ Download original ~164KB]