Just overclocked my Pismo

[coius]

After about 15 minutes of prep/soldering, I bumped up my Pismo to 500Mhz from 400Mhz.

so the setup is:

PowerBook G3 500Mhz/1MB L2

100Mhz FSB

768MB RAM (256MB lower slot, 512MB upper)

80GB 7200RPM HDD, DVD/CD-RW Slot-load optical, brand new 6500Mah battery, original G3 Pismo powerbook power adapter (black brick). 14" LCD. ATI Rage 128M 8MB AGP 2x Graphics card

Airport, 10/100 Fast Ethernet, 2x FW400 ports, 2x USB 1.1 ports, USB 2.0 PC Card, Belkin 802.11b/g PC Card (in case I need G access). Sound in/out, 56Kbps Modem

It's actually quite a bit quicker than the 400Mhz. Windows are slightly more responsive under OS X, typing keeps up now in web browsers, and it loads pages quicker. Not bad...

 

[theos911]

Can you share how you did this?

 

[coius]

On the backside of the daughterboard there are a series of jumpers (under the SoDIMM slot) that have the jumpers labeled with numbers.

it's Jumper 127 that needs to de-soldered and moved to the left (to 130) that will up it. It's a series of three pad

like

--x

x--

x--

x--

--x

x--

above is not a valid jumper position, but you will see three pads, and whether the resisters are from the middle to the right, or middle to the left determines which setting it is

if you see the jumper 127/130 pins, you desolder the jumper going from the middle across to the right pad, then move it over to the left going to middle (and the right pad has nothing) it will set the CPU to 500Mhz

The resister can be put in any polarity (i confirmed this with my dad) so if you get it turned around, it will still work.

Another option to do that is to remove the resister altogether, and use a lead pencil to jumper (supposedly it puts the right resistance on the laptop)

The heat hasn't jumped up on the laptop, I have had no freezing, but keep in mind it will use slightly more power, and the results may not be the same for everyone. Some people have reported that their machine won't start unless the 400 is at 400 or 450, but not at 500Mhz. Mine started right up at 500Mhz.

It is very hard to work with this and you need a really itty-bitty tip or you could damage it. I suppose if you sent me your daughter board, I can do it for you for $15 with return shipping. I would gladly do it for you. I would also test it in my machine to make sure it will work. PM me if you want this, and I can do it. I have my own weller temp-controlled soldering station that I do this with, and i have a replaceable tip that goes down to a fine point.

 

[theos911]

Does anyone know if this is possible on a Wallstreet? How does the Wallstreet's cooling compare to the Pismo's for this purpose?

I will check my Wallstreet for those jumpers when I get a chance.

 

[Byrd]

That's great coius, of the 3 - 4 Pismo machines I've come across, I could never get them to go anywhere - apart from the 400's going to 450. 400-->500 and stock 500 --> 550 resulted in dumping to open firmware and/or random chiming.

 

[CharlesT]

I realize this is an old thread, but in case anyone else finds themselves tempted to try this I'd like to report at least partial success on my two machines. I've got two 400MHz pismos that I received for free (as dead) a couple of years ago, they both came alive after hard drive replacements and new PRAM batteries. I've since gotten better setup with soldering equipment so figured I'd give micro soldering a go. The first one was a little messy, started with the R127 move and it ran at 500 but was pretty unstable. So I backed that one down to 450 and it's been running fine since. The second one went more smoothly. It runs great at 500MHz.

The hardest part is keeping track of the resistor after removal, that sucker is tiny! The resistors are 10K Ohms, by the way, just standard pull-ups.

An interesting observation, the one that would only go to 450, when I had it at 500 it got noticeably warm very quickly, temp is normal at 450. The one that went to 500 with no problem also doesn't seem much warmer than before the change. The bump up to 500 is definitely noticeable running 9.2.2.  At 450, not so much.  But it's worth a shot if you're so inclined. 

 

[unity]

There is no need to track the resistors, they are 0 ohm, meaning no resistance. A solder bridge works just fine. I have overclocked a few in the past, even overclocked one with a G4 upgrade.

 

[CharlesT]

Well, I can confirm that since I lost two of the resistors before I got my ducks in a row. Replaced those with direct shorts. But the resistors do in fact measure 9.9K, at least the ones on my boards did.

 

[Innes]

Super interesting. Does this mimic the setup for the retail 500mhz Pismos or is that a different chip?

 

[CharlesT]

Well, it mimics it in the sense that it performs just like the retail version. As you've probably heard before, many of these chips come from the same batches and are then tested to see how they perform. Those that work at a higher speed are binned as higher speed parts, those that don't go into the lower speed bins. Having said that, in fact there may be more demand for the slower parts so in some cases the higher speed parts are sold as the lower grade. In other cases there may be a considerable design margin on some components so they can perform at higher speeds than they're actually sold for. Finally, a higher speed part must meet it's performance specs at all temperature ranges to be sold as such. So if a 500MHz part fails at very low or very high temperatures it can't be sold as such even though it works fine at more typical temperatures.

So, having said all that, it's basically luck of the draw. Some work, some don't. If I were to stress test my 500MHz overclock under worst case conditions it may well start to fail. But, given these are old, obsolete machines in any case that most of us just use to play with, it's no harm or foul. Just having fun. It is kind of neat to move one tiny part and end up with a machine Apple sold for an additional $1000 back in the day.

 

[mrpippy]

The 500 MHz Pismo also has a much better metal heatsink for the CPU, the 400 MHz heatsink is partially black plastic and just doesn't seem like it would dissipate heat very well

500 MHz:

400 MHz:

 

[galgot]

I have these all metallic heatsink on some 400MHz as well. In fact these are the early Pismo heatsink model, the other part black plastic (well, special kind of)  was the later model introduced on the Pismo models revision in sept 2000 when the HD passed from 6GB to 10GB for the 400MHz and 12GB to 20GB for the 500. 

Does anyone know if this is possible on a Wallstreet? How does the Wallstreet's cooling compare to the Pismo's for this purpose?

I will check my Wallstreet for those jumpers when I get a chance.

I know it’s a 2011 question but yes it seems possible to overclock a Wally.
Found this page (in French) explaining how .

http://www.macbook-fr.com/powerbook/bricolage/overclock_pb_g3_article62.html

Never tried myself. Don’t think I would , Wallstreets are much more fragile than Pismos.

For the Pismos, I remember a trick for overclocking using a 8B graphite pencil to make the connections. Never tied either… But I understand it worked pretty well, and would stay in time as after 2 or 3 month of use the heat will melt the graphite made connections.
here is the original page on the WaybackMachine, in French again sorry and with gone images :

https://web.archive.org/web/20020922213059/http://www.mac-lover.com:80/overclock/overclock.html

but this one as the images, with shorter text :

https://www.mac4ever.com/dossiers/74027_overcloker-un-power-book

 

[CharlesT]

Both of mine have the all metal heat sink so they must be the earlier models. I have to say, the Pismo is by FAR the easiest to work on of any laptop I've ever played with, what a nice design. No wonder folks like them so much. Compare these to a Clamshell or 12" G4, ugh, shudder...

 

[Trash80toHP_Mini]

I've got what I call my PDQuicker, apparently modified by its very sharp original owner to match the Lombard's clock. I still haven't checked it out.

Have there been any results posted on clocking the entire bus on these models?

 

[Byrd]

I think the 400 and 500mhz models use the exact same heat sink, it’s the Wallstreets that use a little black plastic shroud.

Wallstreets overclock to the next speed bump (eg. 233 to 266), I have a 233@300 which has been fine.

 

[galgot]

Indeed the Wallstreets/PDQs have a big metal "heat shield", but not the same.
https://www.ifixit.com/Guide/PowerBook+G3+Wallstreet+Heat+Shield+Replacement/2
It’s really just a metallic plate, not like the thick metal or special plastic real heatsinks of the Pismo. these serve their purpose very well, very rarely I hear Pismo fans running. Whereas the Wallstreet gets hot very easily.

Edit: The Wallstreets have a very small heatsink, that is the black rounded metal plate plugged atop the proc on the daughter card.

That gets itself in contact with the big heat shield plate placed atop.

It's a bit the same system on the Lombard. I suspect they changed that on the Pismo to a bigger and more effective heatsink due to increase in MHz and poor heat dissipation with these small rounded thing...