PowerBook G3 PRAM Battery—i moved it

[AEChadwick]

i finally had time to turn three powerbooks worth of parts into one, solid PowerBook G3 Series (PDQ).

As part of the rebuild, i decided to move the PRAM battery—similar to my PowerBook 1400 project—to save a complete teardown in the future.

Like the 1400, this uses the existing battery pack’s wires, and reroutes the battery into the hard drive bay. There’s much less empty space inside the PowerBook G3. I decided to use some scraps of Pyralux to make my own ribbon cable. It was still a tight fit, just be careful getting the CD drive in and out.

i made custom battery boards to make future replacement effortless. I used Fritzing because it’s easy for small projects. The battery clip is a Keystone 3005. *this same board should work in my Pismo (with less batteries), but i haven't had time to rebuild that... i think it needs a new logic board...

I modeled the existing hard drive bracket so i could design a replacement to fill the space. The original is included here; please remix & share. My replacement attaches the PRAM battery underneath with an SSD upside down on top. The SSD is just an mSATA-to-IDE adapter, removed from its case to minimize profile. I would have liked to get the battery on top, but the original G3 Hard Drive is mounted upside down. My little bracket removes as a single piece, and is still easier than taking the computer completely apart.

I hope this inspires people.

 

[Romko23]

Sorry to ask this in this thread, but where can I get a Pismo G3 good working pram battery ? Its a pia to keep unhooking the optical drive to get it running. My Pismo works great, just want to get a new Pram battery for it.

 

[AEChadwick]

i’m pretty sure the only option is “build your own” (thus this project) but i happy for someone to prove me wrong.

 

[ried]

You can find NOS on eBay from time to time. Here's one I bought in April. Part #922-3829:

 

[kkritsilas]

This may be a dumb question, but what voltages are required from the PRAM battery on the Wallstreet/PDQ and the Lombard/Pismo? Are the 2330 coin cells putting out 3.2V each, or only 1.6-2.0V each? Or is is a current draw issue? Seems to me that a smaller lithium battery could be used, perhaps in a heat shrink tube that would take up far less space. and not require the relocation of the PRAM battery.

 

[AEChadwick]

This may be a dumb question, but what voltages are required from the PRAM battery on the Wallstreet/PDQ and the Lombard/Pismo?

this is an excellent question, and i have no immediate answer. I will lay a multimeter on the wires as soon as i can sit still. In the meantime, here is a schematic of the PDQ PRAM battery.

 

[Cedsrepairs]

It seems to want +3V and -3V and their awful solution was with six 50mAh units
Indeed they seem to want 150mAh (per side, 300mAh total) which is an awful lot to me.

In the nineties, most cells were 3Volts. Which makes replacing this wiht modern units really complicated.
I think it is feasible to replace all of this with two VL3032 (3V 100mAh), but before posting a howto i'll do it on my own PDQ and confirm the method. We're losing 50mAh but I really doubt this power is used anywhere (I might be wrong).

Nowadays, most rechargeable cells a 3.2V (sometimes 3.4V), swapping with those could lead to unpredictable results.
You can't swap with coin cells (non rechargeable) without adding diodes, and even so, this could lead to even more unpredictable results

 

[kkritsilas]

I think 2 CR3032s or a pair of CR123s would probably work. From the schematic above, it seems to want positive on the red wire, negative on the Black, and a voltage half way between the two on the white wire(virtual ground). So using a pair of CR2032s/CR123s (assuming they are 3.2V each) in series, the red comes off the plus side of CR2032 #1, the white is on the connection between the two CR2032s, and the black is on the negative of CR2032 #2. Alternatively, a 6V lithium (2CR5, 6V @1500mAH for example) cell/battery can be used, with a pair of resistors making up a voltage divider network to create the "virtual ground". I still don't see the need for 4 or 6 coin cell batteries. I suppose a set of 4 AA or AAA batteries in a holder could be could used, but now we are defeating the purpose of making the PRAM battery pack smaller.

 

[Cedsrepairs]

I'll work on that when I receive my 512MB of ram, so that I do the two at once

I always look with a lot of "modesty" at what engineers designed, especially back in the days. I'm sure there is a reason why they went for what seems to be a very complicated solution, providing way too much power for what it's supposed to do. So clearly there is something in there in need of >100mA peak. What can it be, I have no idea. Probably part of the sleep process.

The question is, does that reason still exist in the context of retro computing.
It's likely nobody will use sleep anymore on that poor thing (I certainly won't).
However the PRAM is clearly key to have that thing boot reliably. It's also a pain the date vanishes between boots. So these are my modern goals.
I think (and I'm an electronics engineer) that two VL3032 should indeed provide less power in the edge cases, that I have yet to identify, work just as good in the everyday use, and more than that probably charge faster that the original pack (which is great in our retro context). They'll live 20 years without leaking and they are rechargeable. I'll also (as the OP) relocate the battery as the stock location is an absolute pain.

I'll test all that shortly

 

[kkritsilas]

I made an assumption that the original PRAM batteries weren't rechargeable, so if they are, then add a diode (a low voltage drop type, like a schottky, or two) to the stuff that I posted further above.

I think the 2CR5 would be the best way to go, in terms of space usage and practicality. It certainly can provide the power, being 1500 mAH rated, and using the resistor network and the two diodes, should make for a pretty compact solution. 1 X 2CR5 bettery, two resistors of whatever value are deemed appropriate, and a schottky diode to prevent the computer from trying to charge the 2CR5, and good to go. Use the same wiring harness that came in the original PRAM battery pack (mostly so that the connector on the end of the harness mates up properly with the connector on the motherboard, an you should be good to go.

 

[Cedsrepairs]

The pram batteries are indeed rechargeable.
One of the challenges especially on the wallstreet is that there is like zero available space anywhere.
My solution clearly goes along with replacing the hard disk, which frees up around 1/2 of a 2,5" in height, enough to put either two vl3032 (but not much more) or something similar to the original battery ( waiting for the test board of @AEChadwick by mail )

I'm not considering using the original battery location as it requires disassembling 90% of the laptop on a wallstreet, which could do more harm than good .

 

[AEChadwick]

( waiting for the test board of @AEChadwick by mail )

it's on the way, let me know if it works. (I think it's flawed... but it's been to cold in my workshop to experiment)

 

[Cedsrepairs]

Well as I said I should be getting my VL3032 today or tomorrow, so i'll first try that, but that'll be a temporary solution ,
your solution is clearly more elegant (and allows the batteries to be easily replaced) and this can also be proposed to various people (if ti works) whereas you can't objectively recommend people to replace a 6x50mAh system with a 2x100 like I'm gonna do I do it because I know what I'm doing, accept the risks, and have half an electronics lab, but I won't say to everyone with a dead pram to do this.
Your board could be generic solution.

 

[alectrona6400]

Hate to necro a thread but this is INCREDIBLE! Dead PRAMs cause some stupid problems with these. Glad to see someone has thought of this; any updates on the project and where I could find something like this? -K

 

[AEChadwick]

it's incredible in idea, but i’m not 100% sure it works. Turns out the circuitry is weird, like most PRAM batteries only charge if there is a functional battery charging in the PowerBook--no charge to the main battery means no trickle-charge to the PRAM. I've tried this PRAM replacement board in a few machines (i also made a 500-series/1400 version), and it works for a minute then doesn't... i think because the poor little VL3032 discharges and isn't getting any juice.

so until there's a viable battery replacement option for this old PowerBooks... this project is on hold.

 

[Cedsrepairs]

Exactly . It seems the pram is tightly linked to the main battery ; I would say it even charges thru the main battery .
My solution hit the same problems .

 

[3lectr1cPPC]

The wallstreet/pdq batteries I believe aren't too hostile to battery rebuilds like some of them are - they use weird 17670 Lithium cells, but I think they'll be fine if you get some new cells and spot weld them in. There's a battery reset app for these.
I may build a new PRAM battery for mine soon now that I have a spot welder - my battery still works for multiple hours, these old 17670 cell batteries do seem to have great longevity.

 

[dankcomputing]

Could you post a Gerber version of the Wallstreet PCB file? I've been waiting years to be able to rebuild these batteries and this is just the thing to make it as painless as possible.

 

[MikeatOSX]

Thanks to AEChadwick for this great thread!

Some years ago I've written a manual about PowerBook Pram batteries starting with
PowerBook 145B, ending with PowerBook G3 Firewire (Pismo).
Sorry, it's written in german, but you can translate it with Google. Many pictures included.

PowerBook_PRAM_Batterien_v02

MediaFire is a simple to use free service that lets you put all your photos, documents, music, and video in a single place so you can access them anywhere and share them everywhere.

www.mediafire.com