PowerBook 500 Battery Recell

[Mike Richardson]

Has anyone else re-celled a 500 series intelligent battery?

The battery reports only 1752 mAh. The new cells are capable of 2700 mAh. How do I get the battery to fully utilize the cells?

 

[beachycove]

I have recelled one too and also get the 1752 mAh reading in one or other of the 500-series battery utilities. I presume that the 'intelligence' of the intelligent battery software is to blame, but how you'd fix that is way above me.

 

[Bunsen]

What's the real-world performance like though?

 

[beachycove]

Mine gives 80 mins. or so, as I recall. I can live with that, even if I would like to live with more.

 

[wally]

...how you'd fix that...

Two ideas, both untried...

1. Start with fully externally trickle charged cells, see if by repeated usage cycles if the proprietary intelligent chip SC406346DW will "learn" that there is more capacity than the default internal programming. Probably not, but some chips may be able to adapt upwards; they all have to be able to adapt gradually downwards to accommodate cell aging. There may be a learning reset means but I do not know about it.

2. PROVIDED (for safety) that the cell spec allows max charging and discharging current to scale as capacity (most NiMH cells do), hack the current shunt sensing resistor-reduce its ohms proportional to the ratio increase in cell capacity. You would have to make a precise 4 terminal measurement of voltage drop when it was passing say an amp, then either shunt it or replace it altogether with a proper low temp coef shunt resistor of the appropriate value and current carrying capacity. The proprietary programming could remain the same, the lower value shunt would lead it to believe fewer coulombs per second were flowing, and it would have its unchanged expectations about both discharge and charge cycles externally scaled by your reengineering of the shunt. In the 5xx battery the shunt is that zig zag metal thingy that sits in the 5/8 inch long slot cut into the intelligent battery board. For safety, it would be good to have an engineer review your particular hack and compare the spec sheets of both original and new cells.

 

[Bunsen]

If it doesn't turn out that the "Intelligent" controller can be adapted to the higher capacity cells, would it be worthwhile buying lower-capacity (and presumably cheaper) cells in future?

 

[wally]

Often the lower rated capacity choices will have lower internal impedance, and more reliable internal construction (less likely to short if dropped, or short due to aging). During the engineering development cycle there is a battery evaluation group that struggles to rank the cell choices, both by datasheet comparisons and by sample testing the available choices at representative loads that the laptop will present to the battery. For example if the cells are capacity rated at a 5 or 8 hour discharge rate and the laptop fully discharges a battery pack of those cells at a higher rate over approximately 3 to 4 hours, it is possible for some slightly lower rated capacity cells to stay above the cutoff voltage longer because of lower internal ohmic losses. Demonstrated performance and reliability are just the starting point leading to a extended discussion with qualified vendors over process control as well as contract terms and conditions, including who pays for the recall if there is one.

Now for us individuals contemplating buying just one set of cells for one battery, some guesswork is required, and trying to rank choices by reasoning alone is difficult. If you can find the datasheet for the original cells in the internet archives, compare it with your replacement choices. For acceptable delivered (not 5 hour rated) capacity it is desirable that the discharge curves at laptop load were as flat or flatter (equal or lower internal voltage drop) than the original cells. If multiple choices meet this criteria then proceed to testing or guesswork. Certain cells physically smaller than the original might be acceptable if they have enough current rating to survive the charge rate of the original design, and if when installed the thermal sensors are moved to make good contact with an improvised heat spreader to ensure that the sensors sense temperature representative of the pack.

 

[Mike Richardson]

There is supposed to be a way to get the battery to recalibrate its' self and determine the new maximum voltage. You must discharge the battery to a certain voltage according to the LIND tools. I tried this with the prototype and could not get it to recalibrate.

 

[wally]

If the 5xx series battery intelligence is using learned voltage settings as a way to estimate time remaining on a battery, then forget what I said about hacking the shunt. That has a chance of working only if the intelligence works by trying to keep track of mAH flowing in and out of the cells. If that NiMH gauge works on guessing the remaining capacity from voltage, messing with the shunt would likely make things worse by making the gauge algorithm see inaccurate estimates of the PowerBook current drain while it was trying to infer the battery internal voltage drops and compensate for them as a function of varying load.

On the learning thing, be sure to give it multiple attempts to learn in case it is taking some kind of moving average. On some of the newer commercial chips learning is kind of a gradual adjustment towards the eventual new setting.

 

[beachycove]

I watched this thread with some interest, having re-celled my 540c battery successfully (though with similar, mixed results using 2300mA tabbed AA cells) and want to make a small suggestion.

I have a new theory that these and similar (e.g. Duo) batteries could be much more easily re-celled by using appropriately rated NiMh AAA cells held in those standard plastic battery holders with leads that you can buy in most any electronics shop. These can be glued into the original battery case or shell, with original fuses etc. wired between the holders and then taped or glued against the cells in the holders as required. These would fit easily, there would be very little soldering required, there would be no soldering of batteries as such, and thus no heat damage done to the cells.

I have been experimenting with re-celling various 68k PowerBook batteries over the past three or four years, and have found repeatedly on using high capacity AA cells that the power management systems you are working with in these machines simply do not allow the cells to charge to their full capacity. It is NOT hard, in other words, to get the battery life originally promised in Apple documentation for a machine — 2.5 or 3 hours for a Duo, for example, is about what you get when you re-cell a Duo battery with high capacity cells — but I doubt that anyone, anywhere has significantly exceeded an originally specified battery life on a Duo, say, though re-celling a battery. Thus it is pointless to replace original cells from 1993-95, which were typically rated somewhere around the 1000mA mark, with 2600mA cells from 2009, because in a 68k PowerBook, you'll simply never get the benefit of the higher capacity cells. I would bet that the situation with the early PPC PowerBooks like the 1400c that used NiMh cells would be the same.

Later powerbook batteries that contained much more sophisticated power controller boards in their battery packs were different: replacement batteries manufactured five years on for Wallstreets and Pismos, for instance, really do give better battery life, not only because the cells used are higher in rated capacity, but because the packs themselves contain specially manufactured circuitry that works both with the cells themselves and with the power management system in the machine, telling it that it can charge the cells further.

That, as I say, is my theory.

If this is so, then in view of the very considerable difficulty involved in re-celling a 540c battery, why not forego the hassle of a) sourcing tabbed AA or 4/5 AA NiMh cells and B) trying to shoehorn them into one of those impossibly small battery cases, when c) you can very conveniently and cheaply stick some AAA cells in there in trays?

I do not recall the exact configuration of a 540c battery at present, but a Duo Type II battery that I have open on my workbench contains the following: + terminal <-> 3 cells <-> fuse <-> 4 cells <-> fuse <-> 3 cells <-> - terminal, with a couple of additional sensors that are wired from the central contact points on the battery, and that go in between these cells, much as do the fuses. It would be vastly easier to replace all this with AAA cells as described than to try literally to duplicate the original. As those who have tried will know, the tight fit makes the work almost impossible.

So use of AAA cells would be more convenient, but I speculate that it would also be just as effective. The AAA cells you pop in, if rated at 1300mA in particular (these are not common, but they are now available), will reasonably approximate the rating of the original cells in a 68k PowerBook. Therefore, you are likely to get every bit as much juice from them as you do from a set of 2300-2600mA AAs that will never charge to their full potential.

That is now my working hypothesis. I have working 540c and Duo batteries for the moment, but sometime in the Autumn or winter I'm going to give it a go. I have a nice little PowerBook 190, onto which I transplanted an active matrix screen from a 5300, that needs a tiny bit of battery-related TLC, and I now think I know just how to go about it....