PDA

View Full Version : How to properly replace rechargeable lithium battery?



Trixter
May 27th, 2012, 11:40 AM
I'm currently restoring an AT&T 6312 WGS (http://trixter.oldskool.org/2012/05/27/data-preservation-case-study-att-6312-wgs/) and the battery lasts about 10 seconds before completely discharging, which means I have to keep going into the setup program to re-set everything including the hard drive tables. I want to replace this, but I'm not quite sure what the best way to do it is -- I can't just replace it with two AA batteries because the charging aspect might explode them, and also I don't think they'd provide enough voltage. Pictures are below:

90849085

Chuck(G)
May 27th, 2012, 11:59 AM
Jim, is that really a rechargeable battery--don't I see a "Do not recharge" right below the "Do not short-circuit" warnings? I didn't know that rechargeable lithium batteries were around yet for the 6300 series.

Nafartarra
May 27th, 2012, 12:03 PM
I have just replaced 30 min ago a battery pack of a Tandon 386. It was missing, but if I don't remember bad, is was a pack of 3 AA.
I have mounted a pack of 3 AA alcaline batteries.I think that you could mount a pack of 4 AA and a diode, reducing the voltage to around 5.3 and avoiding the recharging.

pearce_jj
May 27th, 2012, 01:25 PM
Certainly looks like "do not recharge". I'd slap in 4x AA plus a blocking diode, just to be sure.

Trixter
May 27th, 2012, 02:08 PM
It may indeed not be a rechargeable battery -- I was going off of the fact that the system could hold the BIOS settings for 10-15 seconds before they faded completely.

My electronics knowledge is weak. I thought a diode was a bit of a "check valve" -- wouldn't I want to use a resistor instead? If not, what diode would I use?

RickNel
May 27th, 2012, 02:28 PM
Sounds like you might have a short circuit if the battery is draining totally so fast. It could be somewhere close to that battery connection, or something loose inside the case shorting to earth - does the battery get hot? You can check for a short across the + and - terminals without even opening the case. If you don't have a multimeter then even a flashlight bulb and battery, with the + connection routed via the +/- battery terminals on the box, would show you whether the battery circuit is shorted.

PgrAm
May 27th, 2012, 02:59 PM
I usually replace these with modern lithium button cells (CR 2023) they can last for a long time. I just tape leads to either side of the battery and let it hand around in the case, you can tape it to the side of the case if you want it to be neat.

Chuck(G)
May 27th, 2012, 03:11 PM
A CMOS setup like this can hold settings for quite some time if there's a capacitor somewhere in the circuit. What really drains the battery is the clock. I've got a piece of equipment here that puts a 1F supercap in parallel with the battery--you have minutes to spare when changing batteries.

I'd use some alkaline cells--they have a very low self-discharge rate (at 68F, less than 2% per year).

One way to determine if a battery is in a charging circuit is to power up the equipment, remove the battery and measure the voltage across the battery holder. If there's a voltage present that's somewhat in excess of the battery's nameplate rating, it's probably a charging circuit.

SpidersWeb
May 27th, 2012, 04:34 PM
It may indeed not be a rechargeable battery -- I was going off of the fact that the system could hold the BIOS settings for 10-15 seconds before they faded completely.

My electronics knowledge is weak. I thought a diode was a bit of a "check valve" -- wouldn't I want to use a resistor instead? If not, what diode would I use?

A little voltage is lost over the diode, it's ideal for this purpose. Just a generic power diode will do, 1N4004 are common and usually rated "max 0.8V drop". 1.5V x 4 = 6V - 0.8 = 5.2V which is pretty close and it'd probably end up a tiny bit higher like 5.4V. But it would probably be fine with two coin batteries or 4 x AA just on it's own, just depends on how cautious you intend to be.

Chuck(G)
May 27th, 2012, 04:47 PM
A little voltage is lost over the diode, it's ideal for this purpose. Just a generic power diode will do, 1N4004 are common and usually rated "max 0.8V drop". 1.5V x 4 = 6V - 0.8 = 5.2V which is pretty close and it'd probably end up a tiny bit higher like 5.4V. But it would probably be fine with two coin batteries or 4 x AA just on it's own, just depends on how cautious you intend to be.

Look at the battery--it's a lithium primary cell--not a rechargeable. Rechargeable lithiums weren't introduced until 1990 (http://www.nature.com/nature/journal/v381/n6582/abs/381499a0.html)--and even then were hard to get. The WGS is a 1980's system.

No diode needed.

SpidersWeb
May 27th, 2012, 06:55 PM
Yeah I know, was just in case he wanted a closer voltage to original.

Chuck(G)
May 27th, 2012, 07:22 PM
I haven't looked at the WGS schematics, but much of the Olivetti/AT&T 63xx line could accept between about 2.5V to 7V as a source for the RTC.

I think the 6312WGS is pretty much the same as the Olivetti M28/280.

Dwight Elvey
May 27th, 2012, 09:03 PM
Hi
First, it is not a rechargable battery pack. If it was designed
for that computer, it would have no charging current.
One could check it with a current meter to be sure.
As for diode drop, I doubt you'd see more than 0.1 volt
drop for the current levels used in a RTC or CMOS RAM.
I would just use 4 cell.
Dwight

pearce_jj
May 27th, 2012, 11:53 PM
My electronics knowledge is weak. I thought a diode was a bit of a "check valve" -- wouldn't I want to use a resistor instead? If not, what diode would I use?

A diode is indeed like the check valve (in a garden tap for example) - it allows flow in one direction only. So if this had been a rechargable battery, putting a diode in line with a replacement alkaline battery pack to allow current to flow *from* the battery pack would do the trick, since it will block the charging current presented by the host when it's powered on. But clearly that isn't needed here since the battery definitely isn't rechargeable anyway.

I recently modded a fruit machine in exactly this way as it happens, the datasheet for the CMOS RAM chips in that showing wide voltage tolerance in standby mode (2 - 5V IIRC).

Trixter
May 28th, 2012, 08:52 PM
Based on Chuck's comment where the specs were from 2 to 7v, I decided that a full 4AA 6V pack was worth a shot, and it seems to be working fine with no ill effects. Thanks for the advice!