• Please review our updated Terms and Rules here

3.6v battery on 286 motherboard

ibmapc

ibmapc

Veteran Member
Joined
Apr 5, 2010
Messages
1,046
Location
Albany, OR USA
I have a 286 Motherboard with a dying 3.6v battery soldered to it. I want to replace it with one of these It's a LIR 2032 (Lithium Ion Rechargable) The holder for it is a perfect fit on the motherboard. My concern is the charging voltage. With the Motherboard powered up, the voltage across the existing dying battery is 4.0 volts(it drops to 3.0 volts after the power is off for a while and the clock loses time dramaticaly). The specs for the LIR2032 says charge voltage is 4.2 volts.
1)Will it charge at the slightly lower voltage?
2)Will the battery overheat if the charge doesn't stop when it's fully charged?
3)I wonder if it will hold its charge long enough to power the Clock and CMOS setup memory if the machine is left off for 2 or more months?

Thanks in advance

Greg
 
Last edited:
I don't think Li-ion or any other lithium based rechargeable battery is a good replacement for NiCad batteries that were used on older AT motherboards. Lithium batteries usually require special charging controller, this to make sure battery is not overcharged (and as result overheated and destroyed).

I would suggest using another NiCad or NiMH battery instead. I think small cordless phone battery is a better replacement.
 
Yep, That's what I was afraid of. Can you suggest a battery and/or source for this?

EDIT
OK, I found this. The spec say's it's not rechargeable, but I question that. I've never seen a NiCD that wasn't rechargeable. Anyone have an opinion on this?
 
Last edited:
I personally wouldn't bother with replacing one of the old rechargeable batteries as it'll just end up dead\leaking later anyways. If you have a battery header you can check to see if there's any voltage on it, and if not then just connect a more common battery to it, such as a CR2032 battery (solder a cable to a holder) or even just AA batteries.
 
Does your motherboard have a 4-pin header for attaching an external battery? If it does, I'd just go for 4 x AA alkalines in a battery holder and get the %^$^# battery off the PCB before it leaks and does some real damage. But if you've got to get a rechargeable battery, just get a cordless phone pack and wire it in. Radio Shack has a pile of different ones--or you can get one cheap from China on eBay. They're insanely cheap.
 
I don't think CR2032 will do a good job. CR2032's voltage is only 3V (vs. expected 3.6V) could be too low - MC146818A RTC minimal operating voltage is 3V. Other than that the charging current will be applied to a non-rechargeable (and pretty small) CR2032 battery. Which is not a good thing, as it could cause the battery to overheat, and so on...

On the other hand 4 x AA will provide about 6V... way too much.

The rechargeable RTC battery is usually connected through a diode and a resistor to VCC (+5V). Diode prevents battery discharge through other than RTC components, and resistor limits the charging current. With 6V (or any greater than ~ 4.4V) battery RTC will be always powered by the battery, even then the computer is ON. Probably it is not a very bad thing, since the capacity of the AA alkaline batteries is enough to keep the RTC running for years. But I would use 3 x AA batteries instead.
 

Attachments

  • RTC - Rechargable battery.jpg
    RTC - Rechargable battery.jpg
    11 KB · Views: 2
Last edited:
These old systems usually are fairly lenient on the battery voltage, you can get away with anything from 2.5V-6V really. I have my Compaq Prolinea 4/33, designed for a 4.5V battery brick, running just fine off a CR2032, keeps time and settings.
 
I soldered a CR2032 to the olld batter ypads with a diode on the + side to prevent the charging circuit from killing the cell.
 
Another vote for 3x AAA battery holder, series diode and leads. Dead easy and avoids the leaking NiCd issue. I've done the same with my fruit machine too as it happens, as well because (in both cases) the likely run time will be insufficient to keep the rechargable actually charged.
 
sergey said:
IOn the other hand 4 x AA will provide about 6V... way too much.
Click to expand...

You're forgetting at least 2 diode drops in the backup battery circuit (usually an NPN transistor). In fact, the 4xAA was packaged (with 4 pin connector) with many generic cases of the time. I still have a number of systems that use that 4xAA setup. The batter holder was packaged with a square of double-stick hoam tape for attachment inside the case.

Alkaline cells can have a very long life (years) in a setup like this. On my PC 6300, I replaced the rechargeable with a CR2032 and a schottky diode and it works fine, but IIRC, the Oki RTC chip used there is rated down to 2.2V.

The problem with NiCd and NiMH cells is that they do self-discharge. On a rarely-used system, this can be a problem.
 
Chuck(G) said:
...On my PC 6300, I replaced the rechargeable with a CR2032 and a schottky diode and it works fine, but IIRC, the Oki RTC chip used there is rated down to 2.2V.
...
Click to expand...
So with the diode, your voltage drops to about 2.3 volts. I don't know what the minimum voltage for my RTC is.
THIS is my motherboard. It looks like JP17 is the external battery header. I'll have to figure out the polarity with my ohm meter.
 
Chuck(G) said:
You're forgetting at least 2 diode drops in the backup battery circuit (usually an NPN transistor). In fact, the 4xAA was packaged (with 4 pin connector) with many generic cases of the time....
Click to expand...
Should I connect 4AA's or 3AA's to the JP17 on THIS motherboard. JP17 is a four pin header with one pin removed for external battery. Note that the original battery is only 3.6 volts. I figured out the polarity with my volt/ohm meter and there is at least one diode in the circuit so that it won't attempt to charge the external battery when the power is on. So the volts will drop about .7v per diode right? So if there are two diodes the 4AA battery will drop to about 4.6 volts. But if there is only one diode it will drop to about 5.3 volts and that has me worried that I might damage something.
 
The photo [here] shows the typical situation.

RTC chip

The RTC chip (built into a chipset on later machines) that either battery supplies is designed for a wide range of voltage.
Different chip types have different ranges.

Internal battery

3.6v NiCad (or a 3.6V NiMH).
The charging circuitry is designed for the charging requirements of a 3.6v NiCad.
Substituting with something else that is not compatible may damage the battery due to improper charging.
In this application, NiMH can be used in place of NiCad.

External battery connector

Your machine is a 286.
A 6V battery is not going to be a problem.
 
It's done! That @#$%^&ing NiCd Battery is GONE!! I'm glad that I got to it before it started to leak and ruin things. Now I'll have to leave it turned off for a few days to see if the clock keeps better time with the 4xAA Alkaliners.

Thanks to all who replied. Your advice was terrific as usual.
 
I have a couple of 486 and 386 boards on which I replaced the leaky barrel-type 3.6V batteries with coin-cell holders for CR2032s. I didn't check the circuitry but I would check the batteries every once in a while and they were always cold to the touch after a couple of hours running - does that mean the old batteries were not recharged, or should I expect fireworks after more power-on hours?
 
iulianv said:
I have a couple of 486 and 386 boards on which I replaced the leaky barrel-type 3.6V batteries with coin-cell holders for CR2032s. I didn't check the circuitry but I would check the batteries every once in a while and they were always cold to the touch after a couple of hours running - does that mean the old batteries were not recharged, or should I expect fireworks after more power-on hours?
Click to expand...
If the barrel battery that you pulled out was a NiCad (most likely) and you are using a standard Lithium (different to Lithium-Ion) CR2032 directly in its place, then your motherboard will be attempting to charge a battery that is not designed to be charged. I do not know what the consequence is.

To stop that Lithium battery being charged, a diode needs be added in series with the battery. The polarity of the diode in relation to the battery is critical. Unfortunately, the forward bias voltage drop over the diode means that the RTC chip will get about 2.5V, or less, not the 3V output of the battery. 2.5V may be too low for some motherboards. Use of a different type of diode (e.g. Germanium) will reduce the voltage drop.

blocking_diode.jpg
 
Volt drop though varies with current... a small signal diode will have almost zero drop at the kind of currents we're talking for RTC maintenance - in the uA range, anyway.
 
You must log in or register to reply here.
Back
Top