External battery connections to motherboards

[mbbrutman]

I'm trying to resurrect a basket-case PC compatible that I've had for ages. It's a 486 but this is more of a general hardware question.

The motherboard has a soldered rechargeable battery on it which has long since stopped working. It has a 4 pin connector for an external battery. (In the picture the four pins for the external battery are on the right side.)

• What's the correct type of battery to use here - rechargeable or non-rechargeable?
• And at what voltage rating?

The motherboard is an FIC 486-GVT-2. The manual says pin 1 is '+' while Pin '4' is negative, but it doesn't have specs. J5 (right below the external battery connector) is used to enable or disable the internal battery.

The resistors and diodes between the battery and the jumpers make me curious. I didn't buzz them out fully but at least some of those resistors and diodes are connected to the external battery connector. Which implies a non-rechargeable battery is expected there. The resistors might also be to drop the voltage, but I still don't know if it wants 4.5 or 6v.

I guess I could disable the internal battery using the jumper, then look for positive voltage on the external battery connector. That would also answer the rechargeable vs. non-rechargeable question.

Thoughts?

 

[modem7]

I guess I could disable the internal battery using the jumper, then look for positive voltage on the external battery connector. That would also answer the rechargeable vs. non-rechargeable question.

I have yet to hear of a motherboard that tries to charge the battery connected to the 'external' connector.

The resistors might also be to drop the voltage, ...

100 ohms.

Looking at a few RTC chip datasheets, the current drawn from the battery (when in use) is typically measured in tens of µA. The voltage drop over 100 ohms is practically nothing (compared to a figure of several volts)

I think a reason for the resistor is for battery/fire protection if a fault on the motherboard (before the resistor) very heavily loads the battery. If we assume a short-circuit, the current drawn from the battery is limited to tens of mA.

... but I still don't know if it wants 4.5 or 6v.

I see a DS12885Q RTC chip on your motherboard.

The DS12885Q has a dedicated pin, "VBAT", for a 3V battery (min. = 2.5V, max. = 4V). The DS12885Q datasheet contains, "The battery should be connected directly to the VBAT pin. A diode must not be placed in series with the battery to the VBAT pin. Furthermore, a diode is not necessary because reverse charging-current protection circuitry is provided internally to the device ..." From what you describe, your motherboard has a diode in series (as well as a resistor). Is this a case where the motherboard engineer just didn't care?

In regard to the VBAT pin, is also, "If a backup supply is not supplied, VBAT must be grounded." I wonder if the engineers decided to not use VBAT, and instead, maintain voltage to the DS12885's VCC pin when the motherboard is powered off. The spec for the VCC pin is {min. = 4.5V, max. = 5.5V}. But that would mean that the rechargeable battery on your motherboard must (when functional) provide say, 5V to 6V (more than 4.5V to 5.5V to cater for the voltage drop over the diode).

I wonder what is going on? Is the VBAT pin grounded on your motherboard? Do you know what the voltage rating is of the rechargeable battery on your motherboard?

 

[konc]

I have yet to hear of a motherboard that tries to charge the battery connected to the 'external' connector.

It's very rare, I've seen exactly one but it exists. After that I've been compulsively measuring all external battery connectors and never encountered another. Thankfully it's very easy to check if charging current reaches the connector.

 

[mbbrutman]

modem7,

The problem with memory is that it tends to fade. I might have seen one board that takes an external rechargeable battery so that means I have to be careful with all boards, especially when the board documentation doesn't specify anything. The existing battery on the board is spot welded to tabs and can not be removed. The rating is not obvious. I'll need to remove the motherboard to start properly buzzing out the connections.

For now I took the experimental route; I looked for charging current while the machine was powered on and there was none. Then I connected leads to an external battery pack (3 NiMH AA) to get a voltage of around 3.6 and tried that overnight, and the CMOS settings held. For a more permanent solution I'll look at the Dallas chip to see what exactly is wired so that I can give it the proper voltage. I'll probably also just use a lower voltage than the max knowing it is for temporary use, and use a CMOS save/restore utility so that I can remove the battery for longer periods. This machine was a particular pain in the rear and it kept thinking it had a CMOS password when it had no battery, so I had to get something in place just to get it booted.

The great thing about the hobby is that everything is turning into an adventure, even simple battery replacements. Stay tuned ...

 

[mbbrutman]

Coming back to this with a more measured approach ...

• Vbat is not grounded.
• A jumper selects between the rechargeable battery and the external battery.
• Both paths route the battery through a diode first, then a 100 Ohm resistor, then to Vbat.

As you pointed out the data sheet clearly says not to use a diode in series to the Vbat pin, but both paths on this mother board do that.

Here is more detail from a Dallas RTC application note:

"Some RTCs use the battery voltage as a reference to determine when VCC is at a valid level. When VCC is below the minimum value, the part will go into write protect, locking out access to the part. While the RTC is operating from VCC, the VBAT input will be at a high impedance. If a battery is not connected to the VBAT input, or is connected with diodes in series, the VBAT input can float high, causing the RTC to go into write protect."

I wonder if there was an update to the datasheet and the original data sheet did not state that.

So I think I'm going to change my approach. While the 3 NiMH AAs were safe for a test, I'm going to get a 3v coin cell holder and wire that to the external battery connector. I'm not going to bother worrying about the diodes - the motherboard made it this far. If I do have problems with the RTC going into write protect while the machine is powered on then the diode can be bypassed.