'328P & DS1307

I did a board for a customer, DS1307 wired up just like the datasheet shows: coin cell battery, 32.768 HKz resistor, 4.7K pullups on SCL & SDA, Squarewave out not connected, power from 5V. Nothing unusual, nothing I haven’t done before.
The battery terminal is measuring 0.4V when the board is powered and no battery is installed.
With no battery, the RTC won’t keep time at all, just outputs a static # that I think is all 1s (0xFF).

Circuit is eating 12mm coin cell batteries - going dead in just 24 hours. What could make them get used up like that?
DS1307RTC.png

Barring no shorts on the PCB: If there is a capacitor across the battery, maybe it is faulty. Try cutting the SCL and SDA pins to see if the battery voltage stays up after 24Hs. Bad battery holder. Bad RTC chip. Bad batch of batteries.

.

No cap across battery.
Battery holder had been shorting battery + to Gnd, that was already resolved by removing the top layer of the battery pads. They sent data showing 3.2V on a fresh battery, <24 hours later it was dead.
Is there any way to test the chip outside of removing it? They are soldered in.
Is suppose it could be the batteries. The layout is not very complex in that area. Pin 4 is Gnd, 7 is open.
I did a screen capture, Eagle 7.7 shows the signal names on traces. Kinda nice.
I don’t think they want to cut traces and then ship the boards with white wires.
RTC_Layout.jpg

Battery holder had been shorting battery + to Gnd, that was already resolved by removing the top layer of the battery pads.

Sounds suspicious.

If the RTC was in a 8 pin socket it would be easy to leave the SCL and SDA pins bent out of the socket.

Maybe jerry-rig an external battery, monitor the current draw then power down the cct. See how much current was being taken in power down mode.

I use DS3231 RTCs modules "Plug and Throw" Plug it in, if it works, good, if it doesn't work throw it away. ;)

.

Maybe jerry-rig an external battery, monitor the current draw then power down the cct.
See how much current was being taken in power down mode.

Masking tape battery to header:
2017-04-02_20-52-23.jpg

The battery holder is thru hole, the pads extend inside the case. Apparently there was enough solder on them and the battery lip was narrow enough that the battery Gnd was making contact. They unsoldered the connector, scraped the top layer of pad off, and got rid of the short that way. The + traces lead away from the battery on the bottom layer.

I'll have them try testing the battery current draw. I have an 4.5 year old sparkfun DS1307 module, it's usually connected to a powered down 12-shift register card that has clock code running. It's battery still has 3.082V on it.

The chip is soldered in.

Thanks for the ideas.