Quite a numbers of factories’ products were found that their functions couldn’t work after they were delivered to customers, even during testing and package. Then checked the batteries (AG or lithium button cell), found that the batteries have swollen, or gotten heat if they were installed for a short time. If tried to change those batteries and tested them again, found that the products can work normally. This problem often occurs in 4E of 9 series of voice IC.
For this problem, we should have some tests as follows:
1. Firstly measure the current of IC, and use an oscilloscope to measure the voltage of OSC pins.
2. Reset test: use a new battery to do power on reset, and then a reset current will generate. After a short while, the current will drop to below 1uA. Continue to power on IC several times, the IC can be reset normally.
3. Anti-interference test: After IC finished resetting, input the pulses with different width (1~20us) to the input pins. Then check whether IC works normally and monitor the change of Isb. Finally we can find that the voltage of OSC pins become low responding to the pulse, which indicates IC exit sleep mode. Meanwhile the current will increase to 2mA. But it will drop to below 1uA quickly rather than maintain at 2mA. Sometimes the reset current can’t return to 1uA but remains at 2mA. If in that condition, IC cannot get back to sleep mode and the batteries will be exhausted gradually.
Short pulses can always awake the voice IC. This has nothing to do with the length of debounce time. Although a long debounce time can let the voice IC not respond to short pulse, it can awake the voice IC.
Conclusion:
This is the problem of electromagnetic interference in voice IC. To solve this problem, we can first parallel connect a 102 capacitor and a 10K resistor to input pins, which can eliminate interference from input port. Then monitor them for 2 hours, you will find that no action occurs in IC, and the current maintains at 0uA. It indicates that the input signals are all eliminated by capacitor.