Charger isolate circuit question

[Mr.B]

Good afternoon all…

I have a proven circuit in a small temperature transmitter.
If you are looking at the schematic – ignore Q2 for the moment and assume the battery –ve is connected to the circuit 0v.
It is primarily supplied by a 18650 Li battery – with internal over/under protection.
This battery can be charged by a standard 5v wall wart using the included CN3063 charger chip.
The MCU is supplied by a 3.3v LDO regulator.

The principle of operation is simple:
When 5v is supplied Q1 FET will turn off, preventing current flow back to the battery.
The 5v supplies the LDO via diode D2.
The CN3063 chip starts charging the battery.
When the 5v is removed, Q1 turns on and allows the battery to supply the LDO.
The 4u7 cap is a requirement of the charger chip.

That’s the background – yes, it works, proven in the field.
Now my question:

These devices are used in sub-zero temperatures.
The staff have been instructed not to plug them into the charger until 24 hours after the device has come out of the freezer.
As this cannot be guaranteed, I want to introduce an extra level of protection for the Li battery.
This is where the FET Q2 comes in…

Can I put Q2 in circuit here?

Further information:
The charger chip does not have an enable input.
The MCU receives the following feedback information:
The CN3063 chip has two open collector indicators – Charged and Charging.
The MCU can determine when 5v is applied by using these.
There is a DS3231 RTC in the circuit (not shown) that can also provide the on board temperature.
The entire circuit is potted in epoxy so an assumption can be made that the battery is at roughly the same temperature as the DS3231.
Once the temperature of the overall circuit raises to a safe level the MCU can activate CHARGE_ENABLE.

Datasheet for Q2 also attached.

Appreciate your feedback.

[Ian.M]

It looks like the TEMP pin can be used to enable/disable the charger.  The datasheet says the setpoint is 46% of Vin, so you need to transition from about 2/3 Vin to 1/3 Vin, which should be fairly easy to do with a potential divider of 2x 10K between the 3.3V rail and a MCU I/O pin.   You cant just pull it hard low as that disables the TEMP pin.

However the datasheet sucks, so some tests with a 10K pot from Vin to 0V, wiper to TEMP, are advisable to find the actual thresholds to confirm the datasheet isn't telling porkies.

[Mr.B]

Thank you for your time Ian.

Yes, the datasheet sucks alright...

I will have a play with the TEMP pin and a pot.

However, the confusing thing thing is:
The datasheet does say to ground the TEMP pin if you are not using it...
Interesting when it also says "If TEMP pin’s voltage is below 46% of input supply voltage VIN for more than 0.15S, this means that battery’s temperature is too high or too low, charging is suspended."

I will have to experiment with the thresholds to see what happens and where.

Attached copy of datasheet - should have done that in the original post.

[Mr.B]

Update for those who are interested.

The datasheet is crap.
The behavior of the TEMP input is complete lies - unless my pot value was too high and the pin needs some current.
Using a 100K pot I panned the full range from 0V to charge input (5V1).
No change, the device continued to charge the battery regardless.
I tested 5 devices - consistent behavior.

I then built a dead bug version using the FET solution.
Works perfectly.

[Ian.M]

I was suspicious about the datasheet, but I don't think you can definitely say the TEMP input is non-functional till you've tried starting with it at 2/3 Vin, using a resistor network closer to typical battery pack thermistors than a 100K pot.  If it checks that pin at initial powerup or battery connection it might latch its disabled state if you start with it grounded.   However, as you tried 5 units, the odds are you didn't start from 0V for all of them so it may well be non-functional. 

[Mr.B]

I tried all 5 with a "power down - set pot - power up" methodology.
I tested 0 to 100% in approximately 5% steps.
You may be right about testing with resistor values closer to that of a real LiPo battery pack thermistor.

[2ni]

Just a silly question: Why is D1 needed?

[Mr.B]

Polarity protection from users who think they can fix broken charger connectors, but consistently get it wrong...