0
\$\begingroup\$

I’m designing a nRF52832 SOC based device and I came across this issue with energy buffering and battery voltage drop.

I am using a CR2032 battery, nRF52832 SoC and a VCNL3040 proximity sensor.

This sensor has a pulse current up to 200mA for a few milliseconds (attached image below - captured one measurement,) but the battery is rated only for max 0.2 mA continuous load and for peak load the capacity drastically decreases, as it can be seen in this article.

The problem is that these high current pulses are causing a voltage drop from the battery. When the battery is older and cannot deliver enough current, the voltage drops below the operating voltage limit causing a hard reset of the device.


One more important thing is that the device is designed for a long battery life. Most of the time, it is in sleep mode. The sensor measures only once in a few minutes. A BLE data transfer is done once in a few days, therefore the average power consumption is around 5 µA.

As I understand it, the workaround would be to attach a capacitor which will cover the current pulses. That would mean - if I assume full consumption of the proximity sensor 200mA for approximately 5 milliseconds (4 peaks of 1.25 ms) and 100mV voltage drop - I will need a 10 mF capacitor (200 mA)(5 ms)/(100 mV). In this case, I will need a capacitor with high discharge current rating and as low leakage current as possible - I imagine leakage current around 1 uA, but I don’t know how realistic this is.


I found a capacitor of desired capacity and low leakage current (CPH3225A,) but the rated maximum discharge current is only 10µA. According to the datasheet it seems that when increasing the discharge current the capacity decreases.

Is there any way around the limitations?

\$\endgroup\$
4
  • \$\begingroup\$ You could disconnect the cap from the battery when you're not transmitting. You'd need a circuit to limit inrush current, but that wouldn't be hard if you already have an onboard microcontroller. Another idea is boost the battery voltage to charge the cap and then put a linear regulator on the output so that you increase the allowable voltage change. Won't be efficient though. \$\endgroup\$
    – Jon
    Feb 4, 2023 at 16:02
  • \$\begingroup\$ Even electrolytic capacitors could do the trick. If their voltage rating is much higher as needed the leakage at low voltage can be low enough. I do not see why you chose a double-layer cap... \$\endgroup\$
    – datenheim
    Feb 4, 2023 at 19:10
  • \$\begingroup\$ @datenheim electrolytic capacitors capacity around 10mF are rather big and also the leakage current is (assume typical) I=0.01CV or 3μA whichever is greater, so that would mean for 16V capacitor having leakage current 300 μA ( 0.01*10000μF*3V ) \$\endgroup\$
    – Letho
    Feb 6, 2023 at 18:25
  • \$\begingroup\$ Right, I underestimated the leakage current a bit. \$\endgroup\$
    – datenheim
    Feb 6, 2023 at 21:23

1 Answer 1

2
\$\begingroup\$

My car key fob has a CR2032 battery and a small lithium rechargeable battery in order to support a relatively powerful transmitter. So that is another way.

In order to see how much capacitor leakage will affect the battery life, you can calculate the internal self discharge current and look up the capacity of the cell.

The capacitor you linked has an internal resistance of 160Ω (presumably typical) so it's not going to be remotely appropriate for supplying 200mA from a ~3V fully-charged voltage. Short-circuit current would be less than 20mA, so it's hardly better than the button cell itself (in fact, it's aimed at replacing such primary cells).

\$\endgroup\$
6
  • \$\begingroup\$ How often do you have to charge that rechargeable battery of your keyfob? \$\endgroup\$
    – Designalog
    Feb 4, 2023 at 16:24
  • \$\begingroup\$ @ErnestoG It charges itself from the CR2032, which lasts around a year or two. \$\endgroup\$ Feb 4, 2023 at 16:29
  • \$\begingroup\$ Are these systems normally using some sort of lipo battery charging IC to interface between both batteries, and is the lipo normally feeding the large-current peaking device (no LDO/SMPS in between)? \$\endgroup\$
    – Designalog
    Feb 4, 2023 at 16:57
  • \$\begingroup\$ @ErnestoG TBH, I didn't reverse engineer the circuit, merely re-soldered the fractured battery connection and put it back together. The LiPo battery would be directly feeding the transmitter circuit for sure. \$\endgroup\$ Feb 4, 2023 at 17:05
  • \$\begingroup\$ I wouldn't have thought of the rechargeable battery -- but I can see sense in it. This would only work for room-temperature or strapped-on-people devices, though. LiPo cells, like most rechargeables, really don't like excessive hot or cold. \$\endgroup\$
    – TimWescott
    Feb 4, 2023 at 18:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.