TPS61094: TPS61094 charge voltage limitation (2.7V vs 3.6V) for Li-SOCl₂ + LIC1020 (80F) in NB-IoT smart meter – which setting is recommended?

Hello TI Team,

We are designing a battery-powered smart water meter using NB-IoT communication, with a target lifetime of ~10 years.

System details

• Battery: 3.6V, 19Ah Li-SOCl₂ (primary lithium)

• Supercapacitor: LIC1020, 80F, 3.8V rated

• Regulator: TPS61094

• Output voltage: 3.3V

• Load profile:

  • Sleep current: ~45 µA

  • Active current: ~70 mA

  • Peak current: ~330 mA (NB-IoT TX, ~1 sec/day)

Challenge

We observed that TPS61094 supports discrete charge termination voltages, and there are no selectable values between 2.7V and 3.6V.

This leaves us with only two practical options:

• 2.7V

• 3.6V

Concern

1. If we select 3.6V:

   • Charging will only occur when battery voltage is near full (~3.6V)

   • Once battery drops below ~3.3–3.4V, the LIC will no longer charge

   • This may lead to loss of supercap buffering during mid/end-of-life

2. If we select 2.7V:

   • Charging will work across most of the battery discharge curve

   • However, stored energy in the LIC is significantly reduced (~50% compared to 3.6V)

   • This may impact NB-IoT pulse performance and efficiency

Questions

1. For a Li-SOCl₂ + NB-IoT pulse load application, which charge voltage would TI recommend:

   • 2.7V (lower energy, full-life operation)

   • 3.6V (higher energy, limited charging window)

2. Does TI have any guidelines or reference designs for selecting supercap charge voltage in long-life metering applications?

3. Is it acceptable to operate the LIC continuously at 2.7V from a performance and lifetime perspective?

4. Would TI recommend an alternative device (within TI portfolio) that allows:

   • More flexible charge voltage (e.g., 3.0–3.3V range)

   • While still supporting ultra-low quiescent current and supercap charging?

5. Does TI recommend any alternative supercapacitor / LIC / HLC devices or a different power architecture (e.g., passive buffering, different regulator topology, or hybrid approach) that would be better suited for:

   • Li-SOCl₂ battery

   • NB-IoT pulse load

   • 10-year lifetime requirement?

Goal

Our objective is to:

• Ensure reliable NB-IoT operation across full battery life

• Maximize usable battery capacity

• Avoid mid-life degradation of supercap functionality

Looking forward to your recommendation.

Thank you.