• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 62 subscribers
  • Views 104 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS61094: Migrating from TPS63900 to TPS61094 for NB-IoT smart water meter using 3.6 V Li-SOCl₂ battery – request for TI recommendation

Abhishek Hulamani
Prodigy 30 points

Part Number: TPS61094
Other Parts Discussed in Thread: TPS63900, , TPS61099, TPS61299

Hello TI Team,

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

System overview

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

  • Output voltage: 3.3 V

  • Application:

    • Water flow measurement using pulse sensing

    • NB-IoT data transmission once or twice per day

  • Load profile:

    • Deep sleep current: ~45 µA

    • Active current: ~70 mA

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

Current design

We are currently using TPS63900 as a buck-boost regulator. While it meets functional requirements, the minimum input voltage limitation restricts how much of the Li-SOCl₂ battery capacity can be utilized near end-of-life, reducing effective battery usage.

Proposed change

We are evaluating TPS61094, primarily due to its lower minimum input voltage (~0.7 V), which appears attractive for extracting a higher usable percentage of the battery capacity.

Questions

  1. Suitability:
    Is TPS61094 suitable for a long-life NB-IoT smart meter with short high-current TX pulses and very long sleep periods?

  2. Low-power behavior:
    How does TPS61094 quiescent current and efficiency compare with TPS63900 in:

    • Ultra-low load / sleep conditions

    • Short high-peak NB-IoT transmit bursts

  3. Battery discharge concerns:
    Are there any startup, stability, or brownout risks when operating TPS61094 close to its minimum input voltage, especially considering:

    • Li-SOCl₂ internal resistance increase at end-of-life

    • NB-IoT current pulse behavior

  4. Architecture guidance:
    For a 3.6 V Li-SOCl₂ NB-IoT meter, would TI generally recommend:

    • Buck-boost

    • Boost-only

    • Or an alternate architecture (e.g., bypass + boost)?

  5. TI recommendation:
    Based on this use case, can TI recommend a better-suited power solution (either within the TPS6109x/TPS639xx family or another TI device) that:

    • Maximizes usable battery capacity

    • Minimizes quiescent current during long sleep periods

    • Reliably supports NB-IoT peak currents

  6. Are there any reference designs or application notes from TI related to NB-IoT meters or Li-SOCl₂ battery-powered devices?

Our main objective is to maximize battery utilization and lifetime without compromising NB-IoT reliability or long-term stability.

Thank you for your guidance.

Please find our current schematic of the TPS63900.

WhatsApp Image 2026-02-10 at 1.48.59 PM.jpeg

  • 0
    TI__Mastermind 19790 points

    Hi,

    These two chips have completely different architectures, so you'll likely need to choose based on your specific application needs. I'm the AE responsible for TPS61094, so I can only offer some application-specific advice.

    The TPS61094 is suitable for your application. For Iq, the TPS61094 has a slightly lower Iq. However, it's important to note that the TPS61094 differs from typical boost converters in that it has a function to charge a supercapacitor. This allows the chip to operate in boost mode during VIN removed, ensuring a constant output voltage. 

    I put a slides below to help you know more about this device. 

    TPS61094 Overview.pptx

    I'm not sure if you need this functionality. If not, you might need to consider other low IQ boost converters, such as the TPS61299 and TPS61099. These can also reduce the input voltage to 0.7V. However, if you have further questions about the two chips I mentioned, you might need to submit a separate E2E thread, which will then have a dedicated AE to handle.

    Hope this can help. Thanks.

    Aurora