Designing IoT Hardware You Throw Away: What You Can Learn from Disposable Medical Devices Like CGMs

Exploded view of a disposable continuous glucose monitor showing its sensor, circuit board, battery, and adhesive patch

A Dexcom G7 has an ARM Cortex-M processor, a Bluetooth radio, an electrochemical sensor front-end, and a custom antenna—and it goes in the trash after 10 days. An Abbott Lingo is similar silicon, similar sophistication, 14 days, same destination: the bin.

That’s not a failure of engineering. That’s the entire product architecture.

If you’ve spent your career designing hardware meant to last years, this feels wrong. You’ve been trained to think about field serviceability, over-temperature derating, firmware update paths, and 10-year component availability. Disposable IoT device design throws most of that out. What replaces it is a ruthlessly constraint-driven process where a 14-day lifetime cascades into every decision you make, from the SoC to the battery chemistry to whether your enclosure uses screws (it doesn’t).

This guide walks through the major hardware decision points for designing single-use wearable electronics, using continuous glucose monitor hardware as a running case study. But the principles apply equally to disposable logistics trackers, environmental sensors, event wearables, or any connected product designed to die on schedule.

This is not a medical device regulatory guide. It’s a hardware architecture guide using medical devices as inspiration.

Start with the Lifetime Budget: Everything Flows from One Number

Before you pick a chip, a battery, or a wireless protocol, answer one question: how many hours does this device need to operate?

For a CGM, that’s 240–336 hours (10–14 days). For a cold-chain logistics tracker, maybe 72 hours. For a festival wearable, 48 hours. This number is the single most important design input you have, because it directly constrains every downstream decision:

┌─────────────────────────────┐
│   DEFINE DEVICE LIFETIME    │
│         (e.g., 14 days)     │
└─────────────┬───────────────┘
              │
              ▼
┌─────────────────────────────┐
│   DERIVE ENERGY BUDGET      │
│   (lifetime × avg power)    │
└─────────────┬───────────────┘
              │
              ▼
┌─────────────────────────────┐
│   CONSTRAIN CONNECTIVITY    │
│   (duty cycle, tx power,    │
│    protocol selection)      │
└─────────────┬───────────────┘
              │
              ▼
┌─────────────────────────────┐
│   SELECT SoC + BATTERY      │
│   (integrated, lowest BOM)  │
└─────────────┬───────────────┘
              │
              ▼
┌─────────────────────────────┐
│   DESIGN ENCLOSURE + PCB    │
│   (minimize assembly steps) │
└─────────────────────────────┘

A fixed lifetime also eliminates entire categories of engineering work. Component derating? You don’t need to worry about electrolytic capacitor aging over 10 years if the device lives for two weeks. OTA firmware updates? Your device’s lifespan is shorter than your sprint cycle. Reliability testing narrows to “does it survive 14 days of body-worn use?” instead of “does it survive 5 years in the field?”

If you don’t nail the lifetime budget first, every downstream decision is a guess.

Chip Selection: Integration Over Flexibility

In disposable IoT device design, the SoC is the system. Your goal is the fewest chips possible on the smallest possible board, at the lowest possible unit cost. Flexibility, processing headroom, and future-proofing are irrelevant. You’re designing for a 14-day device that will be redesigned in 18 months anyway.

“Integration” means something specific here:

  • MCU + BLE radio on one die. No external radio module. The TI CC2642R, Nordic nRF52832, and Renesas DA14585 all combine an ARM Cortex-M core with a 2.4 GHz radio.
  • On-chip ADC sufficient for your sensor front-end. CGMs need to read an electrochemical sensor. A 12-bit ADC at modest sample rates is fine. If the SoC has it, you’ve eliminated an external ADC IC plus its supporting passives.
  • On-chip voltage regulation. Avoid external LDOs or DC-DC converters if the SoC’s internal regulators can run directly from your primary cell.
  • Minimal external passives. Integrated matching networks, single-crystal or crystal-less operation, and on-chip decoupling all reduce BOM lines.

For CGM chip selection specifically, Abbott’s Lingo uses a TI CC26xx-family part. The reasoning is sound: the CC2642R offers sub-1µA standby current, a capable 12-bit ADC, good analog peripherals for sensor interfacing, a mature BLE 5 stack, and competitive volume pricing, reportedly sub-$1.50 at scale for 100k+ quantities. Nordic’s nRF52 series and Renesas’s DA14585 are both viable alternatives with similar integration levels. The point isn’t which chip “wins.” It’s the selection criteria for disposable design:

PRIORITY FOR DISPOSABLE IoT CHIP SELECTION
──────────────────────────────────────────
1. Unit cost at volume (100k+)
2. Sleep/active current ratio
3. Level of integration (radio + MCU + analog)
4. Supply chain reliability
5. Development ecosystem maturity
──  ──  ──  ──  ──  ──  ──  ──  ──  ──  ──
NOT priorities:
- Processing headroom for future features
- Wide operating temperature range
- Long-term availability (10+ year lifecycle)

Notice what’s missing from the priority list. You don’t need a chip that’ll be in production in 2035. You need one that’s cheap, low-power, and highly integrated right now.

Power Architecture: Primary Cells and the Art of the Tiny Energy Budget

Primary cells are the default for disposable products, and it’s not even close. Non-rechargeable batteries eliminate the charging circuit (IC, inductor, protection FET, USB connector), which saves BOM cost, board area, assembly steps, and user interaction. No one charges a CGM.

Common chemistries for disposable wearables include silver oxide (stable voltage, good energy density), zinc-air (high energy density but activated on unsealing), and thin-film lithium (tiny footprint, low current capability). Your choice depends on the current profile and form factor requirements.

Here’s a back-of-envelope power budget for a CGM-like device:

EXAMPLE: 14-DAY DISPOSABLE WEARABLE POWER BUDGET
─────────────────────────────────────────────────
State          │ Current │ Duration  │ Duty     │ Avg Draw
───────────────┼─────────┼───────────┼──────────┼─────────
Deep sleep     │ 1 µA    │ Continuous│ ~99.9%   │ ~1.0 µA
Sensor read    │ 5 mA    │ 3 ms      │ 1/5 min  │ ~0.05 µA
BLE advertise  │ 8 mA    │ 2 ms      │ 1/sec    │ ~16 µA
BLE connection │ 10 mA   │ 15 ms     │ 1/5 min  │ ~0.5 µA
───────────────┼─────────┼───────────┼──────────┼─────────
TOTAL AVERAGE  │         │           │          │ ~18 µA
───────────────┴─────────┴───────────┴──────────┴─────────
14 days × 18 µA ≈ 6 mAh minimum (add 2–3x margin → ~15–20 mAh cell)

Fifteen to twenty milliamp-hours. That’s a silver oxide coin cell like a Renata 377 (28 mAh) with room to spare, or a thin-film lithium cell like an Enfucell SoftBattery. The math is almost absurdly small, and that’s the point. A well-designed disposable sensor connectivity architecture keeps the average draw in the low-tens-of-microamps range.

The critical insight: the battery is often the single largest physical component in a disposable wearable. It dictates board outline, enclosure volume, and product form factor. Design around the battery first, PCB second.

One more advantage of primary cells that’s easy to overlook: shelf life. A silver oxide cell retains 90%+ capacity after 5 years on a warehouse shelf. That matters enormously for a product that sits in retail distribution or hospital inventory before activation. Rechargeable cells self-discharge and require charge-before-use workflows that disposable products can’t tolerate.

Why BLE Wins for Disposable Wearables (and What to Watch Out For)

BLE is the right connectivity choice for disposable wearables. Not “it depends.” BLE. Here’s why:

  • Universal smartphone compatibility. Every iOS and Android phone has BLE. Your gateway device is already in the user’s pocket.
  • Excellent power profile. Intermittent, low-throughput data transfer at microamp-level average currents. Exactly what a disposable sensor needs.
  • Chip-level integration. The SoCs discussed above have the radio on-die. No external module, no additional BOM lines.
  • Mature protocol stacks. GATT profiles exist for glucose monitoring, heart rate, and generic sensor data. You’re not building from scratch.

What you need to design around:

Range is limited. Expect ~5–10 meters for a body-worn device with a small antenna. That’s fine for phone-in-pocket use cases, but don’t pretend it’ll reach across a house.

Throughput is low. Plan accordingly. CGMs transmit a few bytes per reading. If your use case generates kilobytes per second, BLE is the wrong protocol.

Connection intervals are phone-controlled. iOS and Android negotiate connection parameters, and they prioritize their own battery life over your device’s wishes. Design your firmware to handle worst-case intervals (up to 30ms intervals on iOS, potentially longer on some Android implementations). Test against real phones, not just a BLE development kit.

Security is non-optional. Even a 14-day disposable device handles user data. BLE pairing and encryption (LE Secure Connections, not legacy pairing) must be implemented correctly. Disposable doesn’t mean insecure.

BLE is the wrong choice if you need device-to-cloud connectivity without a phone (consider LTE-M or NB-IoT, but your BOM jumps by $5–10+), or if you need multi-hundred-meter range (a different product category entirely).

Mechanical Design and PCB: Optimize for Assembly, Not Repair

No screws. No clips. No battery door. No user-accessible anything. A disposable wearable is a sealed unit, and that simplification is liberating.

PCB considerations:

Target a 2-layer board. With a highly integrated BLE SoC, your component count is low enough that 4-layer boards, and their ~2x cost premium, are often unnecessary. You’ll need careful RF trace routing for the antenna path, but a well-designed 2-layer board with a solid ground pour can meet FCC requirements.

Minimize board outline. The battery footprint usually defines the minimum, so your PCB design is essentially “fit everything in the battery’s shadow.”

Design for panelization. High-volume pick-and-place favors rectangular boards with consistent orientation. Odd shapes and breakaway tabs cost assembly time.

Enclosure considerations:

Ultrasonic welding or adhesive bonding, not snap-fits. Permanent seals are faster to assemble and cheaper than mechanical fasteners.

Skin-contact materials matter even for non-medical products. A logistics tracker lives in a box; a wearable lives on skin. Medical-grade adhesives (3M 4076-series, Scapa healthcare tapes) and biocompatible plastics (USP Class VI) prevent skin irritation. Irritation causes returns, and returns are expensive.

Minimize unique parts. Every distinct component is a BOM line, a supplier relationship, an incoming inspection step, and an assembly station. Fewer parts isn’t just cheaper; it’s faster and more reliable at volume.

The cost equation reframed: At 100k+ units, saving $0.10 on a single component saves $10,000 per production run. Shaving 2 seconds off an assembly step at $30/hour labor saves $16,667 per 100k units. This level of cost-consciousness isn’t a compromise. It’s an engineering discipline.

COMPONENT          │ TRADITIONAL IoT │ DISPOSABLE IoT
───────────────────┼─────────────────┼───────────────
Rechargeable cell  │ ✓               │ ✗
Charging IC + USB  │ ✓               │ ✗
External LDO/DCDC  │ ✓               │ ✗ (use SoC internal)
External ADC       │ ✓               │ ✗ (use SoC on-chip)
Debug connector    │ ✓               │ ✗ (test pads only)
Mechanical screws  │ ✓               │ ✗
Display            │ Often           │ ✗ (phone is display)
Typical BOM lines  │ 40–80+          │ 12–25

What You Can Skip (and What You Absolutely Cannot)

Disposability gives you permission to simplify. But “disposable” doesn’t mean “unregulated.”

You can skip:

  • OTA firmware update infrastructure. Device lifespan is shorter than your release cycle. Flash the firmware at manufacturing and move on.
  • Extended environmental stress screening. No MIL-STD-810 needed for a 14-day indoor wearable. Test for your actual use conditions.
  • ESD protection on user-inaccessible interfaces. If it’s ultrasonically welded shut, nobody is touching the test pads.
  • Over-specifying component temperature ranges. Commercial-grade (0–70°C) is fine for body-worn devices. Don’t pay for industrial or automotive ratings.

You absolutely cannot skip:

  • RF certification. FCC Part 15 (US), CE RED (EU), and equivalent certifications for BLE are non-negotiable. Disposable or not, it’s a radio.
  • Battery safety. UN38.3 testing for lithium cells. Shipping classification (IATA/DOT). Your logistics team will thank you for addressing this early.
  • Biocompatibility testing if the device contacts skin. ISO 10993 cytotoxicity and sensitization testing applies regardless of whether your product is a regulated medical device, because consumer protection regulators and your legal team both care about skin reactions.
  • ESD protection on the antenna path and any conductors exposed through the enclosure. The BLE antenna is electrically connected to the outside world by definition.

The rule is simple: disposability gives you permission to simplify engineering requirements. It gives you zero permission to simplify safety and regulatory requirements.

Building Your Disposable Design Checklist

Here’s the decision sequence, distilled:

  1. Fix the lifetime. Every other decision is downstream of this number.
  2. Calculate the energy budget. Average current × lifetime = minimum cell capacity. Add 2–3x margin.
  3. Select the SoC. Maximize integration, minimize unit cost. BLE + MCU + ADC on one die.
  4. Choose the primary cell. Match capacity to energy budget, form factor to product requirements.
  5. Design the BLE data model. Plan for low throughput, worst-case connection intervals, and proper encryption.
  6. Minimize the mechanical BOM. Sealed enclosure, no fasteners, skin-safe materials if wearable.
  7. Run the cost model. At your target volume, verify every component and assembly second is justified.

Designing a limited-lifespan IoT device isn’t about cutting corners. It’s about constraint-driven simplification: letting a fixed, short lifetime free you from the complexity that long-life products demand. Every component you eliminate is a failure mode removed, an assembly step saved, and a cost reduced.

The disposable IoT category is expanding fast beyond medical devices into perishable goods tracking, construction site monitoring, agricultural sensors, and event and hospitality wearables. Engineers who understand this design approach, who can think in 14-day lifetimes and $0.10 BOM trades, have a skillset the market increasingly needs.

Hubble Network enables Bluetooth connectivity from disposable, low-cost devices directly to satellites—no gateways, no infrastructure, no redesign. See how it works →