BLE Module Sourcing: nRF52840 vs ESP32 Compared

Sourcing BLE 5 modules from China comes down to three decisions: Nordic nRF52840 vs Espressif ESP32 (and which variant), how to handle FCC and CE certification without starting from scratch, and how to assess supply chain stability for a 3–5 year product lifecycle. Getting any of these wrong adds 2–4 months and $5,000–15,000 to your first production run.

nRF52840 vs ESP32: what actually matters for sourcing

Both chips are manufactured in volume in China. The differences that matter for sourcing are not the spec sheet differences — they are the ecosystem, module form, and certification paths.

nRF52840 (Nordic Semiconductor): Cortex-M4, BLE 5.3, Thread, Zigbee, NFC. The go-to chip for IoT devices requiring high RF performance, low power, and protocol flexibility. No WiFi. Sold as a bare die (not useful for most buyers) or embedded in modules from Laird (DVK-BL5340), u-blox (ANNA-B112), Actinius (Icarus), and Nordic’s own nRF9160-DK. In China, manufacturers from Shenzhen produce lower-cost nRF52840 modules based on the Nordic die. The same ecosystem also produces Zigbee modules for mesh networks, long-range LoRa modules for low-power telemetry, and low-cost BLE sensor nodes.

The key sourcing distinction: Nordic does not manufacture the nRF52840 in China — it is fabbed at TSMC and tested in South Korea. Chinese module manufacturers buy genuine Nordic dies from authorized distributors and assemble them into modules locally. The risk is counterfeit dies: a module housing that contains a re-marked STM32 or clone device instead of the nRF52840.

ESP32 family (Espressif Systems): much broader product range — ESP32 (dual-core, WiFi + BLE), ESP32-S3 (adds AI acceleration, USB), ESP32-C3 (single-core RISC-V, WiFi + BLE 5, lower cost), ESP32-H2 (Zigbee + Thread, no WiFi), ESP32-C6 (WiFi 6 + BLE 5 + Zigbee/Thread). Espressif is a Chinese fabless company; manufacturing is at TSMC. Official modules are sold under the Espressif brand and through Ai-Thinker, a Shenzhen-based module manufacturer with direct Espressif supply agreements.

The key distinction: ESP32 modules are far cheaper than nRF52840 modules ($1.80–3.50 vs $4.50–9.00 at 1,000 units) because Espressif’s volume is far higher and the competitive pressure is intense. ESP32-C3 is a common choice for smart door locks, water-leak sensors, and mmWave presence sensors that need WiFi + BLE on a tight BOM. The risk is not counterfeit dies (Espressif is Chinese, so the supply chain is shorter) but rather BOM substitution: a manufacturer quoting “ESP32 module” may deliver ESP32-C3 or ESP32-S2 or a generic clone without notifying you. These are not pin-compatible.

Which to choose: use nRF52840 when you need BLE 5 + Thread/Zigbee + Class 3 RF performance and can absorb the cost premium. Use ESP32-C3 or ESP32-C6 when you need WiFi + BLE at lowest cost, or when you need Matter/Thread without the nRF52840 premium. For pure BLE 5 with no WiFi requirement and budget under $3 per module, ESP32-C3 is the more defensible choice today. The BLE 5 module reference and the ESP32 module overview cover the variant differences in more depth. For the sourcing-specific risks of ESP32 modules — BOM substitution, variant pinning, and FCC modular approval — see ESP32 OEM in China.

FCC and CE certification: the numbers

BLE modules that transmit in the 2.4 GHz band require FCC Part 15 Subpart C certification in the US and CE marking under the EU Radio Equipment Directive. The relevant FCC rule is Part 15, Subpart C, Section 15.247, covering intentional radiators in the 902–928 MHz and 2400–2483.5 MHz bands.

Option 1: Use a pre-certified module

Modules from Laird, u-blox, Actinius (for nRF52840) and Espressif, Ai-Thinker, LILYGO (for ESP32) hold existing FCC grants. You can verify any module’s FCC ID at apps.fcc.gov. If the module’s integration instructions are followed exactly (antenna placement, keep-out zones, power limits), your end product can use the module’s FCC grant for the radio and needs only to pass unintentional radiator testing (FCC Part 15B) for the complete device.

Cost: $1,500–3,000 for Part 15B testing. Timeline: 3–4 weeks.

Option 2: Custom RF design

A custom RF design — proprietary PCB antenna, custom matching network — requires full FCC certification: RF power output, spurious emissions, and in-band emissions across all channels. Cost: $8,000–15,000. Timeline: 8–12 weeks. This path makes sense at 50,000+ units/year where per-unit module cost savings justify the certification investment.

For first products under 10,000 units, Option 1 almost always wins on total cost including engineering time.

CE RED for BLE modules mirrors FCC in the important ways: pre-certified modules (many European modules hold CE marking under EN 300 328) allow the same shortcut. Cost for an end-product CE RED self-declaration using a pre-certified module: €1,500–3,000. Timeline: 3–5 weeks. The FCC certification basics explain the modular-grant mechanism, and if you ship the same device into multiple regions the multi-market certification guide maps out which test reports carry over and which need a fresh submission.

Supply chain stability risks

The most underestimated risk in BLE module sourcing is long-term supply continuity. Choosing a module in 2026 for a product with a 5-year lifecycle means that module must still be available (or a pin/footprint-compatible replacement must exist) in 2031.

Nordic nRF52840: Nordic has a stated 10-year longevity commitment for the nRF52840. The chip has been in production since 2018 and is on a mature node. Risk is low, but module manufacturers (third parties assembling the nRF52840 into boards) can discontinue without that same commitment. Stick to modules from Laird, u-blox, or Actinius — all three have product longevity programs — rather than unnamed Shenzhen module assemblers.

Espressif ESP32 variants: Espressif frequently introduces new variants and has a history of silently introducing hardware revisions (e.g., the ESP32 v3.1 chip revision vs. earlier revisions) that affect RF performance and driver behavior. Pin out compatibility across variants is incomplete. Document which GPIOs are used and flag any that are module-specific; the same discipline applies when pairing a module with JST connectors or pin headers for external sensors. Specify the exact ESP32 variant, the exact module manufacturer (Ai-Thinker, not “ESP32 module”), and the exact firmware SDK version in your BOM. Pin all three in your purchase order.

BOM lock-in risk: specifying a module by function (“BLE 5 module, FCC certified”) without naming the exact part creates a loophole for substitution. We have seen this in practice: a factory delivering ESP32-C3 modules instead of ESP32 modules because “both have BLE 5 and FCC.” Technically accurate, but the GPIO count and WiFi power specs differ. Lock to the exact part number in writing.

What to verify when auditing a BLE module supplier

For suppliers assembling BLE modules (rather than just reselling Espressif or Nordic modules), the factory audit should cover these specific items:

Die authenticity verification: ask to see purchase orders showing acquisition of Nordic or Espressif dies from authorized distributors (Digi-Key, Mouser, Arrow, or authorized Chinese distributors). A factory that buys dies from spot market brokers is a factory that may receive counterfeit parts without knowing it.

RF test coverage: does the factory test RF output power and frequency error on 100% of units, or only on samples? For a BLE module, a failed RF test means a dead radio — the unit will appear to work at close range but fail at normal operating distances. 100% RF test adds $0.15–0.30 per unit and is worth it for any non-commodity application.

Firmware flashing verification: is the exact firmware version hash verified before packing? A factory that flashes firmware and packs without hash verification can ship units with wrong firmware. This is particularly damaging for BLE modules because firmware governs the advertising parameters and connection interval, which affect how your product interacts with host applications.

Red flag we have seen in practice: a Shenzhen module factory showed us their “FCC test station” — a laptop running a BLE scanner app. They described connecting each module and verifying it appeared in the scan as their FCC compliance test. A BLE scanner seeing an advertising packet is not FCC compliance testing. FCC compliance testing is conducted in a shielded chamber with calibrated antennas and a spectrum analyzer. This factory was failing their customers’ FCC compliance requirements without knowing it.

Power, antenna, and integration considerations

The module choice shapes the rest of your design. A decision made only on BOM cost often creates expensive rework later.

Sleep current and battery life. A BLE beacon or sensor running on a coin cell needs sleep current under 10 µA. The nRF52840 can reach 1.5–3 µA in System OFF mode with RAM retention; ESP32-C3 is higher, typically 5–10 µA in deep sleep. Over a two-year deployment, the difference can mean the difference between a 12-month and a 36-month battery replacement cycle. Lock the sleep-current target in your specification and verify it on the first samples.

Antenna integration. Pre-certified modules usually include a PCB trace antenna or a ceramic chip antenna with a defined keep-out zone. Moving the module closer than the keep-out distance to metal, battery, or other components degrades range and can void the modular grant. Include the module’s integration guide as a controlled document in your design review.

GPIO and peripheral mapping. nRF52840 and ESP32 variants have different pin counts, ADC channels, and available interfaces. A module swap mid-program because the original part is unavailable can force a PCB spin. Document which GPIOs are used and flag any that are module-specific.

Firmware SDK lock-in. Nordic uses the nRF5 SDK or nRF Connect SDK; Espressif uses ESP-IDF. The SDK choice affects your firmware team’s productivity and long-term maintenance. When evaluating a module, evaluate the vendor’s SDK release cadence and long-term support policy, not just the hardware price.

Documentation and traceability for long production runs

Once a module is qualified, the work shifts to making sure every production batch matches the qualified reference. Factories often rotate component lots, change die revisions, or update firmware without telling the buyer.

Controlled BOM. The purchase order should name the exact module manufacturer, exact part number, and approved firmware version. Any change requires written buyer approval and a new qualification cycle.

Date-code tracking. Request date codes for the radio die, flash, and crystal on each lot. If a field failure appears six months later, date codes let you narrow the problem to a specific production window instead of recalling every unit.

First-article inspection (FAI). On each new production run, require the factory to send 5–10 first-article units for approval before the rest of the lot ships. Compare RF performance, current consumption, and firmware hash against the golden sample.

Change notification agreement. Negotiate a formal change-notification clause: the factory must notify you 60 days before any process, material, or firmware change. Without this, a silent revision can invalidate your certifications or break host-device compatibility. We have seen modules pass initial qualification, then ship six months later with a different crystal load capacitance that drifted frequency error outside spec — a change the factory considered too minor to mention. For buyers running multi-year programs, our supplier management service enforces AVL and change-notification terms across reorders.

What we watch on real BLE module orders

Last quarter, our client ordered 5,000 nRF52840-based modules from a Shenzhen supplier for a wearable device — the same BLE radio that goes into an OEM fitness band. The factory quoted a competitive $4.20 per module and shipped samples that passed functional testing. We added the order to our incoming inspection plan before the rest of the lot moved to the client’s assembly house.

During the inspection, we audited the lot and inspected 200 units from the first shipment. Eight modules — 4% of the sample — failed RF output power testing, measuring +1 dBm instead of the specified +4 dBm. That difference does not matter on a desk, but it cuts usable range by roughly 30% in a wearable application. We also checked die authenticity documentation and found the factory’s purchase orders traced to an authorized distributor, so the dies were genuine. The problem was assembly variation in the matching network, not counterfeit silicon.

We rejected the lot and asked for a process review. The factory found that one of their two SMT lines had a stencil aperture issue on the RF inductor. After the line was recalibrated and a 100% RF test was added, the second shipment passed. The client avoided shipping 200 potentially short-range units per 5,000-unit batch.

For BLE modules specifically, we always verify three things before sample approval: die purchase orders from authorized distributors, 100% RF testing on the production line, and firmware hash verification before packing. Those three checks catch the failures that basic functional testing misses.

Sample qualification and first-article test plan

Before you commit to a 5,000-unit module order, run a structured qualification on 20–50 samples from the production lot.

RF performance. Measure TX power, frequency error, and sensitivity on a sample of units. A module that advertises +4 dBm but consistently outputs +1 dBm will have shorter range than your design assumes. Frequency error outside ±50 kHz can cause interoperability problems with some host devices.

Current consumption profile. Capture active TX, RX, and sleep current on multiple units. Outliers indicate assembly variation or counterfeit dies. We have seen lots where 10% of modules drew 40% more TX current than the rest; those units would have failed battery-life targets in the field.

Environmental stress. Run a subset of samples through a simple thermal cycle: 8 hours at 60 °C, 8 hours at -10 °C, repeated three times. BLE modules with marginal solder joints or moisture-sensitive components often fail after the second cycle.

Firmware hash verification. Flash the production firmware, read back the binary hash, and compare it to the approved release. This catches firmware mix-ups before they ship.

For buyers new to electronics sourcing in Shenzhen, the complete guide to sourcing electronics from China walks through supplier discovery, audit, and first-order management. For IoT module sourcing, the certification audit is the item most often skipped and most often regretted — running it as part of a pre-shipment inspection with 100% RF testing is the cheapest way to keep dead radios off the boat. The IoT modules industry page has more on what supplier qualification looks like in practice, and the Japanese distributor LoRa gateway case shows how the same die-authenticity and RF-test checks played out on a real radio-module order — see the LoRa module sourcing reference for the equivalent due diligence on sub-GHz radios.