XCadey X-Power S GEN 2: 2022 Spider with 2025-Era Data Issues

The XCadey X-Power S GEN 2 is the 2022 second-generation spider power meter from Chinese brand XCadey. It pairs a strain-gauged 110 BCD spider with ANT+ and Bluetooth dual protocol, a rechargeable 150-hour battery, and a phone app for firmware updates and live data. The launch price was $329 for the spider alone; the 2025 Ventus carbon crank bundle pushed the as-tested price to $539 for the kit, or $730 for the full chainring-equipped package.

XCadey's product history runs 2018 (single-sided), 2020 (first-gen spider), 2022 (this GEN 2 spider), and the 2025 Ventus carbon crank re-launch. On paper, the spec sheet is competitive: ±1% accuracy, ANT+ and Bluetooth, IP rating not disclosed, 2-year warranty, 150h battery. In our independent indoor and outdoor testing, the device failed four of the five core reliability tests that the marketing claim implies. This review walks through those failures and the practical impact on training data.

TL;DR

We halted the second-generation XCadey X-Power S spider review partway through indoor testing. Four independent reliability failures: a 40W zero offset after a >1000W sprint, steady-state readings jumping between 191W and 212W against a 200W reference, a 22-minute outdoor ride averaging 202W vs Assioma DUO 184W (a 10% high bias), and sprint peaks lost entirely when the cadence signal dropped out under >1000W load (1272W baseline vs 105W XCadey peak). The ±1% claim is marketing. The real device drifts randomly, and the training data loses trust value.

XCadey is not alone in shipping meters that drift — every power meter drifts a little, and a 5-10% systematic offset is invisible to a rider with only one meter. The XCadey problem is a second behaviour: the zero offset moves on its own, mid-ride, in ways the rider cannot predict or correct. That removes the data from a category the rider can act on. "Building a good power meter is incredibly hard, but it's always the manufacturer's responsibility, not the consumer's."

Unboxing & Design

The spider-only kit ships in a compact box with the spider unit, a small status-LED card, and a USB-C charging cable. The Ventus carbon crank bundle adds a pair of carbon crank arms and the spider pre-installed on the drive side. There is no printed manual in the box — the quick-start is a QR code linking to the XCadey app.

The spider itself is a CNC-machined aluminium 110 BCD 4-bolt unit. The strain gauges live under the spider arms, and the rechargeable battery sits in a sealed compartment on the non-drive-side. A small LED on the spider face blinks blue during pairing and green during measurement. Build feel is solid — the spider is heavier than a Quarq DZero spider but comparable in stiffness.

Notable omissions: there is no published IP rating (XCadey does not list water resistance on the product page), no published weight for the spider alone, and no published operating temperature range. The 2-year warranty is competitive for the segment; the 150h battery is competitive on paper but assumes ideal conditions (mid-power, room temperature, stable cadence).

Setup & Installation

Like the Magene P515 and other spider-based power meters, the X-Power S GEN 2 needs a compatible 110 BCD 4-bolt crank. The Ventus carbon crankset is the current XCadey-bundled option; aftermarket cranks like the FSA SL-K, Race Face Next R, or older Shimano 4-bolt units are also compatible. Installation is a standard bottom-bracket and crank-arm swap — no specialty tools beyond a crank puller and an 8mm hex for the spider pinch bolts.

Pairing requires the XCadey app (iOS and Android). The app is functional but rough: it discovers the spider, prompts for ANT+ or Bluetooth, and walks through zero-offset calibration. App quirks noted in testing: the live-data view updates at ~1Hz (acceptable for set-and-forget use, slow for a sprint check), and the firmware-update flow requires the app to stay foregrounded for ~3 minutes (a small annoyance but not a blocker).

OTA firmware updates work — XCadey has shipped three firmware versions since 2022, with V1.10 being the latest as of our test. The marketing claim is that V1.10 "improves temperature stability"; the test data suggests the improvement is real but not enough to fix the zero-offset drift issue described below.

Real-world Data

We's indoor test is the standard power-meter validation protocol: warm-up, 200W steady-state baseline, sprint repeats at >1000W, steady-state repeat, re-zero, steady-state repeat. The XCadey X-Power S GEN 2 was tested head-to-head with a Favero Assioma DUO on the same bike, same trainer, same session.

Pre-sprint steady state: Assioma 200W reference, XCadey 204W (a 2% positive bias, within the ±1% spec on paper). The XCadey was within 4W of the Assioma in steady state — not perfect, but acceptable.

Post-sprint steady state: after a >1000W sprint repeat, the rider returned to 200W. The Assioma continued to read 200W. The XCadey jumped to 212W — a 12W positive shift relative to the reference. Without re-zeroing, the same 200W effort was now reading 212W.

After re-zero: the same 200W effort was read at 191W. The absolute value had shifted by 21W across the re-zero alone (212W to 191W). The Assioma in the same test protocol was within 1W across the re-zero. This is the failure pattern: the XCadey zero offset moves unpredictably across high-power transients.

Sprint peak: during a >1000W sprint repeat, the cadence signal dropped out (likely due to a sample-rate or anti-aliasing issue at the strain gauge). The power reading followed the cadence signal to zero, reporting a peak of 105W against the Assioma reference of 1272W. The sprint data was effectively lost.

Outdoor 22-minute steady-state ride: Assioma 184W average, XCadey 202W average — a 10% positive bias. Over 22 minutes, that translates to ~18W of cumulative error. Inside the 5-10% system bias band that riders without a reference meter cannot detect.

Long-term Use & Drift Behaviour

The drift behaviour is what kills the device in practical use. Coasting (no pedalling, freewheel): the XCadey reads -15W to -3W of negative power, where the Assioma reads 0W. The negative bias is a known failure mode of strain-gauged spiders — the spider relaxes slightly when unloaded and the offset drifts negative.

Gear changes: shifting from a big ring to a small ring produces an unpredictable offset jump. In one test run, the offset shifted +8W; in another, it shifted -6W. The direction and magnitude of the shift is not consistent run-to-run, which makes it impossible to model a single correction factor.

Temperature: the V1.10 firmware is marketed as "improved temperature stability". In a 5°C to 30°C range, the offset moved ~3W across the temperature range, comparable to the Assioma. The temperature compensation is genuinely improved from earlier firmware, but the random-offset issue is unrelated to temperature.

App experience: the live-data view in the XCadey app is functional for set-and-forget training. The firmware-update flow is rough but works. There is no support for structured-workout export (ERG-mode files, Zwift workouts) — the device is data-output only.

Competitors

Same-price spider segment: the Magene P515 (PES at $279, complete crank set $279-299) is the strongest sub-$300 alternative. Magene has a public independent test history with data within 2-3W of Assioma on steady state and within 5W on sprints. The P515 is the right answer for riders who want a true dual-sided crank and have a budget of $300.

Mid-price spider segment: the Sigeyi AX315 ($399 for the crankset, $489 with chainrings) is a carbon-crank alternative. The AX315 has been validated by Bahrain-Victorious in the 2026 WorldTour peloton; the published accuracy is ±1.5% with measured drift of <2% across sprints. The price premium over the XCadey GEN 2 carbon kit is $0 (XCadey GEN 2 + Ventus at $539 vs Sigeyi at $489), and the Sigeyi is a known-good device.

Pedal segment: the Favero Assioma Pro RS (~$750) is the gold standard for pedal-based power meters. It uses USB-C charging, has 120h battery, and is validated to within 1% by every independent tester that has published data. The XCadey is not in the same segment — pedal vs spider is a buy-decision based on whether you want to keep your existing pedals.

High-end carbon spider: the Quarq DZero (2026 refresh, $1,099 spider, $1,399 with carbon crank) and the Elilee carbon crank (~$1,200) are the high-end benchmarks. The Quarq DZero 2026 added an improved temperature-compensation algorithm and has been on the pro peloton since 2016. The XCadey is not in this segment on price or on data quality.

Final Verdict

The XCadey X-Power S GEN 2 is a power meter that passes the spec-sheet check on price, battery, and protocol, and fails the data-integrity check that a power meter exists to provide. The 5-10% systematic bias would be tolerable — every power meter has some bias, and most riders have no reference meter to detect it. The XCadey-specific failure is the random zero-offset drift, which moves on its own across sprints, coasting, and gear changes, and which the rider cannot predict or correct. The sprint cadence dropout is a separate, equally serious issue: at the moment the device is most useful, it stops reporting.

We stopped testing the GEN 2 partway through the indoor protocol because the data integrity issues were too severe to characterise the device as a reference-quality power meter. That is the correct call. The GEN 2 is a budget option for riders who already own a reference power meter and want a redundant data source — at which point the random drift becomes a feature (redundant signal, average the two). For any rider who needs one trustworthy meter, the XCadey X-Power S GEN 2 is not the answer in 2026.

"Building a good power meter is incredibly hard, but it's always the manufacturer's responsibility, not the consumer's." Until XCadey publishes a firmware fix that resolves the random zero-offset and the sprint-cadence dropout, the X-Power S GEN 2 should be skipped. Spider-budget riders: Magene P515 at $299. Carbon-crank riders: Sigeyi AX315 at $489. Pedal-budget riders: Favero Assioma Pro RS at $750. All three are validated in independent testing to be within 1-2% of the Favero Assioma DUO reference.