Live telemetry

Real-time vibration.

Tach

Baseline —

RPM 1× tach

—

Phase — Period — Δ —

RMS vibration

—g

Peak — Crest — Temp —

Raw capture

idle

Record raw samples to PSRAM, then download as CSV. Useful for offline analysis in MATLAB / Python / Excel. Fewer axes or a lower sample rate → longer recording at the same PSRAM cost.

Axes

Sample rate — Hz (max 6660)

Duration — seconds (max 78)

Per-axis amplitude

Dominant — highlighted

X axis

—g

Y axis

—g

Z axis

—g

Time-domain waveform — 3 axis

RMS trend — last 120 s

Baseline —

FFT preview — 2,048 samples / 1,024 bins

Frequency domain

Spectrum.

Amplitude at each frequency. This reference capture uses 1,024 samples / 512 bins over 0–1,969 Hz; shipping hardware supports analysis to 3,330 Hz at the current 6.66 kHz sample rate. Peak at 1× RPM means imbalance; peaks at 2×/3× mean misalignment or looseness; broadband noise above 500 Hz usually means bearing wear.

Record FFT — seconds

Magnitude — linear

1× RPM — Peak — Amp —

Order amplitudes

1× imbalance · 2× misalignment · 3× looseness · 4× bearing/gear

1× RPM

—g

2× RPM

—g

3× RPM

—g

4× RPM

—g

Harmonic tracking

Each harmonic of running speed is tracked over time so a developing fault is visible before it dominates the spectrum. A trace ramping up without the others is a stronger signal than overall RMS rising.

Diagnosis hints

Synced to tracker

1×

ImbalanceNormal

Run the balance wizard.

—g

2×

Misalignment / loose mountNormal

Re-torque coupling and check shaft alignment.

—g

3×

LoosenessNormal

Inspect mount fasteners.

—g

4×

Bearing wear (early)Normal

Within normal — monitor.

—g

Angular rate

Torsional vibration.

Avg rotation angular-rate RMS, 0–833 Hz

—°/s

Overall torsional level across the band

Vibration RMS linear accel

—g

Broadband accelerometer level — for comparison

Gyro spectrum — 0–833 Hz

Dominant — Peak — Rate —

Torsional / angular-rate vibration from the gyro (°/s), 1.66 kHz sample rate. Suspended during a balance or raw capture.

Time-domain waveform — angular rate

—

Raw capture — gyro

idle

Record raw angular-rate samples (°/s) to PSRAM, then download as CSV for offline torsional analysis. Fewer axes or a lower sample rate → longer recording at the same PSRAM cost.

Axes

Sample rate — Hz (max 1665)

Duration — seconds (max 313)

Single-plane balance

Stop guessing. Start measuring.

Mount the sensor so the axis arrow points radially — straight at or away from the spindle centerline. Not skewed, not axial. Mark a reference line on the spindle with a sharpie for the tach pickup. The device computes correction weight and angle using the influence coefficient method.

1

Initial

2

Trial weight

3

Correction

4

Verify

Step 1 — Initial run

Measure baseline

Bring the spindle to target RPM. Hold it steady. Pick how long to capture, then tap Measure — longer captures give better SNR. The 1× component is extracted via Goertzel at the measured RPM.

Current RPM

—

Live vibration

—g

Capture duration — seconds (max 78)

Target ISO 1940 grade

How it works

01Stick the magnetic sensor on the spindle housing. Apply reflective tape to the spindle.
02Measure initial vibration — amplitude A₀ and phase φ₀ at 1× RPM.
03Attach a known trial weight. Measure again.
04Device computes correction weight and angle. Place it. Verify.

Step 2 — Trial weight run

Place + measure

Drag the weight marker on the polar plot to where you placed the trial weight, or type exact values below. Start the spindle at the same RPM, then measure.

Trial weight — grams

Trial angle — degrees

Current RPM

—

Vibration

—g

Trial position

VIEW 0°Drag to mark

Drag to mark the weight · rotate the view with ⤹ ⤸ above · 1.0 g @ 0°

Correction weight

—g

Remove if heavy spot · Add if light spot

Correction angle

—°

Measured CCW from tach trigger

Correction vector

1× RPM · live phase

Predicted outcome

Theoretical residual

—g

Estimated quality

—

Predicted —

Estimate only — not a calibrated ISO 1940-1 measurement. Verify with traceable equipment for certification.

Place the correction weight at the indicated position. Re-run the spindle, then tap Verify to confirm and refine.

Placement helper — from available weights ▾

Step 4 — Measure post-correction

Confirm before/after

Place the calculated correction weight on the rotor, then click Measure to capture the new vibration and see the before/after improvement.

Calculated correction

—g @ —°

Residual (live)

—g

Improvement

—

Effective placement: —

Balance run archive

History.

Saved runs

—

Avg reduction

—%

Last run

—

Reduction trend

Last 50 runs

Runs

—

Uptime / date	Label	RPM	Pre (g)	Final (g)	Reduction	Correction

Continuous record

Log & history.

Samples

—

Current RMS

—g

Baseline RMS

g

Interval

s

total record window — —

Uptime

—

RMS history — last 24 h

Alert at +50% baseline

Condition monitoring

Vibration analyzer.

Tach-free condition metrics from the live accelerometer: ISO 10816 velocity severity, a spectral running-speed estimate, and envelope/demodulation bearing-fault screening. Spin the machine, then Analyze.

ISO 10816 zone

—

— mm/s

Est. speed tachless 1×

—RPM

Broadband

—g RMS

crest —

Bearing impulsiveness env kurtosis

—

band —

IMU temp

—°C

axis —

Envelope spectrum — demodulated bearing-rate peaks

low-frequency peaks = candidate defect rates

Run an analysis to see demodulated peaks.

Torsional vibration gyroscope live

Angular-rate vibration the accelerometer can't see. The gyro runs continuously, so this reads live with the rest of the analyzer — it's suspended only during a balance or raw capture. Housing-mounted, this is torsional vibration (not absolute shaft RPM, which needs an on-rotor mount).

Torsional level AC RMS

—°/s

Dominant freq

—Hz

Sample rate

—Hz

axis —

Reading angular-vibration spectrum…

Multi-device dashboard

Fleet.

Live status for every GV-device on the same network. Enter a serial (e.g. GV-XXXX — printed on the device label) or a hostname / IP to add one — it's resolved via mDNS (gv-xxxx.local). The list and polling state are saved locally in this browser.

Add device

Poll every s

Devices

0 devices

No devices yet — add one above.

Device configuration

Settings.

Network

Currently connected

SSID —

IP —

RSSI —

WiFi network

Password

AP IP address

Security

AP password (≥ 8 chars; leave blank = open)

—

Applies immediately — existing AP clients will be disconnected

Static IP

Use static IP for WiFi (instead of DHCP)

Device

Device name

Accelerometer range

Device mode

Switch between vibration analyzer (logger + spectrum only) and balancer (adds the balance wizard). Persists across factory resets. Requires a reboot to take effect.

Danger zone

Reset all settings, clear NVS, or return the device to factory defaults. Cannot be undone.

Report a bug

Tell us what went wrong. We attach a diagnostic snapshot (firmware version, WiFi, memory, sensor & FIFO health, recent events) and the last crash dump if there is one. No vibration data or recordings are included.

What happened?

Include last-crash dump (if present)

—

Tachometer

Live ADC —

Triggered NO

Trigger edge

Trigger threshold — mVsignal level that registers a tape pass — calibration sets this automatically

Last calibrated

Never

Auto-tune threshold

Sweeps for the widest reliable band at your balancing RPM

Run the spindle at the speed you'll balance at, then start. It samples a few hundred revolutions at each threshold — so duration scales with RPM — and walks outward until detection degrades, then centers in the reliable band. The RPM readout pauses while it tunes. Uses the edge direction from your existing calibration.

Firmware

Serial—

Version—

Target—

Free heap—

internal / PSRAM—

low watermark—

CPU — core 0 / 1—

Uptime—

Updates

Latest available—

Last checked—

Check stats—

Firmware .bin (OTA)

Notifications

Webhook (ntfy.sh)

Alert thresholdpercent above baseline

Minimum interval between notificationsminutes — throttle repeat alerts