Frequency Drift Monitor
This frequency drift monitor tracks frequency stability over long periods — from minutes to hours. Watch drift rate in Hz/minute and PPM/hour, lock onto a target frequency and observe frequency deviation in real time, set alert thresholds, and export timestamped data to CSV — all processed locally in your browser. As a passive signal stability tracker it runs continuously in the background without interrupting your workflow. For complementary snapshot analysis, the frequency stability analyzer computes Allan deviation and other short-term stability metrics from a burst of readings.
Frequency Drift Monitor Tool
| Time | Frequency (Hz) | Deviation (Hz) | PPM |
|---|
How to Use the Frequency Drift Monitor
-
Set Your Target Frequency
Enter the expected frequency you want to monitor (e.g., 1000 Hz for a calibration tone, 440 Hz for concert A, or 50/60 Hz for mains). Set the alert threshold to the maximum acceptable deviation in Hz.
-
Start Monitoring
Click "Start Monitoring" and grant microphone permission. Select your preferred microphone from the dropdown. The tool begins tracking frequency immediately and logging data points with timestamps.
-
Watch the Drift Chart
The main chart shows detected frequency over time with auto-scaling axes. Use the Time Scale selector to zoom between 1 minute and 1 hour views. The deviation chart below shows how far the signal drifts from your target.
-
Monitor Alerts and Statistics
The alert indicator turns red when deviation exceeds your threshold. The statistics panel shows mean, standard deviation, min, max, and overall PPM drift across the entire session.
-
Export Your Data
Click "Export CSV" to download a complete log with timestamps, frequency readings, deviation, PPM, and drift rate. Use this for documentation, calibration records, or further analysis.
Understanding Your Results
Current Frequency (Hz)
The large display shows the real-time detected frequency from your audio input. This is the peak frequency extracted via FFT analysis with parabolic interpolation for sub-Hz accuracy. The reading updates continuously while monitoring is active.
Deviation (Hz and PPM)
Deviation in Hz is the simple difference between the measured frequency and your target:
measured - target. PPM (parts per million) normalizes this deviation relative
to the target frequency: (measured - target) / target × 1,000,000. PPM is useful for
comparing stability across different frequencies.
Drift Rate (Hz/min)
The drift rate uses linear regression over the recent data window to calculate how fast the frequency is changing. A positive value means the frequency is rising; negative means it is falling. This helps identify systematic drift caused by temperature changes, component aging, power supply instability, or mechanical resonance shifts. Thermal drift is especially common in cheap ceramic oscillators and analog synthesizer VCOs, where the frequency can shift several Hz as the circuit warms up from cold start to operating temperature. To observe instantaneous frequency at a single moment without long-term logging, the real-time frequency detector is better suited.
Statistical Summary
The session statistics show the mean (average frequency), standard deviation (spread of readings), min/max (extremes), and overall PPM (mean deviation from target in parts per million). Lower standard deviation indicates better frequency stability.
Alert System
When the absolute deviation exceeds your configured threshold, the alert indicator turns red and an optional audio beep sounds. This is useful for unattended monitoring where you need to know immediately if the signal drifts out of specification.
Technical Background
Long-term frequency monitoring requires sustained accuracy over extended periods. This tool combines real-time FFT analysis with statistical methods to characterize both instantaneous frequency and long-term stability trends.
FFT Peak Tracking
The Web Audio API's AnalyserNode computes the FFT in real time. With an FFT size of 8,192 at
a 48 kHz sample rate, the raw frequency resolution is approximately 5.86 Hz per bin. Parabolic interpolation
around the peak bin refines this to sub-Hz accuracy, which is essential for detecting small drifts over time.
Linear Regression for Drift Rate
The drift rate is computed using least-squares linear regression on the time-frequency data within the visible window. The slope of the best-fit line gives the drift rate in Hz per minute. This method is robust against individual noisy readings because it considers the overall trend rather than just consecutive differences.
PPM Calculation
Parts per million (PPM) is a standard metric for expressing frequency accuracy and stability. It is computed
as (f_measured - f_target) / f_target × 106. For example, a 1 Hz deviation
at 1 MHz is 1 PPM, while a 1 Hz deviation at 1 kHz is 1,000 PPM. This normalization allows meaningful
comparison of stability across very different frequency ranges.
Standard Deviation and Stability
The standard deviation of frequency readings quantifies short-term stability (jitter). Lower values indicate a cleaner, more stable signal. For reference, a quartz crystal oscillator typically achieves stability better than 100 PPM, while MEMS oscillators are usually within 20–50 PPM. Temperature-compensated crystal oscillators (TCXOs) can achieve ±0.5–2 PPM, and oven-controlled crystal oscillators (OCXOs) can reach ±0.01 PPM — though measuring those accuracies requires professional test equipment, not a browser microphone.
Analyzing Recorded Audio Files
The file upload area lets you drop a WAV, MP3, FLAC, or OGG recording directly onto the tool. This is useful when you recorded a calibration tone earlier and now want to analyze its drift offline, without needing to re-run a live session. The tool decodes the file in the browser, runs the same FFT peak-tracking algorithm, and populates the drift chart and statistics as if the recording played in real time. To capture a reference tone to analyze later, use the tone generator alongside any screen-recording or DAW software.
Frequently Asked Questions
What is frequency drift and why does it matter?
Frequency drift is the gradual change in a signal's frequency over time. It matters in applications like radio communications, audio calibration, and electronic testing where frequency accuracy must be maintained within tight tolerances. Excessive drift can cause signal loss, detuning, or measurement errors.
How long can I monitor continuously?
The tool can run for many hours continuously. Data is stored in browser memory, so very long sessions (12+ hours at high logging rates) may eventually consume significant RAM. For extremely long monitoring, export CSV periodically and reset the session. The tool uses efficient data structures to minimize memory usage.
What does PPM mean and how is it calculated?
PPM (parts per million) expresses frequency deviation relative to the target frequency.
The formula is: (measured - target) / target × 1,000,000. For example, if your target
is 1000 Hz and you measure 1001 Hz, the deviation is +1000 PPM. PPM allows you to compare stability
across different frequency ranges on a common scale.
How accurate is the drift rate calculation?
The drift rate uses linear regression over the visible data window, which is more accurate than simple point-to-point comparison. Accuracy improves with more data points and longer monitoring periods. For best results, ensure a clean signal with minimal noise, and use a monitoring window of at least 5 minutes.
Can I use this to test oscillator stability?
Yes, this is an ideal use case. Feed the oscillator output into your microphone input (using appropriate signal conditioning), set the target to the nominal frequency, and monitor over time. The standard deviation and PPM readings will characterize the oscillator's short-term and long-term stability. Note that accuracy is limited by your sound card's sample rate and the FFT resolution.
Is my audio data private?
Absolutely. All frequency detection and analysis runs 100% in your browser using the Web Audio API. No audio data is ever recorded, transmitted, or stored on any server. The tool works completely offline once loaded.
Can I use this to monitor 50 Hz or 60 Hz mains power frequency?
Yes, with caveats. Set the target to 50 Hz or 60 Hz depending on your country's grid standard. To pick up mains hum, you'll need a signal path that actually captures it — such as a loop antenna near AC wiring or a direct line-in connection with appropriate conditioning. A standard laptop microphone will often pick up ambient mains hum incidentally, but the level and clarity vary widely by environment. For true grid-frequency monitoring, dedicated hardware meters are more reliable; this tool is best for educational demonstration or situations where a rough, free check of power frequency stability is sufficient.
What is a "good" PPM value and what do typical oscillators achieve?
PPM requirements depend heavily on the application. A basic RC oscillator may drift thousands of PPM with temperature. Ceramic resonators typically achieve 500–5,000 PPM. Quartz crystals reach 10–100 PPM. Temperature-compensated crystal oscillators (TCXOs) achieve ±0.5–2 PPM. For audio pitch stability, ±50–200 PPM (a few cents) is generally acceptable. For digital communications or time-keeping, you typically need <10 PPM. This tool's browser microphone measurement resolution is practical down to roughly 10–50 PPM for steady tones, depending on signal quality and FFT size.
How do I analyze frequency drift from a pre-recorded audio file?
Use the file upload area below the controls — drag and drop a WAV, MP3, OGG, or FLAC file onto it, or click "browse" to select one. The tool decodes and analyzes the recording entirely in your browser, populating the drift chart, deviation graph, and statistics without any microphone or live input. This is useful for reviewing a calibration tone you recorded earlier, or for post-session analysis of an instrument's tuning stability during a long performance. The same FFT peak-tracking and linear regression algorithms are applied as in live mode.
Related Tools
Frequency Detector
Full real-time frequency detection with gauge, pitch accuracy, and multi-peak separation.
Try it →Peak Frequency Detector
Find the loudest frequency in any sound with top-5 peaks and noise diagnosis.
Try it →Tone Generator
Generate precise test tones at any frequency for calibration and testing.
Try it →