Data Acquisition Rate Calculator

The calculator converts your configuration into an effective sample stream, then sizes throughput and storage.

• Effective samples per second per channel: continuous: sps = sampling_rate  or  triggered: sps = trigger_rate × samples_per_event
• Raw bit rate: raw_bps = sps × channels × bits_per_sample
• Compression and overhead: payload_bps = raw_bps ÷ compression_ratio, peak_bps = payload_bps × (1 + overhead%/100)
• Duty cycle and storage: avg_bps = peak_bps × duty_cycle%/100, total_bytes = (avg_bps ÷ 8) × duration_seconds, recommended = total_bytes × (1 + margin%/100)

1. Select an acquisition mode based on your instrument workflow.
2. Enter sampling or trigger settings to define effective sampling.
3. Set channels and bits per sample to reflect resolution.
4. Add compression and overhead to match file format or transport.
5. Use duty cycle and duration for realistic run sizing.
6. Apply buffer margin to choose storage and bandwidth safely.

Example volumes are approximate and assume decimal units.

Sampling Throughput Sets the Ceiling

Start with the effective sample rate per channel. In continuous mode, a 10 kHz stream means 10,000 samples each second for every channel. In triggered mode, 50 events/s with 2,000 samples/event yields the same 100,000 samples/s per channel. This value sets data production before compression, overhead, and duty cycle.

Resolution and Channel Count Multiply Load

Raw bit rate scales linearly with both channel count and ADC resolution: raw_bps = sps × channels × bits. For example, 8 channels at 100 kS/s and 16 bits produce 12.8 Mb/s, or 1.6 MB/s. Doubling to 16 channels doubles the rate; moving from 16 to 24 bits increases payload by 50% at the same sampling.

Triggered Capture Changes the Average

Triggered systems often run below peak. If the duty cycle is 20%, a peak stream of 10 MB/s becomes 2 MB/s on average. Use duty cycle to represent real acquisition time, not operator time. For bursty experiments, compare peak rate to disk write speed, then use average rate to forecast long runs and data repository growth.

Compression and Overhead Shape Real Traffic

Compression ratio is modeled as raw ÷ compressed. A 2.0 ratio halves the payload. Add protocol and container overhead for timestamps, headers, and checksums; 10% overhead multiplies the compressed rate by 1.10. A realistic pipeline might be 20 MB/s raw → 10 MB/s compressed → 11 MB/s with overhead, before duty cycle applies.

Storage Sizing from Duration and Margin

Total volume is average_bytes_per_second × duration_seconds. At 3 MB/s average, a 2 hour run generates about 21.6 GB. Add a buffer margin, such as 15%, to cover file-system allocation, retries, and calibration blocks. The recommended figure is what you provision on the acquisition PC, then mirror to analysis storage.

Bandwidth, Integrity, and Practical Checks

For continuous sampling, approximate signal bandwidth with Nyquist: bandwidth ≈ sampling_rate/2. A 200 kHz sample rate supports 100 kHz content if filtering is applied. Validate that your link (USB, Ethernet, PCIe) exceeds peak rate with headroom, and CPU usage remains acceptable while writing. Re-run the calculator when channels, bit depth, or triggers change.

What does “effective samples per second” mean?

It is the sample stream used in rate math. Continuous equals the sampling rate. Triggered equals event rate times samples per event. Use it to predict throughput per channel.

When should I choose triggered capture?

Use it when you only need short windows around events. It reduces average storage while keeping high peak resolution, but you must still size links and disks for the burst rate.

How do I estimate overhead percentage?

Include packet headers, timestamps, framing, checksums, and file container metadata. If unsure, start with 10–15%, then validate by comparing a short recorded file size to the theoretical payload.

Does compression change peak requirements?

Compression reduces the payload rate, but peak requirements depend on the compressed stream plus overhead. If compression is CPU‑limited, real peaks may rise; benchmark your encoder at the target sampling rate.

What buffer margin should I use?

Typical margins are 10–25% for experiments and 25–40% for field logging. Margin covers file-system allocation, retries, calibration blocks, and unexpected duty-cycle changes.

How should I interpret the bandwidth output?

For continuous sampling, the bandwidth estimate uses Nyquist, roughly half the sampling rate. It is not a filter design tool; apply anti‑alias filtering and confirm instrument response for your frequency range.