BH-Curve — Operator Help
Purpose. This tab computes and plots the magnetic hysteresis loop \(B(H)\) from two scope inputs: - Current through the winding (I) → used to compute \(H = \tfrac{N\,I}{l_e}\). - Voltage across the same winding (V) → integrated to flux \(\Phi = \int V\,dt\), then \(B = \tfrac{\Phi}{N\,A_e}\).
Accurate geometry (N, \(A_e\), \(l_e\)) and probe scaling (shunt/clamp) are mandatory. Deskew compensates instrument delays only—never to “tune” physics.
1) What you can set
Toolbar
- Acquire & Plot — capture current waveforms and draw B-H.
- Auto + Interval (s) — periodic capture; Interval = refresh period.
- T=1 + Avg cycles — extract exactly one electrical period and (optionally) average the last N periods to improve SNR.
- Ref for cycle detect: I / V / Auto — which trace to lock the period on (use the cleaner, more periodic one).
- 💾 Save PNG / Detailed CSV — export plot and per-sample data (t, V, I, H, B) for traceability.
- ⭮ Clear — clear current plot/history.
- Plot focus / Data — toggle control visibility and numeric diagnostics pane.
Core Geometry
- Turns N — turns of the sense winding used in both formulas.
- Ae (mm²) — effective cross-section area (entered in mm²; converted internally).
- le (mm) — effective magnetic path length (mm; converted internally).
Channels & Probe
- Current channel — scope channel carrying current.
- Voltage channel — scope channel across the same winding.
- Deskew Δt (V−I) [µs] — positive means V arrives later than I; the app delays I by +Δt to align instrumentation paths.
- Probe type & value
- Shunt: enter R (Ω); current is \(I = V/R\).
- Clamp: enter sensitivity (A/V); current is \(I = V·(A/V)\).
Sampling & Processing
- Mode: NORM / RAW — RAW can deliver higher point density; NORM is most robust.
- Pts — requested points per acquisition.
- Stop/Fetch — briefly stop the scope to fetch a stable frame, then resume.
- Remove DC / Detrend — suppress offsets/drift before integration (critical for \(\int V dt\)).
- Equal aspect / Tight fit / Overlay prev. — plot cosmetics and comparison.
2) Signal classes & recommended settings
Use these recipes depending on whether your waveforms are sinusoidal or pulsed/oscillatory. The Ref, T=1, Avg, and Mode choices are the big levers.
| Case | Typical waveforms | Ref (cycle detect) | T=1 | Avg cycles | Auto | Mode | Pts | Deskew Δt | Remove DC / Detrend |
|---|---|---|---|---|---|---|---|---|---|
| A. Sinusoidal (50/60 Hz or LF sine) | I, V ~ sinusoidal | I (cleaner zero-cross) | On | 4–16 | On (1–2 s) | NORM | 1–5k | 0–20 µs (as measured) | On / On |
| B. Steady PWM / square (kHz) | I trapezoid, V pulse & ringing | I (if shunt), else V | On | 8–64 (if period-stable) | On (0.2–1 s) | RAW | ≥10k | 0–2 µs typical | On / On |
| C. Burst / single-shot / flyback | One pulse + ring-down | V (edges are crisp) | Off | 1 | Off | RAW | Max | Use prior calibration | On / On |
| D. Aperiodic startup / sweep | Non-repeating | V or I (whichever has features) | Off | 1 | Off | RAW | High | Use prior calibration | On / On |
Rationale
• T=1 + Avg cycles improves SNR only for periodic/stable excitations.
• For pulses/transients, averaging smears physics → keep Avg = 1.
• RAW + high Pts captures fast edges/ringing needed for accurate \(\int V dt\).
3) Cycle detection guidance
- Choose Ref wisely. Use the cleaner, more periodic channel:
- With a shunt, current (I) is usually best for both sine and PWM.
- Voltage (V) has crisp edges; better for bursts, but excessive ringing can confuse period finding.
- What T=1 does. The app detects one period from Ref and resamples both I and V onto that window before computing H and B.
- Aperiodic/Transient. Turn T=1 Off. The app integrates over the whole captured window; ensure the window fully covers the event.
4) Deskew (Δt) for accuracy
- Definition in this app: Δt = (V − I). If Δt > 0, voltage arrives later than current → the app delays I to match V.
- How to set: excite the winding with a clean edge or a stable sine. Adjust Δt so simultaneous physical events align (edge onset, zero-cross). Do not “tune for pretty loops.”
- Why it matters: Any relative delay between I and V distorts the loop area (loss) and the apparent coercivity.
5) Tips per signal class
A) Sinusoidal
- Ref = I, T=1 On, Avg 4–16, Auto On.
- Keep Remove DC/Detrend On to avoid integrator drift from probe offset.
- With mains hum, prefer Stop/Fetch On for consistent headers.
B) Steady PWM / Square
- Prefer RAW and high Pts to resolve edges and ringing.
- Ref = I (if shunt), else V when I is noisy.
- T=1 On only if period is stable (constant frequency & duty). Otherwise set Avg=1.
C) Burst / Single-Shot / Flyback
- Set the scope to Single acquisition; trigger on the pulse edge.
- In the app: Auto Off, T=1 Off, click Acquire & Plot once.
- Use Detailed CSV for later verification.
- If the baseline drifts between runs, re-zero probes (Remove DC/Detrend stay On).
D) Aperiodic Startup / Sweep
- Auto Off, T=1 Off. Choose Ref with the most distinctive features to anchor the time window.
- Consider Equal aspect Off and Tight fit On to see the full loop excursion.
6) Measurement wiring and scope setup (must-do’s)
- Measure across the same winding for V that you use N for in the geometry. Use a differential probe or a truly floating scope input.
- Current sense:
- Shunt: 4-wire/Kelvin if possible; minimize loop area; know the exact R at operating temperature.
- Clamp: enter A/V correctly; disable built-in scope scaling (set the probe to 1× in the scope if the app handles scaling).
- Scope configuration: DC coupling, sufficient bandwidth (avoid 20 MHz limit for fast pulses), adequate vertical scale without clipping.
- Sampling: choose timebase so one or more complete periods (or the entire pulse) are in view. Increase Pts when using short timebases.
7) Validation & sanity checks
- Air-core check: with the core removed, slope \(\frac{B}{H}\) ≈ \(\mu_0\). Good for geometry sanity.
- Datasheet check: small-signal slope near H≈0 should be ≈ \(\mu_r\mu_0\). Saturation knee should appear at expected H.
- Repeatability: with the same settings, repeated runs should overlay within noise. If not, check trigger stability, deskew, and DC offsets.
- CSV traceability: keep Detailed CSV when publishing results. It documents (t, V, I, H, B) for independent re-analysis.
8) Troubleshooting quick list
- Loop drifting vertically → turn Remove DC and Detrend on; check probe zero.
- Area looks too big/small → verify Ae, le, N, and Deskew Δt; ensure V is across the correct winding.
- Noisy, fuzzy loop → increase Avg cycles (only if periodic), increase Pts, reduce bandwidth limiting, improve probe grounding.
- Cycle detection fails → switch Ref from I↔V, reduce noise, or turn T=1 Off for transients.
- Integration blow-up on pulses → window the event cleanly (trigger earlier), keep Detrend on.
9) Safety and data integrity
- Deskew compensates instrumentation delay only. Do not use it to “fix” physics.
- Never apply math scalings twice (scope probe settings vs. app). Decide where scaling lives.
- Use differential probes for high-voltage windings; obey their CAT ratings.
- Exports include instrument ID and geometry metadata; raw measurements are never altered beyond the stated processing (DC removal, detrend, resampling for T=1).
Formula summary
\(H = \frac{N\,I}{l_e} \quad B = \frac{1}{N\,A_e}\int V\,dt \)
with \(N\) in turns, \(A_e\) in m², \(l_e\) in m, I in A from shunt/clamp scaling, V in V across the same winding.
Version: 2025-08-10 20:01 UTC