Open Tools · Loopsheet
Loopsheet user guide
Use Loopsheet for 4–20 mA scaling, series loop voltage budgets, thermocouple conversion, custom RTD and sensor tables: inputs, units, files and stated limits. To use the tool itself, open Loopsheet.
Professional use. Results are for education and preliminary design. They must be checked by a qualified professional, and they are not a substitute for code-compliant design or licensed engineering judgement.
Loopsheet calculates instrument signal mappings, series-loop voltage margins and sensor conversions. Keep a tag, service, process-unit label and notes with each record. Calculations stay in your browser; the only service exchange is the site run-token check.
Getting started
Open the tool, pick one of the example records or choose Add record, then pick its calculation kind: signal scaling, loop voltage budget, thermocouple, custom RTD, sensor table or point check. Enter values in the units shown beside each field, then choose Calculate. The result panel shows the answer, the assumptions it rests on and any diagnostic. Editing any input marks the result as out of date until you calculate again.
For a simple signal check, select linear scaling, enter process limits 0 and 100 and process value 50: the current is 0.012 A, or 12 mA. The Loop sheet tab lists every record with its tag, service and result as a printable page. Save the document as JSON to keep your inputs; export the current results as CSV for a table you can open in a spreadsheet. No live hardware is read or controlled.
Calculations and conventions
The editor takes currents in milliamps (mA), thermocouple voltages in millivolts (mV), supply and drop voltages in volts, resistances in ohms and temperatures in degrees Celsius on the reference scale. Saved files and exports use the canonical units: amperes and volts, so 12.1 mA in the editor is 0.0121 A in a file, and 12 mV is 0.012 V. Numbers are kept exactly as typed up to about 15 significant digits; longer entries are refused rather than rounded. The process-unit label describes your values; it does not perform a conversion. Use consistent process units for limits, entered points and tables.
Signal scaling: linear mode maps 4–20 mA to the two entered process endpoints. Reversed process limits work. Linear values outside the range are extrapolated with a diagnostic, never interpreted as device alarm thresholds. Equal endpoints are invalid. Square-root mode means process fraction equals the square root of current fraction; it does not apply extraction twice. Square-root values outside the configured range are refused.
Loop voltage budget: enter supply voltage, minimum transmitter voltage, fixed device drop, design current and each series resistance. Include both outgoing and return cable resistance, plus the receiver/burden resistance. Use the worst required operating or alarm current from your device data. “Sufficient” means the entered supply covers the entered DC series budget; it establishes no installation safety or code verdict. A negative allowable resistance means even a zero-resistance loop lacks voltage. Empty resistance lists are valid.
Thermocouples: types J, K and T use government reference functions. J supports -210..1200 C, K -270..1372 C, T -270..400 C. Both junctions must be inside that type's interval. Forward conversion subtracts the cold-junction reference voltage from the hot-junction reference voltage. Inverse conversion adds the cold-junction voltage before finding temperature. Reverse the two junctions to reverse the measured voltage. Inputs never extrapolate. Coefficient rounding at polynomial joins can leave tiny gaps; a bracket/residual failure is reported rather than a fabricated temperature. These ideal reference functions do not model your sensor's drift or accuracy.
Custom RTD: enter R0, A, B, C, a validity interval and a source label. Factors and tolerance classes from copyrighted standards or manufacturer tables are USER-ENTERED; use data you have rights to use. There is no bundled standard RTD table. The supported polynomial has A>0, B<=0, C<=0; the entered interval must have positive resistance and slope. Below zero it includes the fourth-order C term; above zero it omits that term. Two-wire mode adds your total lead resistance. Four-wire mode is ideal and requires entered lead resistance zero. Three-wire compensation, self-heating, excitation accuracy and lead mismatch are outside this model.
Sensor table: enter at least two knots in increasing process order. Signal values must strictly increase or strictly decrease throughout. Add the source and use a consistent signal unit. The lookup interpolates one straight segment at a time; it never extrapolates or smooths your data. Inverse lookup is available because monotonicity makes it single-valued.
Point checks: enter process values, observed currents in mA and an absolute current tolerance in mA. The result reports expected current, observed-minus-expected error, maximum absolute error and the count outside the entered tolerance. Equality is included. Empty lists have zero points and zero maximum error. This is a numerical point comparison, not a calibration certificate or uncertainty assessment.
Results, diagnostics and resource limits
Results expose numeric metrics, categorical flags, per-point rows and diagnostics. A bad record in an otherwise valid document keeps a diagnostic; the summary discloses rejected and completed record counts. A structurally invalid document is refused before calculation. No “success” numeric result contains a non-finite value. Inverses expose iterations, residual and bracket width. An exhausted iteration budget returns NOT_CONVERGED with no temperature result. A measured floating-point plateau wider than 1e-8 C also returns NOT_CONVERGED with its measured width and no temperature result, even when the residual is zero. Such a record is counted as rejected, and CSV preserves the diagnostic and width.
The engine supports 200 records, 1,001 points per table/check record, 20,000 aggregate points, and 128 series resistances per loop. JSON files are limited to 2 MiB UTF-8 and CSV exports to 8 MiB. The conservative calculation memory budget is 64 MiB. These are implementation caps, not measured performance guarantees. An inverse permits at most 64 iterations and targets a bracket no wider than 1e-8 C together with a reference-scaled residual check. A small iteration budget can legitimately refuse an otherwise valid inverse.
Generic nonzero scalar magnitudes must be at least 1e-100 and at most 1e100. Additional per-field limits are defined by the engine constants and input validator: DC voltage at most 1e6 V, resistance at most 1e12 ohm, positive design current 1e-12..0.1 A. Custom RTD input windows are -250..1000 C, R0 0.001..1e6 ohm, coefficient magnitude at most 0.1; the actual usable interval must come from your source. Normalised signal fractions are bounded by 1e6. Very large or very small values outside these windows are refused. The engine does not round small results to zero.
Files and exports
Save/Open uses a documented JSON object with format loopsheet, version 1, title, copyright, terms, and records. Each record stores id, tag, service, unit, notes and its complete task, including inactive inverse-budget fields. Record ids must be unique. Unknown/missing fields, unsupported versions, invalid numeric types and oversized files are refused. The engine never substitutes defaults during Open. Tokens and results are not saved. Keep a separate saved copy before changing a record.
CSV uses seven columns: recordId, tag, service, unit, section, metric, value. Inputs include the task JSON, so assumptions accompany result rows. Values with A, V, Ohm, W or C suffixes use those canonical units; dimensionless fractions and counts have descriptive names. Spreadsheet formula-like user text is neutralised, and text line breaks become spaces so records remain rectangular. JSON preserves your original text. Numeric CSV literals preserve the engine value. CSV is an export, not a round-trip project format. A result snapshot must match the current inputs; otherwise export asks you to rerun. No proprietary file format is read or written.
Every export carries the tool attribution and professional-use notice. Opening, saving and editing inputs remain possible without a run token. Numerical runs and Verify require the site's signed token. If token verification fails, the message explains that the tool runs on simulations4all.com. Cancel stops the running job; late replies from a cancelled or superseded run are discarded, and a fresh worker is started for the next run.
Limitations
Loopsheet is a calculator for individual records, not a design package. It does not read or control hardware, model sensor drift, accuracy classes, self-heating, three-wire compensation or lead mismatch, apply standard tolerance tables, produce traceable calibration certificates or uncertainty budgets, or solve general circuits. The voltage budget is a static DC series sum. Thermocouple conversion is limited to types J, K and T over their reference ranges. Anything outside what you enter, or outside the stated ranges, is refused or labelled rather than guessed.
How we verify
The Verify page runs the tool's reference cases in your browser, using the same inputs, references and finite-tolerance evaluator as the automated test suite. Cases identify their basis and reference family. Hand-derived scaling examples, Kirchhoff voltage/energy balances, independent derivative integration for the custom RTD, and determinant-based table interpolation check the implementation. Published thermocouple checks use separately printed fixed-point values from NIST Monograph 175, rather than recomputing the coefficient polynomial as its own oracle.
Reference-table checks are mathematical verification. Field validation and a published full-loop worked example remain pending source with no reference numbers, and do not count as passes. Seeded property sweeps are reported by family; they do not create extra independent experimental datasets. The tests also probe finite arithmetic, unit/sign conventions, boundary regimes, exact-decimal decision boundaries, file persistence, CSV consumers, gate refusal, cancellation and maximum work sizes. This implementation's own tests do not replace an independent professional review or browser integration checks.
Compared with other tools
This is a compact local calculator and original record format. It does not provide an enterprise instrument database, wiring management, hardware communication, traceable certificates, statistical uncertainty, standard tolerance tables or a general circuit simulator. Use the related circuit tool for bridge circuits and the statistics tool for measurement-system analysis. The value here is that entered assumptions, equations, reference ranges and verification evidence remain inspectable alongside the result.
Troubleshooting and genuine questions
Why does my square-root input fail? Check whether the transmitter or receiver has already extracted the root. This mode only covers the configured range and applies the current-to-process root once.
Can I supply a standard RTD table? Yes, if you have rights to use it. Mark it USER-ENTERED and record the source. This engine does not reproduce standards tolerance tables.
Does positive loop margin establish safety? No. It checks only the entered static series voltage requirements. Device, cable, grounding, hazardous-area and installation requirements need separate qualified review.
Why is an inverse unavailable? The requested signal may be outside the reference interval, the RTD may not be monotone/positive, or the iteration budget may be too small. Inspect the typed diagnostic rather than treating an unavailable value as zero.
What happens offline? Editing and file handling remain available. Numerical runs rely on the shared gate's signed-token expiry and offline-grace policy; this tool supplies no permissive bypass.
Professional use
For education and preliminary design. Results must be checked by a qualified professional. Not a substitute for code-compliant design or licensed engineering judgement.