Airmass Models and Migration

Capoworks tune format 20032 separates the physical air-mass models from the fuel, ignition and actuator tables that consume their calculated load. This page describes the calibration layout and the manual migration needed when coming from an older binary tune.

The standalone Speed Density, Alpha-N and MAF Air Charge strategies use different physical inputs and different model tables. Do not reuse one table as though it had the same axis in every strategy.

Model calibration tables

The final template defines these tables:

Strategy	Fuel menu table	Final size	Y axis	Cell value
Speed Density	VE	16 x 16	Effective MAP in kPa	VE in percent
Alpha-N	Alpha-N filling	16 x 16	TPS in percent	Cylinder filling in percent at the configured reference pressure
MAF Air Charge	MAF Trim	8 x 8	Pre-trim standardized cylinder filling in percent (100 × cylinder air mass / standard charge)	Air-mass correction in percent

The MAF trim table is intentionally not a 16 x 16 VE table in this layout. Its primary air mass comes from the MAF sensor; the trim table corrects that measured filling. The configuration template places RPM on the X axis and mafTrimLoadBins on the Y axis. That Y coordinate is the pre-trim standardized cylinder filling shown above, calculated from the MAF-derived cylinder air mass before the trim cell is applied; it is not MAF sensor output in percent. With two configured MAF sensors, the model sums two healthy sensors, uses twice one healthy sensor if the other is unavailable, and returns no valid model result if neither sensor is usable.

The later per-cylinder model fuel trims are separate from the tables above. They have a 16 x 16 Speed Density family using effective MAP and RPM, and a 16 x 16 Alpha-N family using TPS and RPM. Each family has one shared pair of axes and one trim table per configured cylinder. MAF, Lua and mock air-mass modes do not apply those model fuel trims.

What each model does

Speed Density calculates air mass from RPM, effective MAP, charge temperature, displacement and cylinder count, using the VE table. Effective MAP is the measured MAP when available, optionally estimated from the MAP Estimate table when measured MAP is unavailable or transient estimation is enabled.

Alpha-N calculates reference filling from RPM and TPS using Alpha-N filling, then applies one of two pressure policies. In standalone Alpha-N mode:

• With Standalone Alpha-N: multiply MAP enabled, the reference filling is multiplied by effective MAP divided by 101.325 kPa. This branch does not apply BARO compensation.
• With that option disabled, the reference filling uses 101.325 kPa. If Alpha-N barometric compensation is enabled, it is instead multiplied by ambient BARO divided by Alpha-N calibration reference pressure; if compensation is disabled, no BARO scaling is applied.
• An invalid BARO or calibration reference pressure falls back to 101.325 kPa and marks the model result as degraded.

SD + Alpha-N always uses the pure-reference Alpha-N branch described above, including its BARO policy; the standalone MAP-multiplication setting does not change that branch.

MAF Air Charge converts measured mass airflow into cylinder air mass and a standardized filling coordinate. It applies the MAF Trim percentage before publishing the corrected mass. Its standard-air-charge calculation uses 101.325 kPa and 20 C; the standalone MAF model does not expose the Air temperature source field used by Speed Density and Alpha-N.

Model settings in the final template

These labels appear in the final Capoworks TunerStudio definition:

UI label	Choices or units	Models where shown
Fuel strategy	Speed Density, MAF Air Charge, Alpha-N, Lua, SD + Alpha-N	Fuel settings
Air temperature source	Tcharge or IAT	Speed Density, Alpha-N and SD + Alpha-N
Use MAP estimate table	Enabled or disabled	Speed Density, Alpha-N and SD + Alpha-N
Standalone Alpha-N: multiply MAP	Enabled or disabled	Standalone Alpha-N
Alpha-N barometric compensation	Enabled or disabled	Standalone Alpha-N without MAP multiplication, and SD + Alpha-N
Alpha-N calibration reference pressure	kPa	When Alpha-N barometric compensation applies
Idle VE load source	Model default, Measured MAP, TPS or Effective MAP	When Use idle VE table is enabled

Use MAP estimate during transient is an additional acceleration-enrichment setting. It can make the MAP estimate active while measured MAP is still usable. The estimate table is indexed by RPM and TPS, and its cells are validated in the range 0 to 600 kPa.

If the selected charge-temperature source fails, the model first falls back to a valid IAT value. If IAT is also unusable, it uses 20 C. Both cases set the temperature-fallback diagnostic; a fallback is not reported as a healthy measurement.

Shared corrections

The four VE blend tables are common air-charge corrections. Each correction has its own blend parameter, bias curve, table and optional Y-axis override. The runtime evaluates these corrections from the same captured sensor frame as the physical model.

In SD + Alpha-N mode, the raw Speed Density and Alpha-N masses are combined first, and the common corrections are applied once to the combined mass. They are not applied separately to each branch before blending.

Downstream running-fuel corrections such as the CLT multiplier, IAT multiplier, barometric correction and after-start correction remain in the fuel calculation. The per-cylinder model fuel trim therefore factors modeled fuel, while those later corrections can still move the trimmed fuel quantity. Acceleration enrichment, the Lua fuel offset and fixed cranking fuel remain outside the model trim.

Fuel publication and validity

Each calculation captures its inputs and tune generation once, evaluates the physical model, applies corrections and prepares the complete fuel result. The runtime publishes the per-cylinder result and consumer coordinates together only if the calculation is still current. Stopping the engine, writing the tune or failing the current calculation closes new injection admission; already accepted output callbacks are allowed to finish. Cranking priming has its own gate and remains independent.

For a physical model, the following conditions apply only to inputs required by the active model and its enabled consumers. An input that is neither modeled nor selected for a consumer is not required. For example, MAF Air Charge does not require MAP or TPS merely because those sensors are configured. A fallback can satisfy a requirement only when the model or consumer defines a substitute for that same coordinate; a value from a different available input does not substitute for an unresolved requirement.

• RPM, displacement and cylinder count are finite and greater than zero.
• When the active model or a selected consumer requires measured or effective MAP, that coordinate is in the inclusive range 0 to 1000 kPa.
• When a MAP estimate is used to satisfy a required MAP coordinate, it is in the inclusive range 0 to 600 kPa.
• When the active model or a selected consumer requires TPS, it is inside the configured sensor-error tolerance, then normalized to the inclusive range 0 to 100 percent.
• When the active model or a selected consumer requires charge temperature, it is finite and greater than zero kelvin after the selected source or fallback is applied.
• When MAF Air Charge is active, at least one MAF source used by the model is finite and non-negative, and the interpolated MAF trim is finite and between 0 and 200 percent.
• Axes used by the active model and selected consumers are finite, inside their configured bounds and strictly increasing.
• Corrections and the final cylinder air mass are finite and non-negative.
• Base fuel and cycle fuel are finite and non-negative.

An invalid required model or fuel result does not publish stale fuel. The status changes to Faulted, and new injection is not admitted. A usable fallback, such as a valid MAP estimate or a usable charge-temperature substitute, can still publish fuel, but the status is Degraded and the model fault identifies the fallback. This distinction prevents a substitute coordinate from being treated as healthy sensor evidence.

Migrating a tune to format 20032

The final migration target is the 20032 layout. Older binary tunes are rejected and defaults are restored, so there is no safe universal automatic import of an old binary tune. Do not plan a migration through the intermediate 20026 or 20030 schemas.

Before changing firmware:

1. Keep the old firmware, its matching TunerStudio INI and the old tune together. Do not open the old tune in the final INI.
2. Export or screenshot every table you may need, including its X and Y axes. Record the active VE, Alpha-N filling behavior and MAF Trim corrections for the strategies you used.
3. Record the old per-cylinder fuel-trim tables and their RPM/load coordinates. The former 4 x 4 trims are not copied automatically.
4. Record the model settings, blend settings, table-load selectors and experimental ignition settings that differ from their shown defaults.
5. Save a copy of the old matching INI with the tune. The final firmware's generated INI is the reference for the new field names and units.

After installing the final firmware:

1. Start from the matching generated TunerStudio INI for that firmware build.
2. Select the intended Fuel strategy, then enter the final table axes before entering cell values.
3. Recreate the applicable model tables manually. Alpha-N cells formerly carried through the shared VE table belong in Alpha-N filling; verify MAF corrections against the old active VE values and the final MAF load axis.
4. Recreate the 4 x 4 fuel-trim intent in the appropriate Speed Density or Alpha-N 16 x 16 family. Do not assume that cell-for-cell copying preserves the old operating points.
5. Recheck the model, correction and consumer-load settings. Changing a table selector to a different unit requires its breakpoints to be retuned.
6. Confirm that the final result is not Degraded or Faulted, then validate the calibration from a known-safe operating state before normal use.

There is no source-backed rule that makes an old tune universally safe to import. The physical tables, their coordinates and the per-cylinder trims all changed, so migration requires observed old values and deliberate re-entry.

Source references

This page was checked against the Capoworks firmware source and TunerStudio configuration template at commit 8f4fcfb4. The dedicated model and fuel-publication work originated in commit fef0a82e, with later final-layout changes folded into the pinned revision.

• Final unreleased 20032 migration notes
• Model table and policy definitions
• TunerStudio table, menu and settings template
• Speed Density model
• Alpha-N model
• MAF model
• Captured inputs, corrections and load resolution
• Fuel publication and admission
• Cylinder-model fuel-trim and migration notes