Datalog Analyzer

Drop in an MHD, JB4, bootmod3, or EcuTek CSV log. It finds loaded pulls, plots channels, and screens actual values against logged or entered targets. Everything runs in your browser; the CSV never leaves your computer.

Updated August 11, 2026Log never leaves your browserThresholds explained in the reportMethodology

Load a Log

Drop your CSV here or click to browse MHD, JB4, bm3, and EcuTek exports work as-is. Any CSV with RPM + throttle columns will do.

How to Read Your Own Datalog

A log records selected ECU channels, but it is only as reliable as the channel definition, sensor scaling, sampling rate, and test procedure. This page organizes the data and performs transparent target comparisons; it does not approve a calibration or diagnose a component.

The Channels, in Plain English

Knock and Timing Corrections

There is no universal normal timing-correction number. Use this tool to locate and count events, then interpret them with the logger definition, factory strategy, calibration, fuel, and a qualified tuner. A correction channel cannot independently prove knock or identify a failed part.

Boost vs Target

The tool displays actual minus target and requires persistence before flagging a difference. The threshold is a screening rule, not a diagnosis; control strategy, throttle intervention, limits, sensor scaling, and operating conditions all affect the trace.

AFR Under Load

When a target is logged or entered, TunerBench reports sustained actual-minus-target delta lambda. Without that reference, it displays the measured value but assigns no pass/fail. Confirm whether the logger uses lambda, gasoline-equivalent AFR, or a fuel-specific AFR scale.

Heat: IAT and Heat Soak

Compare start, peak, and change across tests performed under the same ambient conditions and procedure. Sensor placement, ECU compensation, vehicle speed, heat soak, and test duration affect the trace; use platform-specific limits.

Trims and Fuel Pressure

The tool reports closed-loop trim averages and pressure minima. It only flags rail pressure when a commanded-pressure channel is available, and even then it reports a difference rather than naming a failed component.

Log Like You Mean It

Follow the vehicle, dyno operator, and tuner procedure. Perform loaded testing on a suitable chassis dyno or closed course with appropriate supervision, never where it risks public-road users. Keep gear, fuel, temperatures, configuration, and sampling consistent for comparisons.

Methodology and Limits

Channel names are normalized across supported CSV formats; pull detection and persistence thresholds are documented heuristics. A target-comparison flag is not a diagnosis, tune revision, or safety approval. Verify channel mapping and raw rows before relying on any result.

Frequently Asked Questions

What datalog formats does this analyzer support?

CSV exports from MHD, JB4, bootmod3, and EcuTek work out of the box, and most other tools (Cobb, ProTune, generic OBD loggers) work too as long as the CSV has an RPM column plus throttle or pedal position. Column names are matched loosely, so "Boost (psi)", "ecu_psi", and "Charge air pressure" all land in the right slot. The file is parsed in your browser and never uploaded.

How much knock retard is normal in an MHD log?

There is no universal normal value. Channel meaning, sign, filtering, ECU strategy, and acceptable behavior vary by platform and calibration. The tool reports magnitude, persistence, cylinder, and RPM for review against platform documentation or by a qualified calibrator.

What AFR should a turbo car run at wide open throttle?

Use the commanded target in the calibration or a value supplied by the tuner for the exact engine, fuel, load, and condition. TunerBench compares actual lambda with a logged or user-entered target. Without one, it displays the measurement and assigns no mixture pass/fail.

What do fuel trims tell me in a datalog?

Fuel trims are ECU corrections whose definition and valid interpretation depend on operating mode and platform. TunerBench reports a closed-loop average when it can identify light-load samples. It does not use a universal healthy range or infer a component failure from trims alone.

How do I tell a boost leak from a lean fuel system in a log?

A single log pattern cannot reliably distinguish a leak, sensor issue, calibration difference, or fuel-delivery problem. Compare actual values with commanded targets, confirm channel scaling, and use platform-specific diagnostic procedures. TunerBench reports correlations but does not assign a cause.

Why does my AFR read differently on E85 versus pump gas?

Some tools display gasoline-equivalent AFR and others use a fuel-specific scale. TunerBench attempts to infer the scale and normalize to lambda, but you must verify the logger and sensor configuration. Lambda comparison is only graded when a logged or entered target is available.

Is my datalog uploaded anywhere?

No. The CSV is parsed entirely in your browser and is not uploaded. When you share a report, only the compact screening summary is packed into the link; the raw rows stay on your machine.

Related Calculators