Datalogging: What to Log and Why
A datalog records what sensors and the ECU reported while the car was being driven or run on a dyno, second by second. Reading one is how a tune, or a change to the car, gets checked against reality instead of assumed. This guide covers the channels worth logging, how to read a pull, and the signs of trouble to look for.
All 13 sections
The basics
What it is, what it changes, and whether you need it.
What a datalog is
A datalog is a recording, over time, of values the ECU and its sensors are reporting: engine speed, how hard the engine is working, the air-fuel mixture, ignition timing, boost pressure, and more. It is captured either from the ECU itself through a diagnostic connection or from a standalone logging device, and it can be reviewed afterward as a set of lines on a graph, one line per channel, all plotted against the same timeline.
It exists because a tune or a modification changes how the engine behaves, and the only way to know whether it is behaving as intended, rather than as hoped, is to look at what actually happened while the engine was running under load. A tune's tables are a set of intentions; a log is a record of what actually happened when those intentions met the real engine, the real fuel, and the real weather that day.
A log is also not a permanent judgment on a car; it is a snapshot of one drive, one pull, or one session. A car that logs cleanly today can develop a problem later, so logging is something many owners and tuners return to periodically, not a single box to check once and forget.
How a log is captured
Most modern cars can be logged through their factory diagnostic port, using software that requests specific data from the ECU over the same connection a mechanic's scan tool uses. This works for many stock and lightly modified cars without any extra hardware, though the exact channels available depend on what the factory ECU exposes.
A standalone or aftermarket ECU generally logs internally as part of its own software, since it is already managing every sensor input and output directly. Piggyback controllers usually include their own basic logging for the channels they touch. A standalone logger with its own extra sensors, wired directly into the engine bay, is used when a channel is not otherwise available, like a wideband oxygen sensor reading true air-fuel ratio rather than the narrower factory sensor's simplified output.
Whatever the source, a log is only as good as its sample rate, how many times per second it records each channel. A slow sample rate can miss a brief spike in knock or boost that a faster one would catch, which matters more for fast-changing channels like knock and less for slow-changing ones like coolant temperature.
Logging also uses some of the ECU's or logger's own processing capacity, so adding many channels at a high sample rate can occasionally affect performance on lower-powered logging hardware, or fill a memory card faster than expected on a long session. This is rarely a serious limitation on modern equipment, but it is worth checking a device's specifications before assuming every channel can be logged at the fastest possible rate simultaneously.
Why it matters even for a basic build
Even without chasing extra power, a log can catch problems early: a fuel trim drifting because of a developing vacuum leak, a boost target not being met, or a coolant temperature climbing higher than expected on a hot day. Catching this in a log, rather than after a part fails, is the practical value of logging even a mostly stock car.
A log is also how a shop or a tuner communicates what a tune is actually doing, rather than a claim about what it should do. Asking to see a log from a tuning session is a reasonable request before trusting a tune, and a tuner who cannot or will not provide one is worth being cautious about.
Logging is also useful outside of tuning entirely, as a general health check. A used car with an unknown history can be logged for a short drive to check that fuel trims, coolant temperature, and basic sensor readings look reasonable, which can surface a developing issue before it becomes an expensive one.
Reading a log for the first time
A first log can look overwhelming: a dozen or more lines moving at once across a shared timeline. It helps to start with just two or three channels, RPM, AFR, and boost if the car is turbocharged, and get comfortable with how they move together before adding more.
Most logging software lets you scrub through the timeline and see each channel's exact value at a given moment, which is more useful for understanding a specific event, like a knock correction, than trying to read the whole graph by eye. Zooming into the section of the pull where something looks unusual is generally more informative than looking at the whole log at once.
Choosing and fitting
How to pick one, what it needs alongside it, and the usual mistakes.
The channels worth logging
A useful log usually includes engine speed (RPM) and engine load, which together describe where on the engine's operating range a given moment sits, since fuel, timing, and boost targets all vary by RPM and load. Air-fuel ratio or lambda shows whether the mixture is where it is supposed to be, and ignition timing shows how much spark advance the ECU is actually using at that moment.
On boosted engines, boost pressure against its target shows whether the turbo or supercharger is delivering what the tune expects. Intake air temperature (IAT) matters because hotter intake air is more prone to knock and generally costs some power, so a log that shows IAT climbing on a hot lap or a long pull is worth watching.
- RPM and engine load
- AFR or lambda
- Ignition timing
- Knock activity or feedback knock
- Boost pressure vs. target (boosted engines)
- Intake air temperature (IAT)
- Short-term and long-term fuel trims
- Injector duty cycle
Fuel trims and injector duty
Fuel trims are the ECU's own correction to the fuel it is commanding, based on feedback from the oxygen sensor, expressed as a percentage above or below the baseline calculation. Trims that sit far from zero, especially at idle or cruise, usually point to something the ECU is compensating for, like a vacuum leak, a dirty sensor, or an injector delivering slightly more or less fuel than commanded, rather than a problem with the tune's wide-open-throttle tables.
Injector duty cycle is the percentage of time the injector spends open during each engine cycle. As duty cycle approaches 100 percent, the injector is close to its maximum flow and has no more room to add fuel if the engine calls for it. See Injector duty in the glossary and How to Size Fuel Injectors for how this connects to choosing injectors sized for a given power target.
Short-term fuel trim reacts quickly, adjusting within seconds to what the oxygen sensor reports; long-term fuel trim is a slower-moving average the ECU learns over many drive cycles. A short-term trim that spikes briefly and returns to normal is less concerning than a long-term trim that has settled far from zero, since the second suggests a persistent condition the ECU has already learned to compensate for.
Comparing logs to a baseline
A single log tells you what happened during one pull; a log compared against an earlier baseline tells you what changed. Keeping a log from before a part or tune change, on file and in the same format, makes it possible to see exactly which channel moved and by how much after the change.
This matters most for changes that are supposed to be invisible in normal driving but show up under load, like an injector swap or a new fuel pump. A baseline comparison can catch a fuel trim that has shifted even though the car feels the same to drive, well before it becomes a driveability complaint.
| Channel | What a meaningful shift suggests |
|---|---|
| Fuel trims at idle or cruise | A vacuum leak, sensor drift, or an injector delivering off-target |
| AFR at wide-open throttle | A fueling change: injector, pump, or a tune update |
| Boost curve shape | A change in turbo response, wastegate control, or a boost leak |
| Knock activity at a given RPM and load | Less margin than the previous tune had at that same point |
Setting up a logger correctly
A log is only as trustworthy as the sensors feeding it. A wideband oxygen sensor used for AFR logging needs to be positioned correctly in the exhaust, generally close enough to the engine to read accurately but not so close that exhaust pulses distort the reading, and it needs periodic calibration, since sensor accuracy drifts over its service life.
Logged values should also be checked once at idle against known-reasonable numbers before trusting them under load. An AFR reading that looks physically implausible at idle, far outside the range a healthy engine would show in closed loop, usually means a sensor or wiring problem rather than an actual engine condition, and it is worth resolving before using that channel to judge anything else.
Ground loops, loose connectors, and electrical interference from ignition components are common, mundane causes of noisy or unreliable data, especially on channels added with aftermarket sensors and wiring. A log with an oddly noisy or spiky channel is often a wiring issue rather than something happening in the engine, and it is worth ruling that out before reading too much into a single strange reading.
In depth
The numbers, the trade-offs and what goes wrong.
Reading a pull
A pull is a full run from low RPM to redline at wide-open throttle, logged so the whole range can be checked at once rather than just a single steady-state point. Reading one means looking at how each channel behaves across the whole RPM range: whether AFR holds near its target as load and RPM increase, whether boost builds smoothly to its target and holds it rather than spiking or tapering off, and whether ignition timing follows a sensible curve rather than jumping around.
It also means checking that no channel is misbehaving only in part of the range. A tune can look correct at peak RPM while running lean in the midrange, or hold boost well at low RPM while falling short near redline, and a log is what reveals that pattern instead of just the peak numbers.
Cross-referencing channels against each other is often more informative than reading any one channel alone. A dip in AFR that lines up exactly with a spike in boost points to a fueling system running short of capacity right at the moment it is asked for the most; the same AFR dip with no boost change points somewhere else, like a fuel pump or pressure regulator issue.
| Pattern in the log | Likely cause | What to check |
|---|---|---|
| AFR leans out as boost rises | Fuel delivery running short of demand | Injector duty cycle and fuel pressure under load |
| Boost overshoots target then settles | Wastegate control needs adjustment | Wastegate duty cycle and boost controller settings |
| Timing pulls repeatedly at one RPM | Knock margin too thin at that point | Fuel octane, IAT, and the tune's table at that cell |
| Fuel trims drift positive at idle | Unmetered air entering the intake | Vacuum lines, intake boots, and gasket seals |
Knock and feedback knock
Knock activity, sometimes logged as a knock sensor's raw signal or as "feedback knock," the timing correction the ECU applies in response to detected knock, is one of the most important channels on a log because sustained knock is what causes engine damage from a tune running too aggressively for the fuel and hardware in use. See What Does an ECU Tune Do? for how knock relates to fuel, timing, and boost targets.
Occasional, brief knock correction that the ECU catches and pulls timing for is generally treated differently from repeated or sustained knock at the same RPM and load point across multiple pulls, which suggests the tune's margin is too thin at that point rather than a one-off event.
Knock is also not always audible over engine and road noise, and background mechanical noise, a loose heat shield, valvetrain clatter, can sometimes be mistaken for it or mask it in the other direction. This is one of the main reasons tuners rely on the logged knock channel rather than listening alone, especially at higher RPM where engine noise is loudest.
Different platforms report knock differently: some log a simple count of knock events, others log the actual timing correction applied in real degrees, and some report a raw sensor voltage that needs interpretation against a baseline for that specific engine. Knowing which format a given car uses matters before comparing a knock reading against a rule of thumb learned on a different platform.
Red flags to watch for
A handful of patterns in a log are worth treating as a reason to stop and investigate rather than continue: AFR trending noticeably leaner than target as boost builds, boost pressure spiking well above its target before settling, repeated knock correction at the same point in a pull, injector duty cycle sitting near its maximum with no headroom left, or IAT climbing steadily across back-to-back pulls without time to cool.
None of these guarantee damage has occurred, but each one describes the engine running with less margin than intended, which is the condition that precedes damage rather than damage itself. See WHP vs Crank HP: Reading Dyno Numbers for how to read the power side of a dyno session alongside these channels, and start a build to see how a build sheet checks supporting parts against a power target.
- AFR trending leaner than target as boost or RPM climbs
- Boost spiking well above target before settling back down
- Repeated knock correction at the same RPM and load point across pulls
- Injector duty cycle sitting near its maximum with no headroom
- Intake air temperature climbing steadily across back-to-back pulls
- Coolant or oil temperature trending upward without leveling off
Logging on the road versus on the dyno
A dyno pull is repeatable and controlled, but it is also short and happens under conditions, fixed airflow across the radiator and intercooler, a level surface, that do not always match how the car is actually driven. Logging on the road or track catches conditions a dyno session may not: a long uphill pull with less airflow, a hot-lap session with heat soak building over many minutes, or hard cornering loading the engine at an angle a dyno pull never sees.
Neither environment replaces the other. The dyno is where a tune is normally built and verified under controlled conditions; road or track logging is where it gets checked against the conditions it will actually be used in. A tune that looks clean on a single dyno pull but has not been logged under sustained real-world load is not fully verified yet.
Logging on public roads carries its own consideration: a wide-open-throttle pull long enough to be useful usually means sustained high speed, which is illegal in most places outside a closed course or a legal racing surface. Many owners save full-throttle logging for a dyno, a track day, or another closed and legal setting, and use ordinary driving only to log lighter-load conditions like cruise and part-throttle acceleration.
Questions people ask
- Do I need special equipment to datalog my car?
- Many cars can be logged through their factory diagnostic port with appropriate software, and aftermarket ECUs and piggyback controllers usually include their own logging tools. Standalone loggers with extra sensors are used when a channel is not otherwise available.
- What is the single most important channel to watch?
- There is no single most important channel; AFR, knock, and boost against target all matter together, since a problem often shows up as a relationship between channels rather than one value alone.
- What does it mean if my fuel trims are high?
- High fuel trims mean the ECU is adding a noticeable amount of correction to reach its target air-fuel ratio, which often points to a vacuum leak, a sensor issue, or a fueling component not delivering exactly what is commanded.
- Is some knock activity normal?
- Brief, occasional knock correction that the ECU catches and adjusts for happens on some tunes and platforms. Repeated or sustained knock at the same point across multiple pulls is a sign the tune has too little margin there and is worth addressing.
- Can datalogging replace a dyno tune?
- No. A dyno gives a controlled, repeatable environment to tune in; a datalog is how the result is checked, both on the dyno and afterward on the road. They serve different purposes and are usually used together.
- What is the difference between short-term and long-term fuel trim?
- Short-term trim reacts within seconds to oxygen sensor feedback; long-term trim is a slower average the ECU learns over many drive cycles. A long-term trim that has settled far from zero suggests a persistent condition rather than a momentary one.
- Why does injector duty cycle matter in a log even if the car feels fine?
- Duty cycle nearing its maximum means there is no headroom left if the engine calls for more fuel, such as on a hot day or under harder use. A car can feel fine while already running close to that ceiling, which is why it is worth checking in a log rather than judging by feel alone.
- How fast does a logger need to sample to be useful?
- It depends on the channel: slow-changing values like coolant temperature are fine at a low sample rate, but fast-changing ones like knock or boost pressure need a faster rate to avoid missing a brief spike between samples.
- Can I trust a log if the sensors haven't been checked recently?
- Not fully. A wideband AFR sensor in particular drifts over its service life and needs periodic calibration, and a value that looks physically implausible at idle usually points to a sensor or wiring issue rather than a real engine condition.
- Is it legal to datalog a wide-open-throttle pull on public roads?
- Generally no, since a useful wide-open-throttle pull usually means sustained high speed, which is illegal on public roads in most places. Full-throttle logging is normally done on a dyno, at a track day, or on another closed and legal surface.
Read next
- What Does an ECU Tune Do?
An ECU tune changes the fuel, timing, boost and limiter tables an engine runs on. What changes, how it is delivered, and how tuners verify it worked.
- WHP vs Crank HP: Reading Dyno Numbers
Wheel horsepower and crank horsepower are not the same number, and dyno results are not comparable across dyno types. How to read a dyno sheet correctly.
- How to Size Fuel Injectors
What injector flow ratings mean, the formula for sizing a set, and why duty cycle, fuel pressure and dead time all matter.
- Stage 1, 2 and 3 Tunes Explained
Stage 1, 2 and 3 are informal shorthand for how much hardware a tune assumes, not a standard. What each stage usually means and how to verify one.
Reviewed 2026-09-24