PowerFC knock sensor thresholds explained

The PowerFC knock sensor provides an early warning of abnormal combustion noise, but it cannot replace a properly calibrated ignition and fuel tune. On an FD RX-7, the input can trigger ignition retard and, depending on the PowerFC version and configuration, a fuel correction when the signal exceeds the selected threshold. A copied setting is rarely appropriate because every engine, turbo system and sensor installation presents a different signal.

A threshold is meaningful only when compared with a measured baseline. Rotary-engine hardware, exhaust resonance, fuel quality, ambient temperature and wiring can all alter the signal. Start with conservative timing and fueling, record a baseline under controlled conditions, and change one threshold value at a time rather than treating the setting as a cure for knock.

Threshold condition	ECU interpretation	Likely ECU response	Principal risk
Lower threshold	More vibration qualifies as knock	Earlier and potentially stronger correction	False detection of turbo, injector or exhaust noise
Higher threshold	Only a stronger signal qualifies	Later or gentler correction	Genuine knock may pass without correction
Stable baseline	Repeated signal pattern under similar conditions	Predictable, comparable corrections	Hidden changes may be mistaken for sensor behaviour

How the PowerFC knock input works

A knock sensor converts mechanical vibration into an electrical signal. The PowerFC conditions that signal, applies its filtering and monitoring rules, then compares it with the configured threshold. The threshold is therefore an internal signal level, not a boost pressure, an RPM figure, a timing angle or a measurement in decibels.

On most PowerFC setups, a higher numerical threshold requires a stronger signal before the ECU considers the event to be knock. A lower value makes the system more sensitive. Some software versions also provide an enable setting, RPM gates or related filtering, so the threshold may appear to do nothing if the ECU is not monitoring the input during the relevant operating range.

Correct wiring and sensor type matter just as much as the value. Verify the connector, shield grounding and ECU input against the OEM documentation for the relevant chassis and hardware. An open circuit, poor earth or injector pulse can resemble a changing knock pattern, while a resonant component can produce convincing but false detections.

Separating real knock from ordinary noise

A knock sensor does not identify what made the vibration. It only reports that a particular signal crossed the chosen boundary. Turbo wastegate chatter, injectors, engine mounts, exhaust valves and driveline harmonics can all excite a piezoelectric sensor. This is especially relevant to a rotary engine, whose firing characteristics and acoustic signatures differ from those of a conventional piston engine.

The sensor’s mounting position also affects the result. A bracket attached to an exhaust component may measure turbulence rather than combustion pressure, while a loose mount can create intermittent peaks. Check the installation for rigidity and make sure the wiring cannot pick up alternator or ECU noise.

Car history helps narrow the right starting point. The production figures by year can help identify the generation and original configuration of an FD3S. A naturally aspirated 13B, a turbocharged 13B and a 20B conversion should not be expected to share identical noise floors, even when they use the same threshold number.

Australian conditions change the baseline

Australian weather adds variables that cannot be reproduced on a cool workshop dyno. Ambient temperatures around Darwin, Brisbane or Alice Springs can produce high intake temperatures, while the maritime humidity experienced in Sydney and Melbourne changes air density in a different way. Long drives between cities also create heat soak, and turbo temperature will not always match the short pull used during tuning.

Fuel sold at the servo is another local reality. Unleaded 91 with E10 is widely available, but ethanol content and fuel properties can vary between batches and regions. The PowerFC should be checked against the fuel actually being used rather than an assumed octane rating. Switching to premium or race fuel can move the knock window, although a fuel change alone does not prove that the sensor threshold was wrong.

Australian tuning also uses kPa and degrees Celsius, with speed limits and traffic patterns that often limit repeatable high-load runs. RHD packaging can place sensors and looms near heat sources, and a route through Melbourne traffic may not resemble an open run near Phillip Island. At local RX-7 gatherings, sharing a successful pull is useful, but other drivers may be using different fuel, boost and weather conditions.

A repeatable calibration workflow

Begin by confirming that the engine is mechanically sound. Compression, leak-down results, ignition components, fuel pressure, boost control and cooling should be checked before interpreting the knock trace. The threshold should then be calibrated with conservative timing and fueling, not used to conceal a mechanical or calibration fault.

Use these steps to establish a dependable result:

• Record a baseline log at the same RPM and load points, including knock input, timing correction, boost, intake temperature and fuel conditions.
• Check sensor mounting, connector security and shielding before changing the threshold, because wiring faults can imitate erratic knock signals.
• Adjust the threshold conservatively and retest, saving each configuration so the effect of the change remains identifiable.
• If the ECU detects noise without supporting evidence of knock, investigate resonance and grounding before making the input less sensitive.
• If genuine detonation is indicated, correct ignition timing, fuelling, boost or fuel quality rather than simply raising the threshold.

The PowerFC log should show whether a correction occurs consistently or only at one repeatable operating point. A threshold that looks correct on one pull may still be inappropriate at another. Keep the final value with notes describing fuel, weather, boost and software version, since those details make the calibration reproducible.

Preserving the result for other owners

Save the baseline log, final configuration and observations together. Record the PowerFC software version, sensor part number, mounting location, fuel and approximate ambient conditions. This turns an isolated setting into evidence another enthusiast can evaluate rather than a number to copy blindly.

Clear charts and concise captions also help when sharing results in a local RX-7 group. The social media design guidance can support a readable summary of a calibration, provided the underlying assumptions and safety limits remain visible.

Create the baseline log before touching the threshold, preserve every controlled change, and publish the proven configuration with its fuel and weather notes so the next FD owner can reproduce the result safely.