Troubleshooting

By what you see on screen: what it usually means, what to check in order, what to do, and what not to do.

The first check is always the same: what changed since the last good run? The Δ vs baseline tune overlay answers that for the tables; the run strip and Runs & compare answer it for conditions and hardware. Most "sudden" problems are a change you forgot.

Knock counts appear where there were none

Usually conditions, not the map: heat soak from back-to-back pulls, a hotter day, a lower-octane fuel, or an edit that moved the knock limit (a leaner target, more boost).

• Check the Heat soak line under RUN and the IAT at the start of the pull. Ten degrees of soak is close to a degree of knock margin. Cool down and run again before changing anything.
• Check the Δ vs baseline tune overlay. If the last edit was in those cells, undo it.
• Set the Overlay to Knock level / threshold on the Ignition table. Isolated counts in scattered cells with the retard recovering: trace, leave the map alone. Counts on the same cells pull after pull: take 0.5–1° out of those cells only.
• If the counts stay with timing back at stock, the cause is charge temperature, boost or fuel. See Heat and Knock.

Counts climb during a pull and the retard does not recover

Real detonation. The knock control is at or near its maximum retard and the map is past the limit in that region.

• Lift off: Space or ABORT. Auto-abort does this for you if it is on.
• Open Diagnostics and click Select the cells. Remove at least the retard the ECU was holding plus 1°, in those cells.
• Check the Health panel. If it moved, that pull cost metal; the X-Ray channels show how close it was.

Knock on every pull at stock timing after fitting a part

Parts that add airflow or boost raise cylinder pressure, which lowers the knock limit. The factory margin in the hottest row (peak boost, mid rpm) may already be gone. One or two counts per pull with a single retard step is this hardware's floor on this fuel.

• Accept it, or take 0.5° out of that row only. Do not chase zero. The lever that moves the floor is fuel (octane, E85) or a cooler charge (intercooler).

"Light knock" warning: a few counts, retard recovered

• Take 0.5–1° out of the named cells and pull again. If the counts drop to one or two, that was the margin; keep it. If they do not change, timing was not the cause: look at IAT, boost spikes and λ in those cells.

Retard events while the knock level stays inside normal noise

False knock: the Knock threshold table is too low at that rpm, so background noise trips it. Power is lost to retard for nothing.

• Compare the knock level trace with the threshold line in the Log viewer. Raise the threshold at those rpm back toward stock, and never above stock without a reason.

λ lean under load (measured above target)

In order of likelihood: the VE table under-estimates airflow in those cells; the injectors ran out (duty at 100 %); fuel pressure sagged (pump); the ECU's injector or stoich facts are wrong.

• Result card first: injector duty and fuel pressure. Static injectors or a falling Δ pressure are hardware, and no VE edit fixes them.
• If the error is the same percentage everywhere, it is a scalar: injector flow, dead time or stoich. See Injectors.
• Otherwise Suggested VE on the lean cells, and run again.

Rich everywhere by the same amount

• A uniform error is not VE. Check injector flow and dead time against the installed injectors, and stoich against the fuel in the tank. Fixing it in VE bakes a hardware error into the airflow model.

λ error large at idle and light load, small at wide-open

• The dead-time signature. A dead-time error is a fixed slice of every pulse, so it matters most where pulses are short. Set the dead-time scalar for the installed injectors; do not chase it in VE.

λ swinging, torque ragged, power down: misfire

• Too rich: λ below about 0.70 under load. Lean the target or fix the VE in those cells.
• At the rev limiter with a spark cut: expected. Fuel is delivered and not burned, and EGT spikes. Use a fuel cut on turbo cars.
• Injectors static: the mixture goes wherever the air takes it. Hardware.

EGT high, or "EGT protection added fuel"

The exhaust runs hot when the burn is late (retard), the mixture is lean, or the load is sustained. The protection table richens the final target above 850 °C on purpose, and the VE correction already accounts for it.

• Do not correct VE for the enrichment, and do not switch protection off.
• Lower EGT at the source: fix the ignition map so the ECU is not holding retard, a richer target in the hot cells, less boost there.
• Advancing timing lowers EGT; leaning raises it. If you lean a top end for power, read EGT on a cooled pull and leave 10 °C under the warning level for a soaked one.

Injector duty over 85 %, or injectors static

• Hardware headroom is gone. Leaner targets buy a percent or two; nothing else in the tables helps.
• More airflow, boost or E85 needs bigger injectors and usually a pump first. The pre-run check estimates E85 capacity before you run.

Fuel pressure dropping at high flow

• The pump cannot keep up as rail pressure rises with boost. The injectors flow less than the ECU thinks and λ goes lean at the top. Do not correct VE for it. A pump.

The boost cut fired on the first pull after new boost parts

The Boost table is still the stock one and the new turbo, spring or solenoid makes far more boost per percent of duty. The cut did its job. Peak MAP on that run is not a tuning number: it is how far pressure overshot after fuel was already cut.

• Set every Boost cell to 0, keep the cut 25–30 kPa above the spring's base boost, and repeat the pull short and part-throttle in real time. Read Peak MAP from a run where the cut did not fire; that is the number to index the cut off.
• If MAP still climbs past the cut at 0 % duty, the gate cannot bypass the larger turbo: creep. A larger wastegate is the fix, not a table.

Health dropped, but there was no knock and EGT looked fine

Piston crown heat. The crown is heated by cylinder pressure and cooled by fuel; a leaner target at full load and high rpm raises its temperature long before it raises EGT, because the exhaust probe measures the gas leaving, not the metal it left. The diagnostics now name the component, the cause the teardown found and the cells it happened in.

• Richen the target in the named cells (0.74–0.76 under full boost at the top of the rev range is where big-turbo E85 builds live), or take boost out of those cells, or spend less time there (a lower rev limit, a shorter sweep).
• Forged pistons raise the limit and slow the damage; they do not remove it. The Health panel's "Limits for this build" lists the crown limit for your parts.
• This is the one limit the gauges cannot show you. Treat λ at full load as the piston thermometer.

It was fine on the hot day and broke on the cold one

Cold air is dense. At the same boost pressure the cylinder takes in more air, cylinder pressure rises, and the piston crown follows pressure; on a boosted engine the 5 °C day is the worst case, not the 38 °C one. On E85 nothing warns you, because the knock counter stays silent and the IAT trim only pulls timing when the air is hot.

• Set the margin on the cold day, not the mild one: the boost duty, the top-end λ and the last degrees of timing are the three levers, and the cold pull decides how much of each the engine can carry.
• Real ECUs compensate boost against air temperature; this app does not (yet), so the duty table has to be safe at 5 °C. The Health panel's "Limits for this build" and the X-Ray crown gauge show how close you are.

Closed-loop: boost never reaches the target, or sits above it

The controller can only move the duty within its trim limit. If it is pinned at the limit, the base duty table is wrong by more than the limit, or the hardware cannot make the target at all.

• Read the diagnostic: "saturated" with the duty at 100 % means the turbo or the exhaust energy cannot make that target here; lower the target. Saturated below 100 % means the base duty is too low; raise the base cells toward base + trim and let the trim settle to zero.
• Above target at the negative limit: base duty too high, or creep. If it will not come down at 0 % duty, the gate is too small.
• Not saturated and still off: the gains are low. Raise P by 0.10 %/kPa and I by 0.20 %/kPa·s at a time.

Closed-loop: boost hunts around the target

• Too much gain. Halve P first; if the swing is slow (a cycle per second or more), halve I instead. A base table that is nearly right needs small gains; let the base table do the work and the controller the trimming.

Boost overshoots on spool, creeps, or trips the boost cut

• Overshoot: the wastegate duty at the spool rpm is too high for how fast the turbo arrives. Lower the duty in the 2500–3500 rpm cells, not everywhere.
• Creep at high rpm: the wastegate cannot bypass enough. Lower the top-end duty or accept a taper.
• Boost cut: it is doing its job. Raise it only after the duty table is under control, and keep it inside the MAP axis.
• Less boost than expected with turbo speed high: altitude or a hot day. The compressor works harder for the same MAP; lower the target.

Power lower than last time with no edits

• The Heat soak line under RUN and the start IAT. Ten degrees of soak costs about 4 % on a turbo car.
• The fairness notes in Runs & compare: ambient, pressure, start temperature, hardware, mode, correction standard.
• Deltas under 0.7 % are noise, and the result card says so.

A timing step gave nothing

• You are at MBT (light load) or at the knock limit (high load). Take the step back out and stop. There is no more in that block; see Timing.

Results differ run to run with no edits

• Peak power varies about ±0.3 % between pulls, knock counts by one or two. Three cooled pulls before you believe a trend. More than that, look for soak or a fairness note.

The engine failed

• Read the teardown: the headline, the cause chain (how much margin each factor took) and the prevention list. That is the lesson; do not skip it.
• Rebuild from the Health panel. Restore the last starred tune from Runs & compare, not the one that broke it.
• The last seconds of the log are kept with the failure marker. Scrub them with X-Ray on.

After Reset to stock the car runs worse

• The stock tune is for stock hardware on the recommended fuel. With parts fitted it is wrong in exactly the ways the retune banner lists. Start the method at stage 0.

Steady-state: coolant climbing, timing dropping

• A long hold at load heats the coolant, the ECT trim pulls timing, torque falls. That is the trim working. Let it cool, hold for shorter periods, and change the ECT trim only with a reason.

Everything is a warning and nothing makes sense

• Export the project, Reset to stock, fit the parts again, start at stage 0. Twenty pulls in the right order beat a hundred in the wrong one.