Hard Starting After a Tune: Cranking Fuel, Startup Enrichment, and What to Check First

A tune that runs beautifully once the engine is going but takes three attempts to catch on a cold morning is one of the most common post-flash complaints, and one of the most misdiagnosed. People replace fuel pumps, crank sensors and starters chasing it. The overwhelming majority of the time the hardware was fine and the calibration simply never had its startup tables adjusted to match everything else that changed.

The reason is structural. Startup fuelling is not calculated the same way running fuelling is, it lives in different tables, and it does not learn. As of September 2026 that is still true on every HP Tuners-supported platform we work with, which is why the same problem keeps appearing on cars whose owners did everything else right.

Why startup is a separate calculation

When you turn the key, the computer has almost nothing to work with. The engine is not spinning fast enough for the airflow sensors to give a meaningful reading, the oxygen sensors are cold and legally ignored, and there is no stable manifold pressure signal to build an air estimate from. The normal fuelling maths — volumetric efficiency or mass airflow in, injector pulse width out — has no valid inputs.

So the calibration does not use it. During cranking it falls back to a fixed lookup: at this coolant temperature, deliver this pulse width, regardless of what anything else says. Once the engine catches and stabilises, a second set of tables tapers extra fuel away over the first few seconds, and only then does normal closed-loop control take over.

That means three distinct phases, each with its own tables:

• Cranking — a fixed pulse width by coolant temperature. No feedback, no correction.
• After-start enrichment — extra fuel added the moment the engine catches, decaying over time or cylinder events.
• Warm-up enrichment — a longer, gentler multiplier that fades as the engine reaches operating temperature.

Fuel trims do not touch the first two. This is the single most important fact in this article. You can drive a car for a month with a broken cranking table and it will never get better, because nothing in the system is watching that phase and correcting it. It must be edited by hand.

What actually changed to break it

Almost every startup complaint traces back to one of four modifications, and each breaks startup for a slightly different reason.

Bigger injectors. Cranking fuel is commanded in milliseconds of open time. Larger injectors flow more per millisecond, so an unchanged cranking table delivers proportionally more fuel. Fit injectors 40% larger and the cold start is roughly 40% richer than the calibration author intended. On some platforms the cranking table scales automatically with the injector flow constant; on others it does not, and knowing which you have is the difference between a five-minute fix and a week of guessing. Getting the running fuelling right through correct injector characterisation does not necessarily carry across to the startup tables.

A camshaft with more overlap. More overlap means more residual exhaust gas trapped in the cylinder at low engine speed, and cranking speed is the lowest speed the engine ever sees. The mixture that would have lit a stock engine will not reliably light one with a big cam, so the engine needs noticeably more fuel during cranking than stock and often more spark advance too. A cammed engine that starts hard is usually starving during cranking, not flooding.

An air model change. Switching between airflow-metered and speed density operation changes what the computer does immediately after the catch, which changes how violently the transition into normal fuelling happens. A car that catches then immediately stumbles and nearly dies is usually failing at the handoff rather than during cranking itself.

A forced-induction install. These change residual manifold conditions and, more importantly, usually come with all three of the above at once.

The symptom table

The single most useful diagnostic here is not a tool, it is paying attention to when the problem happens. Startup faults are strongly temperature-dependent, and the temperature at which they appear points directly at the table that is wrong.

Symptom	When it happens	Most likely cause	Where to look
Long crank, then catches and runs fine	Cold only	Cranking fuel too low for the cam or injectors	Cranking pulse width, cold end
Catches, black smoke, smells of fuel	Cold only	Cranking fuel too high after an injector change	Cranking pulse width, cold end
Long crank, catches, stumbles, clears with throttle	Hot restart only	Too much fuel for a heat-soaked engine	Cranking pulse width, warm end
Catches instantly then almost dies	Any temperature	After-start enrichment decays too fast, or air-model handoff	After-start enrichment and taper rate
Runs rough for the first minute, then clean	Cold only	Warm-up enrichment mismatched to the build	Warm-up multiplier by coolant temp
Long crank identical hot and cold	Every start	Not a calibration fault — fuel pressure, battery, or a sensor	Fuel pressure test, battery load test

That last row matters as much as the others. A fault that behaves identically at every temperature is almost certainly not in these tables, because these tables are indexed by temperature. If your car cranks for exactly the same duration on a freezing morning and on a hot afternoon, stop editing and go check fuel pressure and cranking voltage. That is the same logic that governs an LS swap that cranks but will not start at all — the temperature signature tells you which half of the problem space you are in.

Getting a log that actually contains the event

Most people cannot diagnose a startup problem because their log does not include it. Scanner software commonly begins recording once it sees the engine running, which means the cranking phase — the part you need — is already over by the first frame.

Set the recording trigger to key-on, or start the log manually before you turn the key and let it run through the whole sequence. You want the file to open with the engine at zero RPM.

Channels worth having, and why:

• Engine coolant temperature — the axis every startup table is indexed by. Without it you cannot know which cell you were in.
• Engine speed — shows cranking speed, the catch, and whether the engine sagged afterwards.
• Commanded injector pulse width — what the calibration actually asked for, as opposed to what you think you wrote.
• Wideband air-fuel ratio — the only honest answer to "rich or lean". Narrowband sensors are not warm enough to be believed for the first several seconds, so a wideband is close to essential for this specific diagnosis.
• Battery voltage — cranking voltage below roughly ten volts changes both starter speed and injector behaviour.
• Short-term fuel trim — only useful after the catch, to see how hard the system is fighting once it does start correcting.

Two seconds of that data settles the question that everyone argues about: was the engine too rich or too lean before it caught? If the wideband pegs rich immediately, you are flooding. If it hangs lean and the engine only lights once the running tables take over, you are starving. Everything after that is arithmetic. If reading logs is still new territory, our guide to reading a datalog covers the general method before you apply it to this specific case.

The order to change things in

Work top to bottom and change one thing at a time. Startup tables interact, and changing three of them at once means the next start tells you nothing.

First, eliminate hardware. Load-test the battery. Check fuel pressure at key-on and confirm it holds after the pump shuts off — a system that bleeds down overnight gives a textbook cold-start-only long crank that no table can fix. Confirm the engine is mechanically capable of starting: compression, no large vacuum leak, correct plug gap for the build.

Second, fix cranking fuel. Change only the cells for the temperature range where the problem occurs. If it starts fine at 180°F and badly at 40°F, the cells above 150°F are already correct and touching them will break a start that currently works. Move in modest increments — around 10% at a time — and retest with a cold engine, which means waiting, not cheating with a fan.

Third, fix the catch. If the engine now lights promptly but sags immediately afterwards, the problem is after-start enrichment or its decay rate: too little and it dies, too much and it runs rich and rough for several seconds before clearing.

Fourth, fix the first minute. Warm-up enrichment governs how the engine behaves between catching and reaching operating temperature. It is the least urgent and the easiest to over-correct.

Fifth, verify at both ends. A cold-start fix that ruins hot restarts is a common outcome, because people edit the whole column instead of the cells they need. Test genuinely cold, then test after a hard drive with the engine heat-soaked.

What not to do

Two shortcuts appear constantly in forum advice and both make the car worse.

Do not hold the throttle open to start it. On a drive-by-wire car this triggers clear-flood mode, which cuts fuel deliberately. You are not helping the engine start, you are masking a rich condition by removing fuel entirely — and you are training yourself to accept a car that does not start properly.

Do not add cranking fuel to fix a hot restart. Hot restart problems are usually rich, not lean, because fuel has boiled in the rail and the manifold is already carrying vapour. Adding fuel to the warm end of the table makes the second-attempt start worse while appearing to fix nothing.

Where this fits in the wider tune

Startup calibration is unglamorous and it is nobody's favourite part of a tune, which is exactly why it is so often skipped. It has no effect on peak power, it never shows up on a dyno graph, and a shop that hands the car back after a single warm start will never encounter the fault. The owner encounters it the next morning.

If you are running the process yourself, treat startup as a required stage rather than a footnote: after the VE or MAF work is settled and before you call the tune finished, do a genuine overnight cold start with a log running. It costs one morning and it is the difference between a calibration that is technically correct and one the car's owner is actually happy to live with.

TuneVault's Copilot reads the startup tables in your VCM Editor file alongside everything else and flags the mismatches — injector scaling that never reached the cranking table, after-start enrichment left at stock values behind a large cam — as part of the same $39 guided tune that covers the running calibration. It is the same principle as the rest of the process: the tables are all there in your file, the difficulty is knowing which ones the changes you made have quietly invalidated.

The bottom line

A car that starts badly after a tune is telling you which table is wrong through when it misbehaves. Cold-only problems live at the cold end of the cranking table. Hot-restart-only problems live at the warm end. Sag immediately after the catch is after-start enrichment. Identical behaviour at every temperature is not a calibration problem at all.

Log from key-on, change one temperature range at a time, and verify at both temperature extremes before you call it finished. The fix is usually two or three cells and twenty minutes — the expensive part is the fortnight spent replacing parts that were never broken.

Frequently asked questions

Why does my car take longer to start after a tune?

Because startup fuelling is calculated by a separate set of tables from running fuelling, and most tuning changes touch the running tables without adjusting the startup ones to match. Injector changes, cam changes and air-model changes all move what the engine needs during cranking, but the cranking fuel table still asks for the old amount. The engine eventually catches once the running fuel calculation takes over, which is exactly why it starts on the second or third attempt rather than never.

Which table controls cranking fuel?

The cranking or crank pulse width table, indexed by coolant temperature. It is a fixed schedule the computer uses before it has enough sensor data to calculate fuelling normally, so it does not self-correct and fuel trims do not touch it. If cranking is wrong, no amount of driving will teach the calibration to fix it — it has to be edited.

Is a long crank a fuel pump problem or a tuning problem?

Check fuel pressure first, because the symptoms overlap. A pump or leaking-down system that has lost residual pressure gives a long crank that is the same whether the engine is hot or cold and often improves if you cycle the key a few times before starting. A calibration problem is usually temperature-dependent: fine warm and bad cold, or the reverse. Temperature dependence points at the tables; equal behaviour at every temperature points at hardware.

Why does it start fine cold but badly when hot?

Hot restarts are dominated by fuel that has heated in the rail and by residual manifold fuel, so the engine needs less fuel during cranking rather than more. A calibration built around cold starts frequently floods a hot engine, which shows as a long crank that catches, stumbles and clears when you touch the throttle. Reducing the top end of the cranking table, not the cold end, is the fix.

Do injector changes affect starting?

Heavily. Cranking fuel is commanded as a pulse width, so the same milliseconds through larger injectors delivers proportionally more fuel. Fitting injectors substantially bigger than stock without scaling the cranking table gives a flooded, smelly cold start; fitting smaller ones gives a lean no-catch. Correct injector characterisation data fixes the running fuelling but does not automatically rescale the startup tables on every platform.

How do I log a start-up problem?

Set your scanner to record on key-on rather than on engine-run, so the log begins during cranking rather than after the engine catches. Capture coolant temperature, engine speed, commanded and actual fuel pressure if available, short-term trim and the wideband reading. The first two seconds of that log tell you whether the engine was too rich or too lean before it caught, which is the entire diagnosis.

Can a weak battery look like a tuning fault?

Yes, and it is worth ruling out before editing anything. Low cranking voltage slows the starter, which lowers cranking speed and lengthens injector opening delay at the same time. The result is a long crank that appears every cold morning and looks exactly like a fuelling calibration problem. Load-test the battery and check the charging system before you touch a table.