The short version
An injector is an electrically operated valve. When the ECU switches it on, the valve doesn't snap open instantly, and when the ECU switches it off, it doesn't slam shut instantly either. The ECU thinks in terms of the pulse it commanded, but the engine only gets the fuel that actually came out of the nozzle. Injector dead time (also called latency or offset) is the extra pulse time the ECU tacks on so that the fuel delivered matches the fuel requested.
It's one of those settings people rarely look at until something feels off. Once you understand what it's correcting, it becomes easy to spot when it's wrong.
Where the missing fuel goes
Break a single injection event into stages and you can see where the losses come from:
- Coil build-up. Current has to flow into the solenoid winding and build a magnetic field strong enough to move the pintle. Nothing is flowing yet. This is pure lost time.
- Opening travel. The pintle is moving, so some fuel is coming out, but the flow is climbing from zero toward full flow rather than being at full flow.
- Fully open. The only part of the event where the injector flows at its rated capacity.
- Closing travel. The ECU has already switched off, but the pintle takes a moment to seat. Fuel keeps trickling out, which is a small bonus amount the ECU didn't ask for.
Add those up and the injector is almost always short on fuel. The losses at the start (build-up plus opening) are bigger than the bonus at the end (closing). Dead time is simply the net shortfall, expressed as time. The ECU converts the partial-flow ramps into an equivalent chunk of "full-flow time," subtracts the closing bonus, adds the coil build-up, and extends each pulse by that amount.
Why it matters most at idle
The key detail is that those losses take roughly the same amount of time no matter how long the pulse is. Here's an illustration with made-up round numbers. Suppose your injector wastes the equivalent of 1 millisecond every time it fires:
- At wide-open throttle with a 15 ms pulse, 1 ms is a small slice. A dead time error there barely moves the air/fuel ratio.
- At idle with a 2 ms pulse, that same 1 ms is half the event. A small dead time error there is a huge fueling error.
That's why dead time problems show up at idle, light cruise and decel, and almost never on the dyno at full load.
What changes dead time
| Factor | Effect | How the ECU handles it |
|---|---|---|
| Battery voltage | Lower voltage means the coil takes longer to energize, so dead time grows. This is the biggest influence by far. | Dead time is entered as a table against voltage |
| Fuel pressure | More pressure pushes harder on the pintle, which changes how long it takes to open and close. | An extra table dimension, needed if your fuel pressure moves around a lot |
| Opening and closing shape | The ramps aren't straight lines, they're curves, which makes dead time hard to calculate on paper. | Measured data from a flow bench beats estimating |
If your fuel pressure is steady (for example a manifold-referenced regulator doing its job), the voltage axis does most of the work. See why you need a fuel pressure regulator and fuel pressure sensor diagnosis if you're not sure your pressure is stable.
Reading your VE table for clues
The fastest way to spot bad dead time data is to look at the shape of a tuned VE table. With correct injector data, VE rises smoothly with load and RPM and looks like something an engine could physically do.
| What the VE table looks like | Usually |
|---|---|
| Smooth, sensible values from idle to full load | Injector data is right. Leave it alone. |
| Light-load and idle cells pushed unusually high to get the engine to run right | Dead time is too short. The tune is propping up missing fuel with VE. Increase dead time. |
| Light-load cells heading toward zero, or even hitting zero | Dead time is too long. Air is obviously still flowing, so a VE near zero can't be real. Reduce dead time. |
| Normal VE, but AFR shifts when voltage changes (for example with the fans or headlights on) | The voltage axis of the dead time table is off |
Finding dead time when there's no data
Good injector data from the manufacturer or a flow bench is always the first choice. Haltech's Bosch injector data covers a lot of common units. When all you have is a flow figure, or nothing at all, you can still get close with a methodical process. The idea is to get flow rate and dead time close before you start editing VE, so VE ends up describing the engine and not hiding injector errors.
Step 1: Use idle VE as a reality check
On a mostly stock engine, leave the VE table alone and adjust flow rate and dead time until the engine idles and drives reasonably with AFR in the right neighborhood. To judge whether your idle VE is believable, compare it to how much vacuum the engine pulls at idle. Engines with a lot of cam overlap pull less vacuum and have lower idle VE. A rough guide:
| Idle vacuum | Roughly in inHg | Expected idle VE (approx.) |
|---|---|---|
| -10 kPa | 3 inHg | 25 |
| -20 kPa | 6 inHg | 30 |
| -30 kPa | 9 inHg | 35 |
| -40 kPa | 12 inHg | 40 |
| -50 kPa | 15 inHg | 45 |
| -60 kPa | 18 inHg | 50 |
| -70 kPa | 21 inHg | 55 |
| -80 kPa | 24 inHg | 60 |
The logic: an efficient engine needs less air to hold idle, so the throttle is closed further and vacuum is deeper. Higher vacuum at idle points to a more efficient engine at that manifold pressure.
Step 2: Dial in flow rate where dead time barely matters
Load the engine heavily at modest RPM, somewhere around 1,500 to 2,500 rpm, so pulse widths are long and dead time is a small fraction of each event. Don't beat on it. If the AFR reads lean here, lower the injector flow rate setting (the ECU then commands longer pulses). If it reads rich, raise the flow rate.
Step 3: Dial in dead time where it matters most
Changing flow rate will have moved your idle AFR. Go back to idle and shift the whole dead time table up or down until idle AFR is back on target. Then recheck flow rate under load, and repeat both. It's a see-saw: each pass gets smaller until neither adjustment is doing much.
Step 4: Fill in the voltage axis
With the alternator running you'll typically see about 14 V. Disable the alternator and let the battery pull down toward 12 V while you watch idle AFR, adjusting dead time at the lower voltage. Two good points give you a solid estimate of the slope for the rest of the table. Chasing accuracy much below that is rarely worth it on a street car. If the car is running at 10 V, a small fueling error is the least of its problems.
Do this before you tune VE, not after. If you change dead time on a finished tune, every light-load cell you already tuned will move.
Dead time FAQ
Can wrong dead time stop my car starting, cost power or slow boost?
No, not in any meaningful way. A no-start points to something more basic, like trigger, fuel pressure or injector wiring. If AFR and timing are right at full load, missing power is mechanical, and dead time has nothing to do with how quickly a turbo spools.
My engine runs fine but the VE table looks strange. Is it dead time?
Very likely. That's the classic symptom: the car drives well because VE has been bent to compensate for bad injector data. Fix the data and the table will look like an engine again.
AFR is unstable only at extremely light load. Dead time?
Probably not. That's more often the injector working in its non-linear zone, where pulses are so short the pintle barely opens before it's told to close. That's a separate problem from dead time.
If you'd rather have us sort out the injector data for your build, book a free 30-minute consultation. The ECU-side setting itself is covered in our Nexus dead time article.
