FuelTech FT550 and FT450 on a K-series: trigger and first log

• This page covers port-injected K20 and K24 engines on FuelTech FT450, FT550 and FT600. The trigger and cam details are for the early K hardware covered here. That hardware uses a 12+1 crank, a 4+1 exhaust home and four-tooth intake feedback. A later K24W, K20C or aftermarket wheel can differ, so inspect the wheel before you select a pattern.
• Commissioning establishes the signals, wiring and references needed before fuel tuning can begin.

• The 2013–2015 Accord K24W is direct injection and uses a different fuel system. K20C, L15, B/D and s300 work sit outside this page. KTuner is not covered. The rest of this page uses “early K” for the 12+1 hardware in the table, not for every K-series engine.
• The page does not print a base map or a fuel cell range. No complete FuelTech K calibration with a trigger scope, a timing-light record and a VVT zero is published here. This page gives the method and leaves the missing values open.

• The ECU has to answer two different questions. Where is the crank in its rotation, and which of the two revolutions is the engine on? The crank wheel answers the first. The exhaust cam home answers the second when sequential operation is used. The intake cam wheel answers a third question: where is the moving cam relative to its reference?
• On the early K20/K24 hardware covered here, the crank wheel is 12+1: twelve regular teeth plus one additional synchronization tooth. It is not a 12-1 missing-tooth wheel. The crank and both cam sensors are Hall-effect. The exhaust wheel is 4+1 and provides the home signal. The intake wheel has four equal teeth and provides moving-cam feedback. The four-stroke firing order is 1-3-4-2.

Swipe or use the arrow keys to read every label.

• FuelTech’s RPM table lists 12+1 for a Honda Civic Si with an index of 210 or 330 degrees. Those are starting references, not two final settings. The correct reference depends on the wheel, the sensor edge and the decoder. A Haltech 311 degree figure belongs to a different decoder and does not transfer because the engine name matches. The timing light procedure in section 5 settles the working reference.
• The crank connector pin order differs across early K sensors. The K20 blue and K24 grey connectors carry the same three functions in a different order. Check the view before you repin: sensor pins, wire side and mating face are different drawings.

The stock early K coil has an integrated igniter. The ECU drives a logic signal; the coil gets switched power and a head or engine ground. Coil primary current does not pass through the logic output.

A wrong coil or edge setting can damage the coil or the ECU, so verify the edge before the first crank.

The FT450 has seven configurable inputs and ten outputs. It can run four independent injectors and coils, but pairing frees outputs for other functions. It is not enough for four independent injectors, four independent coils, VTC, VTEC, a pump and a fan at the same time.

Count the functions before you choose the loom.

A sequential four-injector, four-coil install with VTC and VTEC already uses all ten FT450 outputs before the pump and fan. That is why a paired loom exists. It frees outputs by sharing them and changes how the engine is configured.

The pairing follows the firing order. Cylinders 1 and 4 are 360 degrees apart in the 720 degree cycle, and so are 2 and 3. One coil output can fire both in wasted spark. In Semi-sequential injection, each pair fires twice per 720 degree cycle.

The fuel model must match that event count.

• Selecting sequential mode does not unsplice the loom. With paired coils and sequential ignition, each coil output fires once per 720 degrees. The cylinder 360 degrees away loses its compression spark, and the engine runs on two cylinders. A paired coil loom needs wasted spark; a paired injector loom uses Semi-sequential injection for the paired strategy described here. Four independent coils and injectors need four outputs each and an ECU with the outputs to drive them.
• The 12+1 pattern can run wasted spark and Semi-sequential injection without cam sync. Sequential fuel or ignition needs the exhaust home reference. Intake feedback is still required for closed-loop VTC.
• The FT450 does not drive an electronic throttle. A cable-throttle K swap can use it. A drive-by-wire K20Z3 or K24Z7 needs FT550 or FT600 for native throttle control, or a cable conversion. The example FT450 loom also asks for high-impedance injectors and a separate charge loom. Those details still matter after the engine starts.

• The FT600 specification table and buyer guide count 20 inputs; its product description counts 21. Budget from the configurable inputs in the connector diagram, rather than treating the headline count as spare sensor capacity. The FT450 and FT550 counts agree between the manual and the listing.
• Use the FuelTech FT450 K20-24 wiring diagram for the ECU and loom revision you have. Do not merge pin numbers from a third-party loom with the FuelTech drawing because both may use the same connector family. The functions may be the same and the pins may not be.

FTManager 5.71 is the current PowerFT build on 4 November 2026. FT500 and FT500LITE compatibility stops at 4.91; use the documentation for the ECU and software version fitted. FuelTech Vision is the FT470 and FT700 family; its VE learning and direct-injection features do not transfer to FT450, FT550 or FT600.

1. Write the identity sheet before you open the wizard. Record the engine code, head and cams. Add the crank and cam wheels, sensor connectors, injector part and flow, and coil part. Note the throttle type, MAP source, wideband controller, VTC and VTEC outputs, and fuel pressure. A “K24” label does not settle the trigger or the injector data.
2. Set the engine limits. Maximum engine speed sets the last RPM point and the percentage-use display. Maximum boost sets the map extent, not the boost. For a turbo engine the manual suggests 10 psi above the maximum you will use; for a naturally aspirated K, 0.0 psi.
3. Select the 12+1 crank pattern, Hall input and the edge that matches the sensor. Use a 210 or 330 degree index seed. Do not treat the seed as timing. The decoder and the timing light settle the reference.
4. Set the injection and ignition mode to match the loom, not the other way around. Use wasted spark and Semi-sequential injection for the paired loom. Use sequential operation with cam sync when the outputs are independent. The mode is a description of the wiring.
5. Enter the total injector flow for the bank. Four 80 lb/h injectors are 320 lb/h under the same pressure. Do not enter one injector’s flow as the bank total. Check the pressure the flow was rated at.
6. Enter the injector deadtime from the injector data. The 2.9 manual uses 1.00 ms for high-impedance injectors and 0.60 ms for low-impedance injectors on a peak-and-hold driver. Those are defaults, not your injector’s specification.
7. Enter dwell from the coil data. The wizard shows a 3.600 ms at 12.0 V example. That is an example, not a K coil specification. Coil type alone does not establish dwell; the wrong edge can damage the coil or the ECU.
8. Calibrate TPS with the engine off. Check MAP against atmosphere and ECT and IAT against ambient. Associate the wideband before you generate the map, then generate the base map and print the connector view.

The generated map is a starting calibration, not a measured tune. It uses the compression category, fuel type, total flow, cam profile and deadtime you declared. If those inputs are wrong, the map is wrong in a way the O2 loop cannot repair.

The ignition calibration screen locks timing at 0 or 20 degrees at any rpm. The manual’s procedure holds a steady fast idle near 2,000 rpm and warns that rpm swings make the light wander. Do this before any fuel tuning or load run.

1. Establish a trustworthy cylinder-one compression TDC mark. Some K pulleys have a mark you can use; verify it. If it is not trustworthy, establish true TDC mechanically and make a mark. With a dial indicator, bisect equal piston heights on either side of TDC; the piston dwell makes the apparent peak imprecise. A mark that is a degree off is a timing error at every rpm.
2. Start the engine, open the ignition calibration function and lock the timing at 20 degrees. Hold a steady fast idle near 2,000 rpm.
3. Point the light at the mark and adjust the compensation until the light shows the locked angle. Use 20 degrees on a correctly configured light, or 40 only after confirming that its display doubles. If the light reads 24 degrees when the screen is locked at 20, the error is 4 degrees. The manual specifies a −4 degree compensation for that example; remeasure after the change.
4. Test the light before you trust a doubled reading. With the engine unloaded, change calibration compensation by 5 degrees, observe the light, then restore the original compensation. Keep the timing lock active during this check. If the light moves 10 degrees, the light is reading double. With doubling confirmed, a 46 degree display against the expected 40 needs a −3 degree compensation change, not −6.
5. Check the reference at a second steady speed. A fixed offset stays fixed. A reading that drifts with rpm points to an edge, delay or trigger problem, not a number you can fix in the compensation box.
6. Record the locked reference, the light reading and the final compensation. Then release the timing lock and confirm the screen and the light agree at the same steady speed.

The 210 and 330 degree entries in the RPM table are starting points. The manual says the values are only a start and the ignition calibration must always be performed. The same warning applies to a copied index from another engine. The timing light is the final check.

In advanced 3D mode, Main fuel table holds pulse time against load and rpm. The load axis can be MAP or TPS. Basic 2D mode uses a load table plus RPM compensation. The table is not a Hondata VE table and it is not a self-learning VE model.

A display that looks like a volumetric-efficiency table or a duty table is a conversion of the pulse time, not a second fuel model.

Honda calls the intake phasing VTC; FuelTech’s function is Variable camshaft (VVT). The 4.2 release added closed-loop VVT, and the older on/off function is now Variable camshaft (VTEC). VVT moves the intake cam continuously; VTEC switches the rocker to the high lobe.

A calibration that treats one as the other will not work.

The FT450 has no yellow outputs, which is why blue 5 is the one with the built-in suppression. If you use another blue output, add the diode the manual shows. A grey output is not recommended for VVT. This is a hardware limit, not a software preference.

1. Establish the mechanical zero and the safe travel of the phaser before you command a target. Check chain indexing and oil pressure. The ECU does not know piston-to-valve clearance, and an aftermarket cam can turn a software target into contact.
2. Keep the two cam signals in their own jobs. The exhaust 4+1 home goes to the cam sync input. The intake four-tooth feedback goes to the VVT position input. Do not use the intake wheel as the home reference.
3. Set the sensor position angle and the target reference. FuelTech’s VVT target is relative to the configured sensor position and solenoid direction. The manual’s advance example has a sensor at 45 degrees and a 10 degree table value, which commands 55 degrees. That is an example of the arithmetic, not a Honda target.
4. Keep VTEC off and command a small phase step with the engine warm. Log target, actual, duty, rpm, ECT and sync status. Confirm the cam moves in the direction you expect and returns.
5. If the phase moves backward, stops at a fixed angle or oscillates, diagnose the reference and direction first. Check oil pressure, the solenoid and the sensor before you change a PID term. A stable loop can hold the wrong physical zero.
6. Only after the small step repeats should you build the target table and the PID. Then enable VTEC and verify the mechanical switch separately. The output bit is a command, not proof that the rocker moved.

A log is useful when it contains the channels that can answer the question. The internal logger records up to 256 channels. The 2.9 manual’s sampling screen shows 25, 50, 100 and 200 Hz, and the tradeoff is memory.

Use faster rates for available sync-status and cam channels than for coolant temperature. These are processed values; a 200 Hz log is not a tooth waveform. Use trigger diagnostics or an oscilloscope for individual signal edges.

The current logger can save the map into the log file. Turn that on. It lets you match the log to the calibration later. Advanced diagnostics can record pulse edge, tooth count, timeout and noise information. Those channels are more useful than a single rpm trace when the engine will not sync.

1. Set the log start and stop trigger. A manual button is fine for a first start. An rpm trigger is better for a road log. Confirm the memory policy before you record over a log you need.
2. Choose basic or advanced logging. Basic uses one rate for every channel. Advanced lets you set the rate per channel. Use advanced when you need a fast trigger trace and a slow temperature trace in the same file.
3. Download the log through FTManager’s Datalogger and Download path. Save the native .ftml file before you export anything. The native file is the record; the CSV is a copy.
4. Export a CSV and read the first rows. Confirm the timestamp, channel names, units and sample rate. The exact 5.71 datalog export menu and header layout were not captured for this page. Do not invent a header name or a unit that the export does not show.
5. Compare one value in the native log with the same value in the CSV at the same timestamp. If they differ, fix the export before you analyze the run.
6. Keep the map, the native log, the CSV, the software build and the loom revision in one folder. The map and build let you trace commands back to their settings.

Validation starts at warm idle and moves one step at a time. A first start that holds idle is not a tune. Check the sensors, the timing reference and the outputs before you add load.

1. Warm the engine and check the sensors before you touch the fuel table. A plausible but wrong ECT or MAP reading will make every correction wrong.
2. Record a warm idle log and a part-throttle log. Compare lambda, trim and pulse. If the trim is adding fuel at idle, fix the base or the deadtime before you move to load.
3. Run the VTC small step and the VTEC check separately. Do not change the cam target and the fuel table in the same run.
4. Do the hot restart and the electrical load test. A car that starts cold and dies hot is not ready for a road log.
5. Only then add load in steps. Keep the same route, the same fuel and the same channel set. Change one thing per run.

Send the map, the native log, the CSV, the software build, the loom revision and the sensor list when you want the file checked. Include the map so the log can be compared with the calibration.

A number on one row does not transfer to another engine or loom. The scope column is part of the number.