Almost nobody calls us because their Haltech will not make power. They call because the car cranks and never fires, because the factory dash went dark after the swap, because the ABS light came on and stayed on, because a sensor reads perfectly with the key on and goes to garbage the moment the engine runs. None of those are tuning problems. They are wiring, grounding, and communications problems, and they are the majority of what we solve.
The short version. Most "the tune must be bad" complaints turn out to be electrical. There is a reliable way to tell the difference before you spend a cent on calibration: if the symptom moves with electrical load or with temperature, you are looking at a wiring or grounding fault. If it is repeatable at an exact RPM and load every single time, it is calibration. Almost everything below follows from that one distinction.
On this page
Why this gets solved here
Plenty of shops will sell you a Haltech. Fewer will take ownership of the part where the ECU has to coexist with the rest of the car. Here is the specific reason we can, stated plainly enough that you can hold us to it.
Haltech's own support stops exactly where this work starts
Haltech has said publicly that one-on-one custom CAN network configuration sits outside what their support team provides, and they point users to a training course instead. That is a reasonable line for a manufacturer to draw. It also means that the moment your build needs a custom CAN frame — a factory cluster kept alive, an OEM ABS module read for wheel speeds, a series-mandated timing box fed the exact broadcast it demands — you are on your own unless someone fills that gap. Filling it is a service we sell, not a favour we do. That is what the GCAN Mapping Service (TBS-00323) is: a live session where the frames get built and verified in NSP with you watching.
The person who wires the car is the person who tunes it
This is the structural problem with how most builds are organised. The harness gets built by one party, months later the calibration gets done by another, and neither conversation informs the other. The wiring decisions determine what the tuner can see and control. The tuning requirements determine what the harness has to include. When those are the same person, the sensor you will need in six months gets wired in now, and the ground that would have corrupted your knock signal never gets landed in the wrong place.
We build harnesses for the chassis we see repeatedly
Not just reselling Haltech's looms. When a platform comes through often enough to learn its specific traps, we design and build our own for it — which is why there are parts on this site like an RX-8 K-swap engine harness (TBS-00184) and a Series 1 RX-8 chassis jumper (TBS-00317). Those exist because we got tired of solving the same chassis-side problem by hand every time.
Diagnosis happens live, and the car never ships anywhere
Wiring and CAN faults are the easiest class of problem to solve remotely, which is the opposite of what most people assume. The evidence lives in the log file and the software configuration, both of which we can see in real time over Zoom while you are standing at the car. We tell you what to put a meter on and what we are looking for, and we interpret the result on the spot. No transport, no drop-off window, no car sitting in someone's queue for three weeks. Read more about how the remote process actually works.
The five failure classes behind most Haltech problems
1. Power and grounding architecture
The single largest source of unexplainable behaviour, and the one people are least willing to believe. An ECU is a measuring instrument: every analog reading it takes is a voltage compared against its own reference and its own ground. Put resistance anywhere in the return path, or let a high-current load share a return with a low-current sensor, and every measurement in the car starts drifting with engine load.
What makes this so expensive is that it never presents as an electrical fault. It presents as a tune that will not settle. A ground offset of a few tenths of a volt shows up as a throttle position that reads slightly wrong, a MAP signal that shifts under load, and a wideband that reports lean when the engine is not. Chase those individually and you will spend months. Ground architecture is a topology that has to be designed as a whole — which returns land where, what shares a path with what, and in which order — and it is one of the first things we redraw when a car arrives with symptoms nobody can pin down.
2. Trigger and sync
An engine that cranks and never fires, or that pulls cleanly to 5,000 RPM and then falls on its face, is almost always a trigger problem rather than a fuel or ignition problem.
There are more ways to get this subtly wrong than most people expect. Reluctor sensors are polarity sensitive, so a signal can be present, look healthy on a meter, and still be read at the wrong point. Shielded cable has exactly one correct grounding arrangement and several plausible-looking wrong ones, and the wrong one converts your most noise-sensitive signal into an antenna. Routing matters, because ignition primary current is the noisiest thing in the engine bay and it does not respect a tidy loom.
The diagnostic work here is in capturing what the ECU is actually seeing rather than what the sensor is producing — those are different questions, and only the first one explains the behaviour. Haltech gives you filtering and arming controls in software, and they are genuinely useful, but they are a repair for a marginal signal rather than a substitute for a correct one. Knowing which situation you are in is the whole job.
3. Driver stage mismatches: injectors and ignition
The category where people destroy hardware rather than merely waste weekends.
Injector drivers and ignition outputs are each built to drive a specific kind of load. Put the wrong kind on them and the failure is rarely immediate — it is cumulative, which is exactly why people get away with it for a season and then suddenly do not. By the time an output fails, the cause is weeks in the past and nobody connects the two.
The ignition side has a harder edge to it. A Haltech ignition output is a switching signal, and whether it can drive your coils directly depends entirely on what is inside them. Get that wrong and you do not get a fault code, you get a dead output. Dwell compounds it: every coil family has its own requirement that shifts with supply voltage, and a figure carried over from a different coil is how igniters get cooked. The symptom is a misfire that only appears at high load, which reads convincingly as a fuel problem to anyone working from the wrong assumption. Establishing what is actually fitted, and what it needs, is the part that has to happen before anything gets connected.
4. Drive by wire and safety interlocks
DBW is where sloppy wiring stops being an inconvenience and becomes a safety issue. A Haltech DBW system is built around redundancy: the pedal and the throttle each report their position twice, and the ECU continuously cross-checks the pairs against each other. That cross-check is the mechanism that catches a failing wire before the throttle does something nobody asked for.
The failure mode worth understanding is that a car with that redundancy defeated drives completely normally. There is no warning, no light, and no symptom — right up until the moment there is. That is why DBW commissioning is a deliberate verification step in our process rather than something confirmed by the car driving away under its own power.
5. CAN bus: two separate problems people constantly conflate
Physical layer is whether the bus is electrically sane. It is governed by a small number of rules about how the pair is run, how the network is terminated at its extremities, and how devices hang off the backbone — and those rules are not negotiable or approximate. The reason this catches so many builds is that a bus assembled incorrectly very often works on the bench and in the driveway. Add length, engine bay noise, vibration and real current, and it starts dropping devices under conditions you cannot reproduce on demand.
The useful news is that the physical layer can be characterised quickly and definitively. There is a specific electrical signature a correctly built bus has, and checking it is the first thing we do on any CAN complaint, because it cleanly separates "the wiring is wrong" from "the wiring is fine and the devices do not understand each other." Those two problems look identical from the driver's seat and have nothing in common underneath.
Logical layer is whether the devices understand each other. A perfect bus does nothing if frame IDs, byte positions, scaling, byte order, and update rates do not match what the receiver expects. Haltech covers the common cases with pre-built device support and vehicle CAN profiles. For everything else there is Generic CAN (GCAN) on Nexus ECUs and the iC-7 and uC-10 displays, which gives you the ability to define messages — but not the knowledge of what to define, which is where most builds stall.
Symptom to cause: how we read it
The table below is how we narrow a complaint down before touching anything. It is not a repair list — several of these share symptoms and can only be separated by looking at live data — but it shows the reasoning.
| Symptom | What it usually indicates | What separates it from the alternatives |
|---|---|---|
| Cranks, never fires | Trigger or sync, not fuel or spark | What the ECU records during cranking, which is often not what the sensor is producing |
| Pulls to 4-5k then cuts | Trigger signal degrading as noise rises with RPM | Whether the cut tracks RPM or tracks load |
| Idle hunts only when hot | Ground offset growing with fan and pump load | Whether sensor readings shift as loads switch in |
| Wideband disagrees with the tune | Controller ground offset, or an exhaust leak upstream | Whether the disagreement scales with load or is constant |
| Dash or tach dead after the swap | Cluster waiting on messages the new ECU does not send | Whether anything at all is arriving on the bus |
| ABS or stability light on after the swap | Module lost expected engine data | Which specific frames that chassis and model year require |
| Devices drop off the bus intermittently | Physical layer: termination or topology | The bus's electrical signature, measured cold |
| Injector or ignition output failed | Load mismatch on that driver, or a shorted device | What was fitted versus what the output was configured for |
| Random dropouts over bumps | Connection integrity, or loads run through the wrong supply | Whether it correlates with road surface or with electrical load |
Which Haltech hardware solves what
Before any of the modules below matter, the ECU itself has to be the right one. There is really only one question that decides it, and it is worth answering honestly rather than optimistically, because it is the one choice a firmware update cannot undo later.
Which Haltech ECU does your build need?
The one decision you cannot reverse with a firmware update
Will your build ever need to send or read a CAN message that Haltech has not already built support for?
A factory cluster, ABS or EPS kept alive on a modern chassis. A third-party dash, logger or timing box. An OEM module you want to read.
No — everything is on Haltech's supported list
- Elite 750 / 1500 / 2500. Fully capable within supported devices.
- Nexus Rebel LS / JZ. Plug-and-go; no user-defined CAN.
Yes — or possibly later in the build
- Nexus S2 / S3. Direct successors to the Elite 1500 and 2500.
- Nexus R3 / R5. Highest capability, VCU.
Elite will never receive Generic CAN. Haltech has confirmed the platform's onboard memory is full. Choose the top path when you needed the bottom one and you buy an ECU twice.
Generic CAN is a Nexus feature. The Rebel is a Nexus but is deliberately simplified and does not support user-defined frames, which is why it sits with the Elite range here.
Once the ECU is settled, most remaining capability questions are answered by an expansion module rather than a different ECU:
| Problem | Hardware | What it changes |
|---|---|---|
| A fuse box you cannot diagnose | Nexus PD16 PDM (HT-198000) | Sixteen solid state, software configured, individually logged outputs with programmable current limits and soft start. A chafed wire becomes a logged fault instead of a fire. Works with Elite and Nexus ECUs; not a standalone PDM. |
| Out of inputs or outputs | IO 12 Expander (HT-059900 / HT-059901) | Adds 4 digital pulsed outputs, 4 digital pulsed inputs, and 4 analogue voltage inputs over CAN. Two can run with different CAN IDs for up to 24 extra channels, subject to ECU compatibility. |
| A dash full of toggle switches | CAN Keypad 8 (HT-011501) / 15 (HT-011502) | One CAN drop replaces a bundle of switch wiring. Each button maps in software to any ECU function. |
| Analog wideband that drifts | WB1 (HT-159976) / WB2 (HT-159986) | Digital AFR over CAN removes the ground-offset problem at its root. WB2 gives per-bank readings, which matters more than people expect on V engines and split exhausts. |
| Factory cluster gone or unusable | iC-7 (HT-067010) / uC-10 (HT-068000) | Programmable alarms, logging, and CAN sensor support. With GCAN both now run standalone with CAN sensors and no Haltech ECU in the car. |
| A device Haltech has never heard of | GCAN on Nexus, iC-7, uC-10 | The ability to define your own frames. Free on eligible hardware. Elite does not get it. |
| Keeping the automatic in an LS swap | Nexus Ti6L (HT-196002) | Nexus control of GM 6L80 / 6L90 with a terminated harness and an NSP setup wizard. Nexus series only, including Rebel. |
| Connector mismatch between devices | CAN adaptor looms (HT-130046 family) | DTM-4 to flying leads, 6-pin circular, JST 5-pin, or Minifit Jr. Often the bus needs the right adaptor and a termination review, not rewiring. |
The pattern worth internalising: when a build runs out of capability the instinct is to buy a bigger ECU. Usually the correct move is an expansion module on the CAN bus — cheaper, it preserves the calibration you already paid for, and it puts the channels near what they serve instead of running another twelve wires through the firewall. A bigger ECU is right when you need more injector or ignition drivers, not when you need four more analogue inputs.
Platform-specific guides
The principles above apply to everything. What follows are the traps that exist on one chassis and nowhere else — the things that are not in any generic wiring guide. Each has its own write-up.
- GM LS and LT — 24x versus 58x reluctor, the cam sensor that moved, drive-by-cable versus drive-by-wire pedal mismatches, keeping the 6L80, and why late-model donor chassis fault when the factory ECU leaves.
- Toyota 2JZ and the JZ family — VVT-i versus non-VVT-i, the factory igniter decision, R35 coil conversions, and keeping a JZA80 interior alive.
- Nissan RB, SR and VQ — coil and igniter differences across the family, consult and diagnostics, ATTESA on GT-R chassis, and why Elite versus Nexus matters more here than people think.
- Mazda rotary: RX-7 FD and RX-8 — leading and trailing ignition, staged injection, Series 6 versus Series 7/8 differences, and what to actually do about the sequential twin turbo system.
- Ford Coyote and Barra — four Ti-VCT phasers and the channel count problem, and late Mustang chassis as a CAN project.
- Subaru EJ and FA — single versus dual AVCS, the immobilizer trap, CAN clusters from 2008, and why boxer grounding is unforgiving.
- Honda K-series swaps — VTEC plus VTC control, DBW versus cable by engine code, and the recipient chassis being most of the work.
The general rule for any modern chassis
From roughly the mid-2000s onward, the engine control unit is not an island. ABS, stability control, the body control module, electric power steering, the transmission controller, and the cluster are all listening for engine data on the vehicle bus. Remove the factory ECU and those modules do not merely lose a feature — they actively fault.
There are three honest answers, and which one is right depends on the car:
- Broadcast what they expect. A Haltech vehicle CAN profile if one exists for your platform, custom frames via GCAN if it does not. Keeps the car feeling factory. Requires knowing precisely what your chassis and model year expect, which is a per-vehicle investigation.
- Keep the factory ECU in a limited role alongside the standalone, servicing the modules that need it. Inelegant, effective on some platforms.
- Accept the loss and design around it. Perfectly reasonable on a dedicated race car. Not reasonable on something that still has to drive on a Sunday.
What does not work is discovering which answer you needed after the harness is built.
Misconceptions that cost real money
A handful of beliefs account for a disproportionate share of the expensive mistakes on Haltech builds. The three most costly:
- "Elite ECUs will get Generic CAN eventually." They will not — Haltech has said the platform's memory is full. This is the one ECU decision you cannot reverse with a firmware update, and getting it wrong means buying an ECU twice.
- "One termination resistor is enough, it works on my bench." An incorrectly terminated CAN bus frequently does work on a short bench harness, then fails once length, engine bay noise, and real current are added.
- "I will add the PDM later." A PD16 replaces the fuse and relay layer rather than bolting on beside it, so adding one later means rebuilding the power distribution section of the harness you already paid for.
We have written up all fifteen, grouped by ECU choice, CAN bus, wiring and grounding, and diagnosis, with what each one actually costs: Haltech Wiring Myths: 15 Misconceptions That Cost People Money.
How we work through it
- Define the target first. Street car that must keep its dash and ABS, or a dedicated race car where those are expendable? That answer changes the entire electrical design and has to come before anything is bought.
- Inventory what has to keep talking. Cluster, ABS, traction, transmission, immobilizer, climate, keyless entry. On modern chassis this list drives the ECU choice, because it determines whether you need GCAN.
- Design power and ground as a topology before any signal wiring is planned, because every measurement the ECU later makes depends on it.
- Plan the CAN network on paper. Which devices, which bus, which bit rate, and where the network's extremities physically sit.
- Build the harness to the plan, and document it so it can be diagnosed later by someone who was not there. This is the step most builds skip and most owners eventually pay for.
- Commission systematically. Every sensor, every output, and every safety interlock confirmed before the engine turns over — in a defined order, because some of these checks invalidate each other if run out of sequence.
- Bring it up live. Remote session, first start, trigger verification, DBW failsafe testing, CAN device verification, then calibration.
Motorsport wiring is a real discipline with a real body of knowledge behind it — High Performance Academy teaches it formally through Wiring Fundamentals, the wider Motorsport Wiring track, and CAN Bus Communications Decoded. If you want to understand the theory behind what we do, that is where to start. If you would rather the car simply worked, that is what we are for.
Frequently asked questions
Is my problem a tuning problem or a wiring problem?
If the symptom changes with electrical load such as fans, pumps or lights, moves with temperature, or appears and disappears over bumps, it is electrical. If it is repeatable at a specific RPM and load every time, it is more likely calibration. Logging usually tells us which within a single session.
Can you diagnose a CAN bus problem remotely?
In most cases, yes. The configuration, device list, and error behaviour are all visible in the software, and a log shows what is actually arriving on the bus. Where a physical measurement is needed we will tell you exactly what to put a meter on and interpret the reading with you on the call.
Do I need a Nexus, or will an Elite do the job?
If your build needs user-defined CAN messages you need a Nexus, because Elite ECUs do not receive Generic CAN. If your devices and vehicle are already on Haltech's supported list, an Elite is still a very capable ECU. Our Haltech ECU Guide walks through the decision, and a free 30 minute consultation answers it for your specific car faster than reading will.
Will a Haltech keep my factory dash and ABS working?
Often yes, but it is chassis-specific and not automatic. It depends on whether Haltech has a vehicle CAN profile for your platform, or whether we build the frames with GCAN. Ask before you buy, because the answer can change which ECU you want.
Do I need a PD16, or will relays do?
Relays work. A PD16 adds programmable current limits, soft start, logging of every output, and a diagnosable electrical system instead of a fuse box full of guesses. It requires an Elite or Nexus ECU and does not operate standalone.
Can I keep my automatic transmission?
On GM 6L80 and 6L90 with a Nexus, yes, via the Ti6L. Confirm your transmission controller's operating system version before ordering. Other platforms vary considerably, so ask about your specific combination.
Do you build custom harnesses, or only sell parts?
Both. We build harnesses in house, including platform-specific looms for chassis we see repeatedly, and we run remote tuning and CAN configuration sessions over Zoom.
My car is a street car. Does any of this apply?
All of it applies, and one thing more: Haltech products are sold for sanctioned off-road and competition, non-emissions-controlled vehicles. In many places it is unlawful to tamper with emissions equipment, and compliance is the installer's and user's responsibility. Check your local rules before you buy or install.
Get the integration sorted before it costs you a season
Tuned By Shawn is a Platinum Haltech dealer. We build wiring harnesses in house, we run remote tuning sessions over Zoom, and CAN configuration is part of both. If you are planning a swap, fighting a dash that went dark, chasing an intermittent nobody can find, or working out whether you need a bigger ECU or the right expansion module, we would rather talk before the harness is built than after.
- Start with a free 30 minute consultation to scope the build.
- Book a live session on our scheduling page.
- Need custom CAN frames built and verified in NSP? That is the Haltech GCAN Mapping Service (TBS-00323).
- Browse the Nexus range or the full Haltech catalog.
Share:
Haltech Controlled 8-Speed Sport Direct Shift Transmission AA80E
Haltech Wiring Myths: 15 Misconceptions That Cost People Money