Every one of these costs somebody money, and most of them cost it twice: once when the wrong part gets bought, and again when the harness has to come back apart to correct it. We hear all fifteen often enough that it is worth putting them in one place.
If you only take one thing: the expensive mistakes on a Haltech build are almost never tuning mistakes. They are decisions made at the planning stage — ECU choice, power distribution, CAN architecture — that cannot be undone later without rebuilding something you already paid for.
The technical background behind these is in our main Haltech wiring and CAN bus guide.
Buying and ECU choice
"Plug and play means I do not have to do any wiring."
A plug-in adaptor harness eliminates the engine harness work. It does not eliminate the decisions. You still have widebands, fuel pumps, fans, boost control, and every sensor your build needs beyond stock. Plug-in gets you a clean, reversible foundation quickly — which is genuinely valuable — but it does not turn a modified car into a stock one. Budget for the rest.
"Elite ECUs will get Generic CAN eventually."
They will not. Haltech has stated the Elite platform's onboard memory is already full and there is no room for it. GCAN is a Nexus feature, and the Nexus Rebel does not get the full version either.
This is the most consequential ECU decision on the platform, because it is the one you cannot reverse with a firmware update. If there is any realistic chance your build will need a custom CAN frame — a factory cluster kept alive, a third-party dash, a data logger, a module you want to read — buy a Nexus now rather than an ECU twice. The Nexus S2 and S3 are the direct successors to the Elite 1500 and 2500.
"I have run out of capability, so I need a bigger ECU."
Usually you need an expansion module, not a bigger ECU. Out of inputs is an IO 12 Expander question. Out of switched power is a PD16 question. Out of dash is an iC-7 question.
An expander is cheaper than an ECU tier, it preserves the calibration you already paid for, and it puts the channels physically near what they serve instead of running another twelve wires through the firewall. A bigger ECU is the right answer when you need more injector or ignition drivers, or genuinely more processing — not when you need four more analogue inputs. Which side of that line your build falls on is worth establishing before you spend.
"I will add the PDM later."
Adding a PD16 later means rebuilding the power distribution section of the harness you just paid someone to build. Solid state outputs replace the fuse and relay layer entirely, so it is not a module that bolts on beside the existing system — it replaces part of it.
This is a design-time decision. Make it before the harness is built, or accept that you are paying for that section twice.
CAN bus
"CAN is just two wires, polarity does not matter."
The two lines are not interchangeable, and how the pair is physically run is part of the specification rather than a matter of tidiness. Get it wrong and you produce an intermittent that only appears under load — the worst category of fault to chase, because you cannot reproduce it on demand in the garage.
"One termination resistor is enough — it works on my bench."
Termination is not a suggestion and it is not a single component. A CAN network has a defined electrical arrangement at its extremities, and a bus that departs from it behaves unpredictably in ways that scale with length and noise.
The reason this myth survives is that an incorrectly terminated bus frequently does work on a short bench harness. The bench tells you it is fine. Then you add length, engine bay noise, vibration and real current, and it stops being fine under conditions the bench could never have shown you. A correctly built bus has a specific electrical signature, and confirming it is the first thing we check on any CAN complaint.
"I can lay the bus out however is convenient."
CAN has a topology it expects, and convenience is usually the enemy of it. Networks laid out for ease of installation rather than to specification produce errors that look random and are not. A device hung a long way off the main run is a problem waiting for a rough enough road.
Layout is part of the design, decided on paper before anything is built — not discovered during the install.
"GCAN will figure out my car's CAN messages for me."
It will not, and this is the most common misunderstanding about the feature. GCAN gives you the ability to define messages. It does not decode a vehicle for you.
To talk to a device or a module you need its frame IDs, byte positions, scaling, byte order and update rates. That information comes from the manufacturer, from a DBC file, or from someone doing the logging and reverse engineering to establish it. Haltech has said one-on-one configuration help sits outside what their support team provides, and points users to training instead. Supplying that missing knowledge, and building and verifying the frames, is precisely what our GCAN Mapping Service (TBS-00323) exists to do.
Wiring and grounding
"I can just splice into the factory harness."
You can, and every splice is a future intermittent. Scotch-locks and twisted-and-taped joints in an engine bay fail on a timeline measured in seasons rather than years, and they fail intermittently rather than cleanly — which means they cost far more diagnostic time than they ever saved in build time.
Splicing also destroys something valuable: a known-good harness to test against. Once the factory loom has been cut into in six places by three different people, there is no reference left to compare anything to.
"Bigger wire is always safer."
Gauge is selected against the current a circuit carries and the voltage drop over its actual run length — it is a calculation, not a philosophy. Oversizing everything adds weight, degrades terminations, makes bundles too stiff to route properly, and strains connectors never designed to carry it.
"Any wideband will do."
An analog wideband outputs a voltage referenced to that controller's ground. If that reference sits even slightly away from the ECU's own, the AFR the ECU records is not the AFR the sensor measured — and you will build an entire calibration on top of a number that is quietly wrong. Nothing about this announces itself; the tune simply never quite settles.
A CAN wideband such as the WB1 or dual-channel WB2 transmits a number rather than a voltage, which removes the failure mode rather than managing it. On V engines and split exhausts, per-bank readings matter more than most people expect.
"My idle is weird, so the tune must be wrong."
Sometimes. More often it is a ground offset, an air leak, or a throttle that has not been recalibrated since something mechanical changed.
The tell is behavioural. If the symptom moves with electrical load — fans cycling, pumps running, headlights on versus off — or changes with temperature, it is electrical. Tuning around a wiring fault produces a calibration that works on the day it was made and drifts afterwards, because you have encoded a fault into a map and the fault does not hold still.
Diagnosis and process
"A standalone ECU will kill my ABS and power steering."
On an older car there is usually nothing to kill. On a modern one it is a real risk and a solvable one.
From roughly the mid-2000s, the ABS module, stability control, body control module, electric power steering, transmission controller and cluster are all listening for engine data on the vehicle bus. Remove the factory ECU and they do not gracefully lose a feature — they fault. The answer is to broadcast what they expect, using a Haltech vehicle CAN profile where one exists or custom frames via GCAN where it does not. Establishing exactly which frames your chassis and model year require is a per-vehicle investigation, not a product you buy off a shelf.
"Remote tuning cannot help with a wiring problem."
This one is backwards, and it costs people weeks. Wiring and CAN faults are among the easiest problems to work through remotely, because the evidence lives in the log file and the software configuration — both of which can be seen live.
A session with you at the car and us in the software resolves most integration problems faster than shipping a car anywhere and waiting for it to reach the front of someone's queue. Where a physical measurement is needed, we tell you what to put a meter on and read the result with you.
"The wiring guy and the tuner are separate jobs."
They overlap far more than either party usually admits, and this is the structural reason builds go wrong.
The wiring decisions determine what the tuner can see and control. The tuning requirements determine what the harness needs to contain. When those are two parties working months apart who never speak, the sensor you will need in six months does not get wired, and the ground that will corrupt your knock signal gets landed somewhere convenient. When it is the same person, neither happens.
How to tell which of these you are looking at
Most of the above produce symptoms that overlap, which is why self-diagnosis so often lands on the wrong one and why "I changed three things and it got better" is not an answer you can build on. The separation is rarely a single test — it comes from correlating what the ECU logged against what the car was doing at that moment, and knowing which of the plausible causes produces which signature.
That correlation is the part we do. A live session usually narrows a long-running mystery to a specific cause inside the first hour, because we are looking for a pattern rather than working through a checklist.
Get a second opinion before you buy
Tuned By Shawn is a Platinum Haltech dealer. We build harnesses in house and tune remotely over Zoom, so the ECU choice, the power distribution plan, and the calibration are one conversation rather than three. Nearly every mistake on this page is cheap to avoid at the planning stage and expensive to correct afterwards — which is the entire argument for having the conversation early.
Share:
Haltech Nissan Wiring: RB and SR Coils, ATTESA and Elite vs Nexus
Haltech LS Swap Wiring: 24x vs 58x, DBW Pedals and Keeping the 6L80