The car was 90% finished for nine days in a row.
That’s not a joke about student engineering. It’s the calendar. Our Formula Student EV car drove under its own power in the middle of September, and the only thing left on the list was wiring — add an energy meter and some sensors so we could finally find out why the car was slow. I budgeted an afternoon for it. It ran from September 26th to October 4th, and several of those days ended at sunrise.
If you’ve ever installed a dash cam, an amplifier or a light bar on your own car and watched a two-hour job eat an entire weekend, you already know the feeling. Car wiring harness planning fails in a specific, predictable way, and it has nothing to do with how fast you work.
The short version: wiring time scales with the number of interactions between circuits, not the number of devices you’re adding. One component is one line on a shopping list and five or six unresolved decisions inside the car — power, ground, fuse, route, physical position, and clearance from everything already in there. Count the decisions instead of the parts and your estimate stops being fiction.
Why We Were Rewiring a Car That Already Worked
During test week in mid-September the car moved for the first time. It was also slow — not broken-slow, just wrong-slow, in a way nobody could explain by staring at it. We had opinions. We had no data.
You can’t argue with a car that has no instruments, so we decided to instrument it: an energy meter on the high-voltage side and sensors we could actually log. That call was right. The one parameter that was strangling the car was invisible until something measured current. The wiring bill for that decision is what this post is about.
The Back of the Car Was Already Full
Here’s what was living behind the driver before we added a single new wire: two brake lights, the motor’s coolant pump, a radiator, the low-voltage feed for the energy meter, an amplifier and speaker, a brake switch and an amp switch. On top of that sat the high-voltage side — motor and controller — plus a precharge circuit. All of it terminated into a hub that had grown to three physical levels, because up was the only direction left.
And a teammate’s data logger board wasn’t even fitted yet. The car was saturated before the thing we were wiring for had arrived. That sentence is the whole failure, and I didn’t see it at the time.
Car Wiring Harness Planning Fails on Interactions, Not Parts
The first thing that ate a night we hadn’t planned: the energy meter’s wiring ended up running close to the main contactor and the negative pack cable. On a schematic that’s fine — those circuits have no relationship. In the car it’s a three-dimensional problem, because two runs with nothing to do with each other end up inches apart purely because that’s where the space was. We fixed it by changing the height of the run and adding an insulating shield between them.
The second example is the one that actually changed how I think. We wanted wheel speed and suspension travel. Four wheels, four sensors. On a shopping list, that’s one line.
In the car it was this. The logger’s microcontroller takes 3.3 V inputs. The proximity sensors we had run on 12 V, so each channel needs a pull-up resistor and a comparator to square the signal down to something the chip will accept. Then the comparator’s datasheet says it only behaves properly up to roughly 1.5 V below its own supply, so each channel also needs a small divider on the non-inverting input to keep it inside that window. Two series resistors, four times over, built on perfboard. And every perfboard ground has to travel back and join the rest of the grounds — which is another connector, another route, another place to physically sit.
Four sensors on a list. Something close to the car’s entire existing low-voltage wire count in reality. We ran out of days and never fitted them, which means the suspension data we started all of this for is data we still don’t have.
Grounding deserves its own warning here, because it is the single most common way a working bench circuit turns into a broken car circuit. I lost eight hours the following season to exactly one missing ground wire, and I had already learned this lesson once.
What “90% Done” Was Hiding
Five of us worked those nine days, and we split cleanly into two camps. Some people wanted every detail settled before anything touched the car. The rest of us — me, most of the time — wanted to build to 90% and let the car tell us what was wrong. Both approaches are defensible, and on this team they genuinely covered for each other.
Wiring is where the second approach breaks down, because it postpones exactly the class of problem that can only be solved by physically moving something. Those problems don’t queue politely. They all come due on the same night, and there is no version of that night that is short.
We never fully tidied that hub. It went to the competition in whatever state it reached at some unrecorded hour in early October, and it was still in that state on the last day the car ever ran. The relay that ended our endurance semifinal was sitting on the third level of it.
I’m not claiming the mess killed that relay — our conclusion was current, not workmanship, and we never cut the part open. What I’ll own is quieter and worse. I had stopped looking at that hub. It had become scenery. You stop inspecting the part of the car you’re tired of, and anything you can’t see clearly is something you can’t check.
How I’d Plan the Same Job Now
This is the part you can steal for your own car, whether you’re wiring a race car or adding a second battery to a truck. Good car wiring harness planning is mostly subtraction and bookkeeping, and almost none of it happens with a crimper in your hand.
- Count circuits, not devices. Write down every new device, then next to it write power, ground, fuse, switch, route and physical location. Six lines per device is a realistic starting point. That number, not the device count, is your schedule.
- Decide where every ground goes before you buy anything. Grounds are the wires people plan last and the wires that cause the most mysterious faults. Pick your ground points on the car first, then design backwards from them.
- Give high-current runs their own corridor. Battery cables and contactor feeds get a path, and signal wiring is not allowed in it. We learned this by having to raise a run and shield it after the fact, which costs several times what routing it correctly would have.
- Delete before you add. This is the one that actually worked. Our plan for the following car was to switch to an air-cooled controller — which deletes the pump, the radiator and all of their wiring — swap the amplifier and speaker for a single motorcycle-type unit, move the high-voltage side onto a proper distribution panel, and put those divider networks on a small PCB instead of perfboard. Our estimate was that the harness would come down to roughly a quarter of its size. That’s a projection from the parts list, not a measured result.
- If your harness has levels, it isn’t a harness. Stacking is what a wiring loom does when it grows without a plan. The moment you find yourself building upward, stop and delete something instead.
- Crimp properly, once. Rework is where the nights go. If you want a genuinely good free reference for how connections should be made, NASA publishes its workmanship standard for crimping, cables and harnesses as a public PDF. It’s stricter than a car needs and it will still change how you crimp.
One safety note, because our car carried a high-voltage pack: isolate the pack and open the high-voltage circuit before you move anything, and never rework live wiring because a deadline is telling you to. Tired people working next to a battery that can push hundreds of amps is how student teams get hurt.
The thing I’d tell a first-year on any build team: the wiring is not the last 10% of the project, it’s the schedule. Every choice made upstream of it — a water-cooled controller instead of air, an amplifier instead of a self-contained speaker, a sensor that needs conditioning instead of one that doesn’t — gets paid for in wire, at night, by whoever is holding the crimper. That was me, and I found out one afternoon at a time. Car wiring harness planning isn’t a wiring task at all — it’s the design review you should have run three weeks earlier.
If you want the tool that made all of this survivable, it’s the same one every time: a multimeter and the patience to check every junction. The whole season is written up here, one failure at a time.