Motor Controller Residual Voltage: How We Scored Zero on Day One of the Competition

Day one of the competition went well. That is the part that still bothers me.

We passed scrutineering. We ran the slalom. We ran the acceleration event. Nothing broke, nothing caught fire, and the car we had spent a year fighting with did exactly what it was supposed to do. Then the results table went up at 9 p.m. and our slalom line was a zero.

Not a bad time. Zero. The cause was motor controller residual voltage — 20 volts still sitting in the controller ten minutes after we had shut the car off. I am the electrical systems lead on this student-built EV race car, so that zero has my name on it.

The short version: the DC-link capacitor inside a motor controller stays charged after you cut power. Ours held 20 V indefinitely. The competition’s energy meter measures current against a zero point, and a meter that is not actually at zero when it zeroes itself reports garbage. The rulebook’s answer to garbage was simple: no score.

Day One: Scrutineering, a Slalom Run, and an Acceleration Run

We had cleared scrutineering the afternoon before, but only barely, and not because of anything glamorous. Our speaker module had come loose in transit and the wire had snapped, so the car was making noise with the pedal untouched. Four hours of resoldering and begging parts off other teams, and we got through inspection with the inspector being generous about the clock.

So day one was the first time all year the car was simply allowed to be a race car. Slalom first.

Then the acceleration event. This is the same car that, a month earlier, was pulling 30 amps at full throttle while we blamed the gear ratio. On day one it was running with the battery current limit at 185 A, roughly 9 kW, right up against the competition’s 10 kW ceiling.

We packed up feeling like a team that had finally arrived. Then the results came out.

Why Motor Controller Residual Voltage Zeroes Your Score

That year the organizers introduced an energy meter for the first time, together with a new rule capping every car at 10 kW. Every team ran one, and the meter had to read zero before your run counted.

We knew about the 10 kW cap — we had tuned the controller around it a week earlier at Jamsil. And here is the part I would like to skip: at that Jamsil test, the meter was already resting at 20 V. I saw it. I decided it was the meter’s default and went back to the thing I actually cared about, which was making the car fast.

Nobody on our team asked why a car that was switched off was reporting 20 volts.

That is the whole failure, and it happened three weeks before anyone scored anything. The bug was visible, in a number we looked at, and we filed it under “probably fine.”

We Thought It Was a Scoring Error. So Did Everyone Else.

Our first theory that night was that the results table was broken. It felt reasonable, because a lot of teams had zeroes. It was the rule’s first year, the organizers were still working out how to enforce it, and half the paddock did not fully understand what the meter was doing.

So we did the comfortable thing and assumed the problem was upstream of us. We asked. We asked again. Eventually an official told us plainly: your energy meter was never zeroed, so your run does not count.

Then came the part that actually hurt. The organizers pulled the energy meter logs, checked them, and wiped them. No data left to look at. We could not debug from the record, because there was no record — only the verdict.

Sleep was optional that night. A car that cannot score is a car that goes home.

Ten Minutes After Shutdown, It Still Read 20 Volts

I put a multimeter across the controller’s B+ and B− terminals — the same tracing routine I use on any wiring problem, except this time I was hoping to see nothing.

20 volts. Ten minutes later, still 20 volts. It was not draining at all.

That is when it stopped being a scoring dispute and became an electrical problem, because a number that does not move is not a fault you can wait out. Inside the controller is a DC-link capacitor, a bank of capacitance across the power terminals that absorbs switching spikes. Cut the power and the motor stops, the lights go out, and that capacitor sits there holding charge with nowhere to send it. There is no bleed path unless somebody builds one.

What we tried over the next two hours, roughly in order of dignity:

1. Turned everything off and back on again. Still 20 V.
2. Disconnected the pack and waited. Still 20 V.
3. Ran a wire from the terminal to the ground. As in the actual ground, the floor of the paddock.

Charge needs a path, not a destination. Bare concrete is not a path. Two hours of that.

Then someone spotted a 1 kΩ resistor on the table. We soldered a lead onto each end, held it across B+ and B−, and watched the multimeter fall: 20 V, 12, 6, 2 — about a minute to zero. That was the entire fix. A part worth less than a stick of gum, three feet away the whole night.

We carried that resistor in a pocket for the rest of the competition and discharged the controller after every single run, before the meter was zeroed. It looked ridiculous. It worked. We scored from then on, and we made the semifinal.

If Your Competition Makes You Run an Energy Meter

Every motor controller has a DC-link capacitor sitting across its bus, and key-off does not empty it. If your event zeroes an energy meter before each run, that leftover charge stops being an electrical curiosity and becomes a scoring problem. Here is what we do now, and what I would tell any team arriving at their first event with a meter bolted to the car.

1. Measure the bus after shutdown at every test, not just at the competition. Multimeter on B+ and B−, DC volts, wait five minutes, measure again. If the reading doesn’t fall, there is no bleed path, and it is not going to appear on its own.
2. Carry a bleeder resistor, and treat it as a race part. Ours was a 1 kΩ resistor with a lead soldered to each end. It took our 20 V down to zero in about a minute. It cost less than a stick of gum and it was sitting on a table three feet away while we panicked for two hours.
3. Never short the terminals with a screwdriver, and never run a wire to the actual ground. A dead short welds tools and wrecks the capacitor. Bare concrete does nothing at all. Charge needs a path, not a destination — we proved that the slow way, in a paddock, at night.
4. Discharge, zero the meter, then photograph the reading before every run. The organizers pulled our meter logs, checked them and wiped them. There was no record left to argue from — only the verdict. If the zero isn’t in your own photos, as far as the results table is concerned it never happened.
5. Read whatever is new in the rulebook first. The energy meter and the 10 kW cap were introduced that year, and half the paddock had zeroes. New rules are where the zeroes live, because nobody has failed at them yet and written it down for you.

One thing that scales past scoring: our pack sat at 48 V and held 20 V after shutdown. If your team runs a proper high-voltage tractive system, the same capacitor is holding hundreds of volts, and the discharge procedure is not a thing you improvise at 11 p.m. with a resistor you found on a table. Service disconnect, insulated gloves, a written procedure, a second person watching. Our worst case was a zero on a scoresheet. Yours might not be.

And the lesson this whole season kept teaching me, in different fonts: a weird number you decide to ignore is a bill you have agreed to pay later. We saw 20 V three weeks early. We shrugged. It cost us every point available on day one.

We got the meter zeroing properly by mid-morning on day two, in time for the endurance race. Somewhere around the middle of it, the water cooling stopped working and the controller — our third one that season — started climbing toward the temperature where controllers stop being controllers.

Everyone told me not to open the pump.

Series: This is part of Field Notes — everything that broke on our Formula Student EV car, in the order it broke.