How to Check Drive-Belt Tension on a Custom Electric Motorcycle

You check drive-belt tension on a belt-driven electric motorcycle by measuring the vibration frequency of the belt span — with a Tension Tester or the Gates Carbon Drive smartphone app — on the lower run, near the midpoint between the sprockets, rather than by pressing the belt with a thumb. The reading only means something against a target for your exact model and belt, so that target has to come from your own vehicle's documentation. The reference case used here is the 2025 Zero S/SR/SR-F/SR-S North America owner manual.

What follows is that manual's procedure plus the reason each step exists, because the steps transfer to almost any belt drive and the reasons tell you which parts of them you are allowed to keep when your build differs.

Why Belt Tension Reads as a Frequency, Not a Squeeze

A belt span clamped between two sprockets is a stretched string, and it behaves like one. Pluck it and it rings at a frequency set by three things: the tension in it, the length of the free span, and the mass each unit of length carries. Add tension and the belt's restoring force pulls it back toward center harder, so each oscillation completes faster — the reason a tight belt of a given length sings a higher note than a slack one. That relationship is what makes a phone microphone a legitimate tension gauge: it measures pitch, and pitch stands in for tension once length and belt mass are fixed.

Your hand is a poor instrument for the same reason. Two bikes can carry identical tension and feel completely different under a thumb, because a longer span or a heavier belt drops the frequency without any change in tension at all. Which means a feel judgement learned on one build does not transfer to the next, while a frequency number does, since it is tied to the geometry it was measured on rather than to the hand that measured it.

One consequence runs through the rest of this procedure: because frequency falls as span length rises, a reading taken at the wrong point on the belt is a different measurement, not a noisier version of the right one.

What the 2025 Zero Manual Actually Specifies

On tools, the manual gives you exactly two options: a Tension Tester, or the Gates Carbon Drive smartphone application. Either one produces a frequency, and either one is the check itself. There is no second, better reading hiding in a force gauge elsewhere.

Where to measure is pinned just as tightly. Lower belt span, close to the center between the sprockets, which rules out the top run and rules out the stretch of belt right beside either sprocket.

Both details are model-bound. They come from one document for one market, and a different custom build may call for another tool or another measurement point, so that vehicle's documentation wins over this one every time.

The Check in Order, From Setup to Reading

Start with the schedule, not the belt. The manual calls for tension to be checked and adjusted at the intervals in the maintenance schedule, so the first thing you settle is whether this check is due. A measurement taken between intervals is a harmless habit, but the number you act on belongs to a scheduled adjustment you can record and compare against the last one.

Then choose your route: Tension Tester or Gates Carbon Drive app.

Then find the lower run, walk to the midpoint between the sprockets, and note exactly where your finger or wrench sits, so the next check starts from the same inch of belt.

Then pluck the belt with your thumb or a wrench so that it vibrates like a guitar string, bring the phone in until the microphone is as close to the belt as you can get it without contact, take the reading, and write it down. Comparison with the target comes after that, not during.

The order is fixed because every step narrows what the number can mean. Get the span wrong, the point wrong, or the vibration damped, and you still produce a plausible-looking figure that answers a different question than the one you asked.

The Four Details That Make or Break the Reading

Location gets fudged more than anything else. Lower span, close to the center, is the instruction, and the reason to respect it is that the two runs of a belt drive do not behave identically, while the ends of a span sit close enough to the sprockets to pick up their stiffness. Pluck beside a sprocket and you measure the hardware as much as the belt.

Microphone placement comes second. Hold it as close as possible without touching, because contact damps the very oscillation you are trying to hear. A finger resting on the belt turns a ringing string into a dull thud, and the app will still report a figure — just not your tension.

Third, pluck rather than push. A sideways shove displaces the belt without setting up a clean oscillation; you want the note, so strike it and let it ring.

Fourth, repeat from the same spot and write both numbers down. Two readings that land close together tell you the method itself is stable, while a wild spread tells you the pluck or the phone moved between attempts.

How to Confirm the Reading Is Actually Right

A number is not an answer until it sits against a target. The manual treats tension as something checked and adjusted to specification, which means the reading's only job is to tell you how far you are from the figure your model calls for.

After that, the road gives you two signals, and they are not symmetrical. The low side is audible: the belt teeth skip over the teeth of the rear sprocket, and you hear it, usually at low speed with load on the drive. The high side is silent, showing up only as wear and drag, so a bike that sounds perfect and rolls freely can still be over-tensioned.

Which leaves a narrow conclusion: silence confirms nothing about the high side, so the number stays the primary check while the sound is a secondary alarm that only fires in one direction.

Ratcheting, Drag, and Why Both Extremes Cost Money

The two failure modes are priced differently, and that asymmetry is the most useful thing in the whole procedure. Ratcheting is loud and consequential: the manual calls the sound unpleasant, notes that the same skip can damage the carbon tensile cords, and states that if ratcheting has occurred the belt should be replaced before the next ride. That last part is a hard rule with a hard bill attached.

Over-tensioning, by contrast, increases wear of the drive system and makes it drag. No noise, no shudder, no warning light — just a slightly slower motorcycle and a shorter service life that shows up as a bill much later.

So if you are choosing where to sit inside the specified range, I would stay away from the firm end rather than chase it. A slightly low belt announces itself with a noise you hear at walking pace; a slightly high one never announces anything, so you pay for it in belt and sprocket life over thousands of miles of not noticing. That preference only holds inside the range your documentation gives — outside it on either side the manual's warnings apply in full, and none of this replaces the target frequency for your belt.

Where the Manual Stops and Your Build Begins

Split the procedure into two parts, the method and the number. The method — lower span, midpoint, pluck, microphone close but not touching, frequency instead of thumb pressure — carries over to any belt drive you can reason about, because it comes from how a stretched span behaves rather than from one particular motorcycle.

The number does not carry over. Target frequency depends on span length, belt mass and the tension the manufacturer wants in that specific belt-and-sprocket combination, which is why you re-derive the target from your own build's documentation instead of borrowing one from someone else's bike.

One maintenance item sits right next to tension and is easy to confuse with it: the manual says to clean the belt with mild soap and water when you wash the motorcycle. Cleaning is not a tension adjustment, and a wash-day glance is not a substitute for the scheduled measurement.

The material for this model gives no target frequency value, so none appears here. That figure belongs to your documentation, and inventing one would be worse than leaving the gap visible.

The One Thing to Confirm First

Before you touch the belt, confirm you actually hold the target frequency or tension specification for your exact belt-and-sprocket combination. Everything else — span location, pluck, microphone position — exists to produce a number, and a number without a model-specific target cannot tell you whether the belt is right, wrong, or merely plausible.