An under desk treadmill is quiet when five things are engineered together: the motor and its drive electronics, the deck-to-frame isolation, the belt-to-deck friction pair, the roller bearings, and the way vibration leaves the machine and enters the floor. A supplier who answers the question with a single decibel figure has not answered it. Noise is a system property, and the number is only meaningful when the measurement conditions travel with it.
This matters commercially more than any other compact-treadmill specification. In a home, a loud machine is an annoyance. In an office, a loud machine is a rejected purchase — and in many cases a product that gets banned by the facilities manager after the trial period, which is a worse outcome for you than a return.
Where the noise actually comes from
1. Motor and drive electronics
A brushed DC motor produces noise from mechanical commutation — brushes running on a commutator — plus the fan or flywheel attached to it. A brushless motor removes the brushes and commutator and switches current electronically, which removes that mechanical contact source. The controller then becomes a factor in its own right: the pulse-width-modulation switching frequency and the quality of the filtering determine whether you hear an electrical whine that rises and falls with belt speed. Two machines with the same motor can sound completely different because of the board driving it.
2. Belt and deck friction
The sliding contact between belt and deck is a continuous broadband source that grows with user weight and speed. Its behaviour depends on the belt weave and backing, the deck material and finish, and — critically — the lubrication regime. A dry running surface is the single most common reason a used unit becomes noticeably louder than it was out of the box, which is also why lubrication guidance belongs in the manual you ship.
3. Rollers and bearings
Roller and bearing quality shows up as a tonal, whirring component rather than a rumble. Bearing grade, roller diameter, concentricity and belt tracking all contribute. A poorly tracking belt chirps against the roller flange — a defect the end user hears immediately, even when the measured average level looks acceptable.
4. Deck-to-frame coupling
How the deck is mounted to the frame decides how much impact energy reaches the structure. Isolation mounts, damping elements and the stiffness of the frame determine whether a footfall is absorbed or transmitted. This is also where the marketing claim “shock absorption” and the acoustic behaviour are the same engineering problem seen from two angles.
5. Structure-borne transmission into the floor
A machine can be quiet in the air and loud in the room below. Rigid contact between the frame and a hard floor transmits vibration into the building structure, which then radiates as low-frequency sound. In a workplace this is often the complaint you cannot fix with a quieter motor, because the noise is coming from the slab rather than the machine.
. When you evaluate a supplier, test in the room you care about, not on a workbench. A machine that measures well on a concrete test floor can disappoint on a suspended office floor.
Why “under 50 dB” is not a specification
A decibel reading without conditions cannot be compared with anything — not with a competing quotation, not with an office background level, and not with your own acceptance test. Worth knowing: because the scale is logarithmic, an increase of about 10 dB is heard as roughly twice as loud, so a 7 dB gap between two suppliers describes a very large difference in experience, while a 1 dB gap is not worth arguing about.
Insist on these conditions before you accept any figure:
- Distance and position — how far from the machine, at what height, and at which side.
- Belt speed — and, if a range is given, a reading at each end of it rather than only at the quietest point.
- Load — an unloaded machine is quieter than a machine carrying a user, so state whether a test load was applied and how heavy it was.
- Surface — the floor type the unit stood on, and whether a mat was used.
- Ambient background level — a reading can never be lower than the room it was taken in; state the background.
- Instrument and class — the sound level meter specification and its accuracy class. IEC 61672-1:2013 is the standard that defines the frequency weightings, including A-weighting, and the Class 1 and Class 2 accuracy classes.
- Metric — average level over a stated period, or a maximum level. They are not the same number.
. Ask for dB(A) as a baseline, because A-weighting is the weighting normally used for this kind of measurement. Ask for the measured figure per model together with the conditions it was measured under — distance, belt speed, load and floor surface — and treat any dB figure quoted without those four conditions as a marketing number rather than a specification.
What a noise failure costs you, line by line
Quietness is a cost driver, not a comfort feature. In each channel the loss shows up differently:
- Return rate (consumer channel): noise is one of the few complaints a buyer can assess in the first ten minutes, which makes it a frequent reason for a marketplace return — and a frequent reason for a review that suppresses conversion for every later buyer of that listing.
- Dealer rejection (office channel): a reseller who places one loud machine into a client site will usually not place a second order, and the lost revenue is the whole range rather than one unit.
- After-sales cost (RMA): noise claims are hard to adjudicate. Without a documented baseline, a claim becomes a negotiation, and the supplier who never recorded a figure ends up either replacing units it cannot fault or refusing claims it cannot disprove. Both are expensive.
- Positioning: if your range is sold on workplace suitability, a quiet specification is what lets you hold a price above the cheapest listing. A machine that cannot support the claim cannot support the price — a point we make in why we do not make the cheapest treadmills.
- Compliance and duty cycle: a unit that is straining thermally is usually also a unit that is loud, because the motor is working outside its efficient band. Duty-cycle figures and acoustic performance tend to degrade together, which is why the definitions in peak versus continuous horsepower belong in the same datasheet as the noise figure.
The request list to send a supplier
Send this as a table and require a written answer to every row. “Not available” is a usable answer; a rounded guess is not.
| Parameter or document | Why the buyer needs it |
|---|---|
| Motor type (brushed or brushless) and rated speed | Determines whether a mechanical commutation source exists at all |
| Controller switching frequency and filtering description | Governs the electrical whine that rises with belt speed |
| Measured dB(A) at a stated distance, at each speed setting | Gives you a comparable number instead of a marketing claim |
| Test load and floor surface used | Unloaded, hard-floor readings are the most flattering and least useful |
| Ambient background level during the test | A reading below the background is not a reading |
| Instrument specification and accuracy class | Distinguishes a calibrated Class 1 meter from a phone app |
| Deck material, belt type and lubrication specification | Predicts how the unit will sound after six months of use |
| Isolation or damping design between deck and frame | Determines structure-borne noise into the floor |
| Recommended duty cycle and thermal cut-out behaviour | Loud machines are often thermally stressed machines |
| Acoustic test report from a laboratory, if one exists | Turns a claim into a document you can hand to a corporate buyer |
| Spare-parts availability for rollers, belts and bearings | Noise often returns as a wear part, not as a fault |
Where a laboratory report does not exist, ask instead for a controlled in-house test protocol and a recorded video, and write your own acceptance criteria around it.
How to verify the claim before you commit
- Ask for a video of the unit running at each speed, with a sound level meter in shot and the distance visible. It is not laboratory data, but it is hard to fake and it exposes gross problems.
- Measure the office background level first. A machine can only be sold as quiet relative to the room it will sit in.
- Order one sample, run it on the floor type your customer will use, and measure at desk height rather than at knee height.
- Measure again after an hour of continuous running. Many machines that start quiet get louder as bearings, belt and motor warm up.
- Record your own numbers and keep them. They become your acceptance criteria for the mass-production run and your evidence base for later claims.
The safety framework behind the product itself is the ISO 20957 series, where Part 1 covers general safety requirements for stationary training equipment and Part 6 adds treadmill-specific requirements and test methods — acoustic performance sits alongside that framework rather than inside it, so do not let a quietness discussion replace the safety file. The document set a corporate buyer typically requests is collected on our audit and certification evidence page.
FAQ
What noise level makes an under desk treadmill suitable for an office?
There is no single pass mark, because suitability is relative to the room. The practical test is whether the machine’s level at desk height, under load, is low enough to be masked by normal office background noise at the neighbouring desk. Ask for the figure with conditions, then compare it against the background level you measure in the actual space.
Are brushless motors always quieter?
They remove one noise source — mechanical commutation — but they do not remove belt friction, bearing noise or floor transmission, and a poorly filtered controller can introduce its own whine. Brushless is generally the better starting point for a quiet specification, but the rest of the machine still has to be engineered for it.
Does a mat really make a difference?
For structure-borne noise, often a large one, because it interrupts the rigid path between frame and floor. It does nothing for airborne motor or belt noise. If your customer’s complaint is that the floor transmits sound downward, the mat and the machine’s isolation design are the relevant variables.
Can I trust a dB figure measured with a phone app?
Use it for a rough comparison between machines in the same room and never as a specification. A specification reading should come from an instrument meeting IEC 61672-1 Class 1 or Class 2, in calibration, with the weighting and time setting stated.
Why do two machines with the same motor sound different?
Because the motor is one of five sources. Deck damping, belt and deck friction, bearing grade and the controller’s switching behaviour all change the result. This is why a supplier’s datasheet should describe the assembly, not just the motor.
Will a quieter machine cost me more?
. Quietness is usually bought with better bearings, better damping and a better controller, so the bill of materials moves — but the return is lower return rates, better dealer retention and the ability to hold a higher retail price in the workplace channel. Treat it as a cost of goods positioned against a margin, not as a premium accessory.
What if the supplier refuses to give measurement conditions?
Treat the absence as a data point. A supplier who publishes a figure without conditions either has not measured it or has measured it under conditions that flatter the product. Both are reasons to require your own sample test before committing to a range.
Next step
Send the request table above to your shortlisted suppliers and request a specification sheet with noise figures and their measurement conditions included. The models this applies to are in our under desk treadmill range, with wear components listed under spare parts; the wider compact programme sits under our walking pad supply programmes, and the office channel route is covered in our sourcing guide for office channels.
Email sales001@treadmillfactory.cn or message WhatsApp +86 18868929496 to request the figures.