---
title: "How This Tool Works — Methodology, Physics & Known Limits"
description: "Exactly what the cleaning tool does, which claims are established physics, which are engineering convention, and what we have not tested. Every figure sourced."
url: https://fixmyspeaker.me/methodology/
updated: 2026-09-14
---

# How this works

Written and maintained by [Muhammad Usama](https://fixmyspeaker.me/author/muhammad-usama/) · Last checked: 2026-09-14

This page is the audit trail for everything else on the site: what the tool actually does, which claims are established physics, which are engineering convention, and what we have not tested. If you only read one page here to decide whether to trust the rest, read this one.

**The short version.** Sound-based cleaning is a real but narrow mechanism, well supported by basic acoustics and used by Apple in a shipping product. The specific frequency everyone uses — 165 Hz — is a convention, not a derived constant. We have not run laboratory trials and do not publish success rates, because we have not measured any.

## What the tool does, precisely

Each mode generates a sine wave in your browser using the [Web Audio API](https://www.w3.org/TR/webaudio/). No audio file is fetched; the oscillator is created locally and the frequency sweep is scheduled on the audio clock. The exact parameters:

| Mode | Range | Sweep period | Amplitude modulation | Duration |
| --- | --- | --- | --- | --- |
| Water Eject | 158–172 Hz | 2.2 s triangle | none | 30 s |
| Dust Removal | 200–400 Hz | 3.5 s triangle | 7 Hz tremolo | 30 s |
| Deep Clean | 100–260 Hz | 6 s triangle | none | 45 s |
| Smart Clean | 140–380 Hz | 9 s triangle | 3 Hz tremolo | 60 s |

Output is capped at 0.85 of full scale with a 120 ms fade in and out. The fade is not cosmetic: an abrupt start puts a step transient through the driver, which is audible as a click and is harder on the voice coil than the tone itself.

## Which claims are established, and which are not

This is the distinction most pages in this category blur. We keep it explicit.

### Established physics — cited

-   **Cone excursion is inversely proportional to the square of frequency** for a given sound pressure level. Halving the frequency requires four times the excursion to produce the same output. This is why low tones move the diaphragm further, and it is a standard result in loudspeaker design, not an opinion.
-   **Water is roughly 830 times denser than air** (≈997 kg/m³ against ≈1.2 kg/m³ at room temperature). A film of it on the diaphragm therefore acts as significant added mass, which damps motion and attenuates high frequencies first — the muffled quality people describe.
-   **The surface tension of water is about 72 mN/m at 20 °C.** This is the force that has to be overcome to break a film away from the grille mesh, and it is why the effect is mechanical rather than magical.
-   **A fixed tone in a small cavity forms a standing wave** with pressure antinodes and nodes. Material sitting at a node experiences very little force. This is the reason every mode here sweeps instead of holding one value.
-   **Sound-based water ejection is a shipping feature.** Apple's Water Lock on Apple Watch clears the speaker cavity with a series of low-frequency tones when you unlock it. That is manufacturer documentation, not folklore.

### Engineering convention — labelled as such

-   **165 Hz specifically.** It spread through the iPhone Shortcuts community and propagated by copying. It lands in a useful band — low enough for large excursion, high enough for a small driver to reproduce efficiently — but it was not derived. Anything from roughly 150 to 180 Hz behaves similarly.
-   **"165 Hz is the resonant frequency of a phone speaker."** Widely repeated and not true. Resonance depends on driver size, cone mass and chamber volume, and varies between models. We do not make this claim anywhere on the site.
-   **200–400 Hz for dry debris.** The reasoning — that unbonded particles respond to rate of direction change rather than peak force — is sound, and the pulsed delivery follows the same logic as tapping a stuck lid. We have not measured the optimum.
-   **30-second cycles.** Chosen as a balance between giving the mechanism time to work and limiting sustained thermal load on the voice coil. Not an optimised figure.

### What we have _not_ done

-   We have not run controlled trials or measured ejection efficiency.
-   We have not tested a matrix of devices under repeatable conditions.
-   We do not publish success-rate percentages, because we have not measured any. Where you see figures like "89% success" or "tested across 47 models" on competing sites with no published method, treat them as marketing.
-   We have not measured the resonant frequency of any specific handset.
-   We make no medical, warranty or repair-outcome guarantees.

## Where our advice matches the manufacturer

The strongest check on advice in this category is whether the device maker says the same thing. On the main points, Apple's own liquid guidance agrees with ours:

| Our advice | Apple's published guidance |
| --- | --- |
| Do not use rice | "Don't put your iPhone in a bag of rice" — rice particles can damage the device |
| Do not use a hairdryer | Do not dry using an external heat source |
| Do not use compressed air | Do not dry using compressed air |
| Do not poke anything into ports or grilles | Do not insert a foreign object such as a cotton swab or paper towel |
| Tap it out with the opening facing down | Tap the device gently against your hand with the port facing down |
| Do not charge until dry; respect the moisture warning | Wait before reconnecting power; allow to dry |

We reached those positions independently and then checked them. Where a manufacturer contradicts us, we change the page.

## On water resistance ratings

One detail worth correcting, because it is misreported almost everywhere. Under **IEC 60529**, IP67 has a defined test — immersion at one metre for thirty minutes. **IP68 does not have a single universal condition.** The standard only requires that protection exceed IP67; the actual depth and duration are declared by the manufacturer. So "IP68" on two different phones can mean two different things, and neither means the speaker cavity stays dry — the cavity is open to the air by necessity, because sound has to escape.

## On hearing safety

We treat this as the real risk on the site, and the numbers are not ours. NIOSH sets a recommended exposure limit of **85 dBA over eight hours**, halving the permitted duration for every 3 dBA increase. The WHO states that levels below 80 dB are unlikely to cause damage. A phone at full volume against the ear can exceed both.

That is why the instruction to keep the speaker away from your head is stated on every page that plays a tone, and why we tell people never to wear earbuds while a cleaning tone runs — a sealed ear canal lets none of that energy escape.

## How this page is maintained

Every technical figure above is traceable to a source on our [sources page](https://fixmyspeaker.me/sources/). If a manufacturer updates its guidance or a figure here is wrong, the correction process is described in our [editorial policy](https://fixmyspeaker.me/editorial-policy/) — and corrections are welcome at [our contact address](https://fixmyspeaker.me/contact/).

Last checked: 2026-09-14

## Frequently asked questions

### Have you lab-tested this tool?

No, and we do not claim to have. We have not run controlled trials, measured ejection efficiency, or tested a device matrix. Any site quoting a success rate like "89% effective" or "tested on 47 models" without publishing its method is quoting a number it made up. What we do instead is state the physics that is established, cite it, and label everything else as convention or estimate.

### Is 165 Hz scientifically derived?

No. It is an engineering convention that spread through the iPhone Shortcuts community and was copied by every tool built since, including this one. It works because it sits in a useful band for small drivers, not because anyone calculated it. The claim that it is "the resonant frequency of a phone speaker" is false — resonance varies by driver size, cone mass and chamber volume, so no single number describes all phones.

### Why do you publish what the tool cannot do?

Because the failure cases are the most useful part. A reader with a torn diaphragm who runs twenty more cycles has wasted their evening and still has a broken speaker. Telling them to stop is more valuable than another paragraph of encouragement, and it is the part competitors leave out.

### How do you decide what counts as a fact here?

A claim is stated as fact only if it traces to a primary source — manufacturer documentation, a published standard, or an established physical constant. Anything else is labelled as convention, inference or estimate in the text itself. The full list of sources is on our sources page.

### Do you accept sponsorship or affiliate commissions?

No. There are no ads, affiliate links, sponsored placements or paid reviews on this site, and no analytics or tracking. That is stated in the editorial policy and is easy to verify — view the page source and you will find no third-party requests at all.

## Sources cited on this page

1.  [If you see a liquid-detection alert on your iPhone](https://support.apple.com/en-us/102643) Apple Support Cited for: rice, heat, compressed air and foreign objects; recommended drying procedure.
2.  [How to use Water Lock and eject water from your Apple Watch](https://support.apple.com/en-us/108352) Apple Support Cited for: sound-based water ejection as a shipping manufacturer feature.
3.  [IEC 60529 — Degrees of protection provided by enclosures (IP Code)](https://www.iec.ch/publications/iec-60529) International Electrotechnical Commission Cited for: IP67 defined test conditions; IP68 being manufacturer-declared.
4.  [Understand Noise Exposure — recommended exposure limit 85 dBA over 8 hours](https://www.cdc.gov/niosh/noise/prevent/understand.html) US CDC / NIOSH Cited for: 85 dBA over eight hours, 3 dBA exchange rate.
5.  [Deafness and hearing loss: Safe listening](https://www.who.int/news-room/questions-and-answers/item/deafness-and-hearing-loss-safe-listening) World Health Organization Cited for: levels below 80 dB unlikely to cause hearing damage.
6.  [Web Audio API specification](https://www.w3.org/TR/webaudio/) W3C Cited for: how the tone is generated locally in the browser.