《防摔唔係靠厚：MIL-STD-810G 到底測咩》Drop Protection Isn't About Thickness: What MIL-STD-810G Actually Tests

The most effective line in the phone case industry is "military-grade drop protection, 1.8 metres".

It works because it supplies a number. The customer thinks: there is a standard I can compare against, and 1.8 m beats 1.2 m.

The problem is that this number is usually not a standard at all. It is advertising. And the standard most often borrowed for it — MIL-STD-810G — never states how many metres is safe.

Our own drop-protection range (Shield) is published as "2.4–2.8 mm, 32–38 g, referencing MIL-STD-810G, targeting a 1.2 m drop". I rewrote that line several times, and I deliberately kept the words "referencing" and "targeting". This article explains why I won't write "passed".

1. What MIL-STD-810G actually is

Start with the most basic fact, and the one most often got wrong: MIL-STD-810G is not a certificate. It is a test-method manual.

It is a US Department of Defense environmental test standard published on 31 October 2008 (later superseded by MIL-STD-810H in 2019 — more on that below). Its job is to tell you how to test: drops, vibration, high and low temperature, humidity, salt fog, sand and dust, and so on. Each has a method number and procedures. Drop testing is Method 516.6, and what we reference is Procedure IV.

The crucial point: the manual contains no pass line anywhere. It does not say "if it passes, the product is approved". On the contrary, it explicitly requires the user to tailor — that is, the procuring party decides which methods to run, at what severity, under what conditions, and most importantly, what "pass" means.

So "passed MIL-STD-810G" carries zero information unless you also know three things:

• Which test plan? (What was tested, and what was not.)
• Who ran it? (An accredited third-party lab, or the brand in its own office.)
• What was the pass criterion? (Case not broken? Phone not broken? Screen not cracked? Internals still functional?)

Without those three, "military-grade" means "I once used a famous document". And because no body issues a certificate and there is no central register to check, any brand can print "MIL-STD-810G" on a box without having tested anything at all. That is a structural fact about the standard, not an accusation against my competitors.

2. What 516.6 Procedure IV actually does

Procedure IV has an honest name: Transit Drop. As the name says, it simulates equipment being knocked over during shipping, handling and loading — not you dropping your phone on the stairs.

Public descriptions of the procedure broadly agree on the following:

• Twenty-six drops. Covering 8 corners, 12 edges and 6 faces — 8 + 12 + 6 = 26. That is where "26 drops" comes from.
• Impact surface: commonly described as a 2-inch plywood platform over concrete. So not straight onto concrete — there is a plywood layer that absorbs some of the energy.
• Height: the standard's heights come from a table graded by the weight and size of the test item. For the weight class a phone case falls into, the figure the industry most often quotes is 122 cm (48 inches). That is where our "1.2 m" comes from.
• Samples: industry write-ups say the standard permits up to five samples to share the 26 drops. I could not verify that word-for-word in the standard text, so I treat it as an industry reading — but if it holds, one unit may only be dropped a handful of times, which means cumulative damage goes untested.

There is one more awkward detail worth saying: under Procedure IV the test item is often dropped inside its transit case or packaging. That is the point — it is a shipping test. Your phone in your hand does not come in a cardboard box.

Which is why I will not write "passed". What we write is: we reference the 810G test conditions and set a 1.2 m drop as an engineering target. That is a standard I hold myself to, not a certificate.

3. A real case: the UK ASA ruling against Mous, April 2024

If you remember only one section of this article, remember this one. It is public, it has an official document behind it, and the details are unusually instructive.

On 3 April 2024, the UK Advertising Standards Authority (ASA) ruled on a complaint against the phone case brand Mous. The result was Upheld. Complaint reference A23-1216884.

What the ad claimed

The TV ad, seen on 15 October 2023, featured someone saying:

"50 people … throwing their phones in Mous cases … let's go."

Several people then threw their phones into the air at the same time, the cameras followed the phones mid-air, they landed, and the participants picked them up, checked them, found them undamaged and still working, and looked relieved.

What the ASA examined

The ruling runs to several pages and I suggest reading the original. These are the key findings:

• On "50 people": the ASA wrote that "of the 50 participants in the ad, the majority of the cast were comprised of Mous employees, friends and family, apart from three paid actors". Because of who took part and their association with Mous, the ASA concluded the test was not conducted independently and the footage did not represent an objective study of any damage.
• On "no damage": the test data recorded external damage to the screen, rear casing, camera and power/volume buttons, and whether those functions still worked. But the phones were not disassembled to check for internal damage. Since internal damage would affect functionality, the ASA found the data insufficient to show the phones had not been significantly damaged.
• On the three drop-test reports: the first report did not use a phone at all — it tested damage to the case itself, so it was irrelevant to claims about protecting phones. The second and third comprised 26 drops from a height of 120 cm onto marble backed by concrete, with an acceptance threshold stating that "the phone should perform correctly with minimum aesthetic damage".
• On height: the ASA found the test height was significantly lower than what the ad showed, so even where the reports showed the phones still worked, they did not substantiate the impression that the case could protect a phone from a harder impact than a consumer would likely need.
• On untested criteria: the ASA noted that functions affecting a phone's ability to work fully had not been tested — including vibration, Bluetooth, microphone, Face and Touch ID, and NFC for mobile payments.

The outcome

The ASA concluded the ad exaggerated the efficacy of the product and was misleading, breaching BCAP rules 3.1 (Misleading advertising), 3.9 (Substantiation) and 3.12 (Exaggeration). The action: the ad must not appear again in its current form, and Mous was told not to exaggerate the efficacy of its products and to hold test data supporting its claims.

One more detail worth noticing: that ad had been cleared in advance by Clearcast, the UK TV ad clearance body. Pre-clearance waved it through, and it was still ruled in breach. Clearance is not evidence.

What this teaches a buyer

1. "Lots of people threw their phones and nothing broke" is not evidence. Who the participants are decides whether it is a test or a performance.
2. The height in the ad and the height in the test can be far apart. The ad looked like a three-storey throw; the report said 120 cm.
3. "The phone was fine" depends on how you define fine. No disassembly, no vibration, Bluetooth or NFC testing is a different thing from "fully functional".
4. Check whether the report actually tested a phone. The first report tested only the case — the case surviving says nothing about the phone.
5. Regulators do act, but afterwards. You can sell on an ad for two years and then receive a ruling. As a buyer, you cannot wait for the regulator.

4. The physics of a drop: height is not the key number — stopping distance is

This is the technical heart of the article, and it is the thing most marketing avoids: what determines the force on your phone is not the drop height. It is the distance over which it comes to a stop.

Step one: impact speed. Falling freely from height h, the speed is roughly v = √(2gh).

• 0.7 m (table height): about 3.7 m/s
• 1.2 m (our target): about 4.85 m/s
• 1.8 m (the industry's favourite claim): about 5.9 m/s

Step two: a 200 g phone landing at 4.85 m/s carries about 2.4 joules of kinetic energy. Those 2.4 joules have to go somewhere.

Step three, and the crucial one: the shorter the stopping distance, the larger the force. Average deceleration a = v²/(2s) and force F = m × a. Here is my estimate for a 200 g phone dropped from 1.2 m (averages, not peaks):

Two things fall out of that table.

First, thick is not the same as protective. If a thick case is rigid inside and only deforms 1 mm, it differs from a thin case mainly by being 600 g instead of 1,200 g — and both are enough to break glass. What matters is how much room there is to deform, not how thick the shell is.

Second, why corners are so dangerous. Force needs to be spread. Landing flat on the back spreads the load across the whole panel. Landing on a corner squeezes the same energy into a few square millimetres, where local stress can be tens of times higher. That is why engineered corner structures exist, instead of simply thickening the back panel.

One caveat: these are averages. Real peak forces can be two or three times higher, because materials usually get stiffer as they compress. So the table is for comparison and for understanding the mechanism. It does not predict whether your phone will survive.

5. Why engineered geometry matters more than overall thickness

If "room to deform" is the key, then each of the following design features has an obvious job. This is the logic behind our Shield, and it is the part I most want to explain.

• Air-cushion corners. Voids inside the corners give the impact somewhere to collapse and deform, turning a hard-against-hard collision into deformation first. This is the single most important element of the design.
• Raised camera lip. +0.8 mm on the aramid range and +1.0 mm on the leather range. The purpose is not looks; it is to stop the lens being the first point of contact. A scratched camera lens drops the whole phone's value.
• Raised screen lip. Put the phone face-down on a table. If the glass touches the table, the case is not protecting the most expensive surface. You can test this yourself in five seconds.
• A stiff back panel. Its job is to turn a point impact into a distributed load. A soft back panel deforms with the impact and concentrates force instead.
• A compliant inner frame (TPU). This is the absorber. The stiff panel spreads, the soft frame absorbs. Neither alone is enough.
• Grippy sides. The most effective drop protection is not dropping the phone. Friction is a real function, not a description of feel.
• The magnet array. All three ranges use an N52 magnet array aligned to the charging coil. The magnets are rigid embedded parts, so structurally they also act as local stiffening. That is something you can compute. It is not a coincidence.

And here is a trade-off I should admit: we did not make Shield 4 mm thick and 60 g. Technically we could. The result would be a phone you don't want to take out of the house. We chose 2.4–2.8 mm and 32–38 g — the point between "acceptable every day" and "enough room to deform". That is a product decision, not a technical limit, and I own it.

6. What we claim, and what we don't

Time to be explicit, because this industry mostly does the opposite.

What we do claim:

• Shield thickness 2.4–2.8 mm and weight 32–38 g.
• Structure: air-cushion corners, raised camera lip, raised screen lip, stiff back panel, compliant inner frame.
• We reference the MIL-STD-810G (Method 516.6, Procedure IV) test conditions, and a 1.2 m drop is our engineering target.

What we do not claim (and shouldn't):

• We do not write "passed MIL-STD-810G". As section one explains, the phrase has no definition on its own; and right now we have no published third-party laboratory report to show. Writing "passed" without one would be exactly the thing I criticise above.
• We do not promise your phone will not break. Any brand that does, ask it why it isn't insuring your phone.
• We do not use "military-grade" as a selling point. It describes where a test came from, not a level of performance.
• We do not claim 1.8 m or 2 m. If we ever do, the precondition is a laboratory report we can name publicly, with the test conditions written out in full.

What does the 1.2 m target cover? Roughly the everyday cases: a phone slipping out of your hand, or off a table. Standing and using a phone puts it about 1.2–1.5 m up; a phone sliding off a desk edge starts at about 0.7 m.

What it does not cover:

• Stairs, bunk beds, ladders. Above 2 m the impact energy rises with height, not linearly in consequence.
• Asphalt, gravel, tile edges. The harder and sharper the surface, the smaller the contact area and the higher the local stress. Plywood over concrete and tile are entirely different tests.
• Being hit by things. Run over by a car, stepped on, caught in a door — these are not free falls, and the forces can be an order of magnitude larger.
• Cold weather. Some plastics become brittle and lose impact resistance when cold. A winter drop is not the same test as a summer drop.
• A phone that has already been dropped. The internals may already carry micro-damage. It is not a new phone.

7. A practical guide: seven questions for any case's drop claims

These apply to any brand, including us.

1. Which laboratory ran it, and what is the report number? If they can't answer, treat it as untested. This is the single most effective question.
2. Was the test item a phone, or a case? The first report in the Mous case tested only the case. A surviving case says nothing about the phone.
3. What surface did it land on? Plywood over concrete? Marble? Tile? Asphalt? The difference can be several times over.
4. How high, how many drops, at what orientation? Twenty-six drops over 8 corners, 12 edges and 6 faces is the procedure. What the ad shows is a separate matter. They have to match.
5. What was the pass definition? Case intact? Screen intact? Powers on? Was it disassembled to check internally? Were vibration, Bluetooth, microphone and NFC tested?
6. Measure the geometry yourself. Use calipers on the camera and screen lips; press the back panel to feel stiff or soft; look for voids in the corners; put it face-down on a table and see whether the glass touches. All four take five minutes and beat any claim.
7. Read the warranty. If a brand doesn't cover your phone, it isn't prepared to stand behind the claim. This isn't about demanding a payout — it's a sincerity test.

One addition: weighing and measuring are necessary, but they are not how you judge protection. Thickness is a cost, not protection. A 3 mm rigid shell can protect worse than a 2.4 mm shell with corner voids. What you are really looking at is how many places the impact has to go.

8. 810G has dozens of other methods, and most of them are irrelevant to you

Drop testing is only one method in MIL-STD-810G. The manual has dozens: high and low temperature, humidity, salt fog, sand and dust, vibration, rain, fungus, and more.

That has two implications — one that helps brands, one that helps you:

• It helps brands: a company can say it "complies with multiple MIL-STD-810G tests", which sounds impressive, while the severity of each is tailored.
• It helps you: a phone case is a personal accessory, and most of those methods have nothing to do with you. Your phone will not meet salt fog (unless you sail), it won't be tested for fungus (Hong Kong is humid, but fungus testing is about long-term damp storage), and it won't face gunfire shock (methods like 519).

If a case brand uses high-temperature testing (around 501.5), low-temperature testing (502.5), humidity (507.5) or salt fog (509.5) as a selling point, ask one question: "How close are those test conditions to my daily life?" The answer is usually: not close. Those methods were written for military equipment operating in extreme environments, not for a phone in a jeans pocket.

We reference exactly one method: drop (516.6 Procedure IV). Not because the others are bad, but because I don't want to build credibility out of tests I consider irrelevant myself.

9. One sample tells you nothing about probability

This is the most overlooked and most important point in the engineering.

"Passed the drop test" usually means: "we took one or a few units, dropped them 26 times, and nothing broke."

What that sentence does not say is: if we made ten thousand of these, how many would fail?

With one sample, statistics only allow you to say "at least one passed". You cannot say "99% pass rate". To establish a pass rate you need tens to hundreds of units — prohibitively expensive for a phone case brand, which is why the industry mostly doesn't do it. That is reality, not conspiracy.

Drop results are also very sensitive to which phone is inside. The same case on a different model has a different weight distribution, camera bump and body stiffness. Even the same phone with a micro-crack at the edge of the glass changes the outcome.

So the correct reading of any "passed the test" claim is: "this design went through one successful engineering verification" — not "this product carries a guarantee of protection". Those are very different statements.

10. The case is only half the system

Many people think "buy a good case and the phone is safe". In practice, protection is decided by the case and the phone together.

• Your phone's own stiffness. A glass sandwich phone flexes when it lands. Flex it far enough and the glass cracks. A case can reduce the flex; it cannot make the phone rigid.
• The camera module. On modern phones the camera bump is a hard point. If it sits proud of the case's raised lip, it becomes the first contact point and all the energy concentrates on one lens ring.
• The edge of the glass. The weakest part of a screen is its edge, where micro-cracks and stress concentration live. Corner drops are the most dangerous because they hit that edge directly.
• How you use it. Keys in the same pocket, sitting on it in a back pocket, using it while walking — those do more damage over time than a single drop.

In one line: a good case can turn "this would have broken" into "it didn't", but it cannot change the fact that some phones are fragile to begin with. If a brand tells you that fitting its case makes you invincible, it is selling confidence, not engineering.

11. Testing it yourself, without sacrificing a phone

If you want to check by hand, you don't need to risk a real phone. This is what I do:

1. Face-down on a table. Does the glass touch? How much clearance is there? Measure with a stack of paper or calipers — you want at least 0.5 to 1 mm.
2. Back-down on a table (camera side). Does the camera ring touch? Measure the clearance the same way.
3. Press the corners. Push hard with a thumb on all four corners and feel for the springiness of a void. A fully solid corner has nowhere to send the impact.
4. Flex it (gently). Twist the case lightly with both hands to judge the back panel's stiffness. A soft back panel concentrates impact.
5. Use a stand-in for a drop test. If you really want to try, take an old phone you don't use, or a wooden block of similar weight (around 200 g), put it in the case, and drop it from table height (0.7 m) onto a hard floor. Don't look for "did it break" — look at whether the case edge deformed, whether the corners collapsed, and whether it still fits the phone afterwards.
6. Record the conditions. Height, surface, angle, number of drops. A test without conditions is a rumour.

To be clear: your own test will never equal a lab test — you have no standard specimens, no fixture, no instrumentation. But it is more than enough to tell "seriously engineered" apart from "just made thicker".

12. Three sentences to finish

One: MIL-STD-810G is a test-method manual, not a certificate, and it contains no pass line. "Military-grade drop protection" only means something alongside who ran the test, what was tested, and what pass meant.

Two: What determines the force on your phone is stopping distance, not height. Going from 1 mm to 5 mm takes the average force from roughly 240 kgf down to roughly 48 kgf — and that comes from structure, not from piling on thickness.

Three: We write "referencing" and "targeting 1.2 m". We don't write "passed", because as of today we have no third-party report to point at. You should know that before you buy. I would rather you buy knowing, than buy believing.

I won't repeat the structure and numbers for the three ranges here.