ENGINEERING NOTE · SEISMIC
How a rack passes a seismic test
A cabinet goes onto a shake table with four accelerometers bolted to it. The table measures how much the cabinet amplifies what is fed into it — and that amplification, not the earthquake magnitude, is what decides whether it passes.
CAICT · YD 5083-2005 · intensity 8 and 9 · X and Y directions
The specimen
A test result only means something if the specimen is pinned down.
| Item | Value | Where it comes from |
|---|---|---|
| Product | IT cabinet | Report page 1 |
| Model | TR-61220 | Report page 1 |
| Outline, W × D × H | 600 × 1200 × 2000 mm | Report §2 |
| Cabinet mass, empty | 160 kg | Report §2 |
| Ballast fitted for the test | 300 kg, in five trays of 60 kg | Report §1 |
| Fixing, base | Four M12 bolts into channel steel; channel steel bolted to the shake table | Report §2 |
| Fixing, top | No external connection at the top | Report §2 |
| Frame construction | Fully welded frame; roof and base screwed to the frame and spot-welded; doors on lock and hinge | Report §2 |
Why the ballast matters more than it looks. An empty cabinet is light and stiff; the same cabinet with 300 kg in it has a different mass and answers the shake table differently. The report fixes the ballast at 300 kg so the test is a defined, repeatable condition — not because 300 kg is the right load for your hall. Your rack will be loaded differently, and its response will differ. That is exactly why the report states the condition instead of stating a conclusion about every possible load.
The instruments
Not "we shook it" — a table, a recorder, and four calibrated sensors.
| Instrument | What it does here |
|---|---|
| Seismic simulation shake table, 2.5 m × 2.5 m | Applies the input motion in the horizontal directions |
| Data acquisition and analysis unit | Records every channel at the same time base |
| Four capacitive triaxial accelerometers | Measure in three axes at each of the four positions |
| Calibration — accelerometers | The report lists the calibration expiry date of each unit |
| Calibration — shake table | The report lists the table's calibration expiry date |
| Test environment | Ambient 23–25 °C, relative humidity 40–41 % |
- On the table (1)The input. What was actually fed in.
- On the cabinet base (2)The first place a cabinet can fail — the fixing.
- Mid-height (3)Where a tall cabinet bends.
- On the cabinet top (4)Where the movement is largest, and where amplification shows up first.
In the order it happens
Sweep, shake, sweep again, inspect.
-
Step 1 — Sweep before the test
White noise, 1–35 Hz, at 0.1 g. The cabinet's natural frequency and damping ratio are measured. This is the baseline: what the cabinet looks like before anything is done to it.
-
Step 2 — Apply the earthquake
A synthetic earthquake wave is fed in horizontally, first in X and then in Y, at the level required for intensity 8 under the report's test basis.
-
Step 3 — Sweep again
The same white-noise sweep is repeated and the natural frequency and damping are measured a second time. The change between the two sweeps is the instrument that detects damage the eye cannot see.
-
Step 4 — Inspect and judge
The cabinet is examined for structural damage. Then the level is raised to intensity 9 and the whole sequence runs again, on the same sample.
Input against response
The same shake, two very different answers.
Horizontal X, intensity 8 — peak response divided by peak input
Horizontal Y, intensity 8 — the same input, a much larger answer
| Case | Peak input (g) | Peak response (g) | Amplification | Structural damage | Verdict |
|---|---|---|---|---|---|
| Horizontal X, intensity 8 | 0.660 | 0.889 | 1.35 × | None | PASS |
| Horizontal Y, intensity 8 | 0.660 | 2.159 | 3.27 × | None | PASS |
| Horizontal X, intensity 9 | 1.320 | 1.705 | 1.29 × | None | PASS |
| Horizontal Y, intensity 9 | 1.320 | 4.517 | 3.42 × | None | PASS |
| Vertical Z | — | — | — | — | Not assessed |
Amplification is calculated from the two columns the report records, not stated in the report. The report records the input and the response; the ratio between them is arithmetic on the page.
Why Y answers three times harder
It is not the earthquake. It is where the cabinet sits in the band.
| Case | Natural frequency, pre-test | Natural frequency, post-test | Change | Damping, pre-test → post-test |
|---|---|---|---|---|
| Horizontal X, intensity 8 | 18.359 Hz | 16.016 Hz | −2.343 Hz (−12.8 %) | 9.34 % → 9.04 % |
| Horizontal Y, intensity 8 | 6.250 Hz | 5.078 Hz | −1.172 Hz (−18.8 %) | 7.64 % → 9.30 % |
| Horizontal X, intensity 9 | 16.406 Hz | 15.234 Hz | −1.172 Hz (−7.1 %) | 10.56 % → 13.49 % |
| Horizontal Y, intensity 9 | 5.078 Hz | 4.688 Hz | −0.390 Hz (−7.7 %) | 8.02 % → 12.02 % |
| Vertical Z, intensity 8 | above 20 Hz | — | — | — |
| Vertical Z, intensity 9 | — | — | — | — |
Percentages in the Change column are calculated from the two frequency columns.
Why Y amplifies about three times more than X.
- A cabinet is stiff along its width and much softer across its depth. The X direction runs into a structure that is hard to move; the Y direction runs into one that is easy to move.
- That softness puts the Y natural frequency — 6.250 Hz before the test — down inside the band where the input motion carries most of its energy. Being near that band is what turns a modest input into a large response.
- X sits at 18.359 Hz, well above the energetic part of the band, and barely amplifies.
What the shift between the two sweeps means.
- A frequency that moves after the test is not automatically a failure. It is the whole reason the second sweep is in the protocol: stiffness changes when joints bed in, and the measurement is how you find out whether anything else changed with them.
- Damping rose in every case here — most sharply in Y at intensity 9, from 8.02 % to 12.02 %.
- What the report judges on is structural damage. The inspection found none, in either direction, at either level, and the verdict is PASS. Read the two together: the numbers moved, and the structure held.
What the report concludes
Seismic intensity levels assessed, in both horizontal directions
Structural damage found after either level
Verdict recorded for X and Y at both levels
Test basis stated on page 1: YD 5083-2005. Intensity 8 was assessed first and intensity 9 second, on the same sample. The report carries the signatures of the approver, the reviewer and the lead tester on its first page.
What this report does not say
Five boundaries, stated before anyone has to ask.
- 1One model, one size. The report covers TR-61220, 600 × 1200 × 2000 mm. Other sizes in the range are built the same way, but this report does not cover them, and we will not tell you it does.
- 2Horizontal only. X and Y were assessed. Vertical rows in the report are marked as not assessed — that is what the dash means, and it is not a pass.
- 3One ballast condition. 300 kg, in five trays. A different load is a different test.
- 4A submitted sample. The report states that the sample was supplied by the client, and that the result applies only to the sample the laboratory received.
- 5An equipment-level structural test. It is not a certification of a building, a data hall or a facility, and it does not stand in for any country's construction code.
The complete report, unedited and with the report number intact so your engineer can verify it with the issuing laboratory. Tell us the model you are specifying and we will tell you which reports cover it — and which do not.
Take this to the next supplier
Six things to read before you accept anyone's seismic report.
- 01Is there a report number, and can the issuing laboratory be contacted with it?
- 02Who is named as the test body, and does the report carry an accreditation mark?
- 03What is the sample — model, outline dimensions, mass, and the ballast condition?
- 04What was the input level, and what was the measured response? Read both columns, not just the verdict.
- 05Which directions were assessed, and which are marked as not assessed?
- 06Was the cabinet inspected for structural damage after the test, and is that stated on the page?
The same report, printed page by page, is on the seismic load guide. This note is the reading of it; that page is the evidence.
Reports against the model you are specifying
Ask for the full report.
The complete report, unedited and with the report number intact so your engineer can verify it with the issuing laboratory. Tell us the model you are specifying and we will tell you which reports cover it — and which do not.
