Volume 1 › Chapter 1 › 1.4
Chapter 1: What a Vibration Test Is For

1.4 The Standards as Vocabulary, Not Subject

about 6 minutes · 2 figures, 1 interactive · quiz of 5 · Slides

After this section you can: name what a standard gives the lab and what it leaves to the lab, and say what to do when the standard and the machine disagree.

The standards are where the words come from: grms, tolerance, control channel, duration, category, method. A technician who can read the standards can read any test plan. Nobody needs to memorize them.

Three of them account for most of what a lab meets. MIL-STD-810 is the military environmental standard, and tailoring is its whole point; Method 514 (a method is the standard's numbered chapter for one environment) covers vibration, Method 516 covers shock, and the standard's Part One says in plain words that the default curves are for when no data exists. RTCA DO-160 covers airborne equipment; its Section 8 is vibration, with categories (classes set by where the box sits on the aircraft). NASA's General Environmental Verification Standard (GEVS) covers spacecraft components and is the usual source of their qualification and acceptance curves. Other industries have their own: automotive, rail, shipboard, seismic. The shape is always the same.

Every standard gives the lab four things: tolerances on the profile (how far the control channel may depart from the profile before the run is out of tolerance), a rule for where control is measured, a duration, and a list of what the report must contain. The standards agree on more than people think, and they differ on exactly the things that decide a test: the width of the tolerance band, where the control accelerometer goes, and how long the test runs.

The controller reads the control channel, usually an accelerometer on the fixture at the article's interface, compares its spectrum with the profile, and adjusts the drive to the amplifier until they match. A standard says whether control is one accelerometer or several averaged.

Every standard leaves the same things to the lab: the fixture, the mounting, the sensor placement, the notch and limit decisions, and what to do when the run does not go to plan. A notch lowers the profile over a narrow band, usually at a resonance, by agreement with the article's owner. A limit puts a ceiling on a response channel, and the controller lowers the drive when that channel reaches it. The fixture and the mounting are the rest of this volume; the sensor placement, the notch and limit decisions and the run that does not go to plan are Volume 2's subject.

Decide. The standard tells you what to prove and how to document it. It does not run the test. When the standard and the machine disagree, when the shaker cannot deliver the profile or the fixture cannot hold the tolerance (usually because it resonates in that band) or the schedule cannot afford the duration, the standard does not change. The plan does, in writing, before the run, with the customer's name on the deviation. The deviation names the requirement paragraph it departs from and says which band, where, and by how much, the way Figure 1.7 marks the band against the tolerance; it gives the reason and what the departure does to what the test proves. The customer's name on it is the name of whoever owns the requirement, whether that is an outside customer, a prime contractor or the lab's own program office; how a deviation is written is Volume 2's subject. The machine's arithmetic in Chapter 10 finds most of these before the article arrives; the low-level run finds the rest. When it is the low-level run that finds one, the test holds at low level: the shortfall goes in the log, the test engineer takes it to the customer, and nothing goes to full level until the plan changes.

The profile with its tolerance band, and the run

Amber: the profile. Blue: the as-run control trace. Gray: the tolerance band, a limit on the curve.

The same run under three standards' tolerance bands

Toggle the bands. Solid, dashed and dotted bands are three different standards judging one run.

Duration.
sequence step
the test
A long exposure at a low level became a short, hotter test. Where the control accelerometer sits decides which curve the tolerance is judged on; the controller holds that one point to the profile, and the rest of the control loop is Volume 2's subject. The three standards' bands are illustrative; the published tolerances for each standard replace them before this figure is final.
Figure 1.7. The profile from Figure 1.5 with its tolerance band, an as-run control trace under it, and the same trace judged under three standards. Same flight, same envelope, same run, different verdict by document. Below, a bar of the hours the article spends in that step in service against the minutes it spends on the shaker: the duration scaling of 1.3. VD
Tolerance bandControl channel ruleDuration ruleReport contents
MIL-STD-810a band on the curve plus a limit on the overall level; wider above a named frequencynamed control location; tailoring may move itderived from exposure with time compression; tailorabletailoring record, deviations, data
RTCA DO-160a band on the curve by categoryat the fixture interfacefixed per category and test typecategory, curve, deviations
NASA GEVSa band on the curve plus overall levelat the mounting interface, average of control channels allowedfixed per test typelevels, durations, signature comparisons

Words, not numbers. The numbers are in the standards and change by revision; the shape of what each asks for does not.

Figure 1.8. MIL-STD-810, RTCA DO-160 and NASA GEVS side by side on the four things that decide a test: tolerance band, control channel rule, duration rule and report contents. Read from MIL-STD-810H, RTCA DO-160G and GEVS revision B (GSFC-STD-7000B), the revisions current when this was written; the tolerances move between revisions, so check the one the requirement cites.

Common Questions

How wide is a tolerance band, in numbers?

The chapter gives the shape and not the numbers, because the numbers are the standard's and they change from one revision to the next. Figure 1.8 sets the three standards side by side in words and names the revisions it was read from. The band that applies is the one in the revision of the standard the requirement cites, read from the standard itself.

When the customer's requirement and a standard disagree, or the plan cites no standard, which governs?

The requirement does, because it is the document that names the standard in the first place (1.5, The requirement); a standard governs a test only through the requirement that cites it. If two standards would give different verdicts on the same run, the one the requirement cites is the one that applies. If the plan cites no standard, the lab asks before it schedules, the same as for any other gap in the requirement.

Can I add a notch or a limit during the run on my own judgment?

Not on your own. A notch changes the profile, and it is set by agreement with the article's owner; the limits are named in the procedure before the run (1.5, The procedure). A change that comes up during the run is a decision for whoever the procedure gives that authority to, and it is recorded when it happens. Who holds that authority, and how a run that does not go to plan is handled, is Volume 2's subject.

Does the standard's list of report contents become my report template?

The standard's list is the floor, not the whole report. The report also has to carry the requirement and its definition of pass, the procedure as run, the result against that definition, and every deviation with who decided it and why (1.5, The report). Reading and writing a report are Volume 2's subject.

Quiz: Try It

Open book. 5 questions drawn from this section's pool. 4 correct passes the quiz. Nothing is recorded here; in the course your passing attempt opens the next section.