Chatter, Natural Frequency and What the Base Can Fix

Why chatter is a structural dynamics problem, how the base contributes, and which of stiffness, damping and frequency separation is worth changing first.

Chatter is self-excited vibration: the cut excites the structure, the structure moves, the movement changes the chip thickness, and the changed chip thickness feeds the vibration. Once the loop has more energy going in than damping takes out, it grows until something gives.

Understanding that loop tells you which fixes are real and which just move the problem.

The three levers

Only three properties of the structure matter, and they are not equally useful.

Stiffness sets how far the structure moves for a given force. More stiffness means less movement per unit of cutting force, so less modulation of chip thickness. Genuinely helpful, and usually the most expensive lever to move once a design is committed.

Damping sets how fast energy leaves the system. This is the one that decides stability, because chatter is a question of whether energy accumulates. Damping is why cast iron beds outperform bare steel weldments of identical geometry.

Frequency separation is keeping the structure’s modes away from what the process excites. Cheap when you can do it, and it is a relocation rather than a cure: the mode is still there and still undamped.

Mass is often listed as a fourth lever. It is really a way of buying the other two: mass lowers amplitude and shifts frequencies down, but on its own it removes no energy at all.

Find out what is actually moving before you spend money

The compliance you care about is the whole loop from tool tip through spindle, column, ways, base, foundation, fixture and workpiece back to the cut. Chatter happens at whichever element in that loop is most compliant and least damped at the relevant frequency.

On a well-designed machine that is very often not the base. A long thin end mill, a slender boring bar, or a poorly clamped workpiece will typically be an order of magnitude more compliant than the structure. Replacing a machine base to compensate for a tool overhang is an expensive lesson.

So measure first. A tap test with an instrumented hammer and an accelerometer, done at the tool tip and again on the base, takes an hour and tells you where the compliance is and roughly how damped it is. Every decision after that is informed rather than guessed.

What the base is responsible for

The base earns its keep in specific ways:

• Providing a stiff, damped foundation for everything above it. If the base is soft, every element above it moves whether or not it is itself stiff.
• Keeping its own modes out of the operating range. A first mode that lands on a tooth passing frequency you use every day is a design fault.
• Not amplifying. A large undamped panel in a base can behave like a loudspeaker cone, taking vibration from elsewhere and making it worse.
• Coupling properly to the floor. Levelling, grouting and isolation are part of the structural loop, not an installation afterthought. See levelling and grouting.

Which fix, in which order

Work through this before redesigning anything:

1. Measure. Tap test at the tool and the base. Establish where the compliance is.
2. Fix the tool loop if that is the culprit. Shorter overhang, bigger shank, better holder, different flute count. Cheapest by a wide margin.
3. Change speed. Because stability lobes exist, a higher spindle speed is often more stable than a lower one. Counter-intuitive and free.
4. Change the workholding. A part ringing in a fixture is frequently misdiagnosed as machine chatter.
5. Check the installation. Loose levelling screws, a missing grout pad or a foot not bearing can turn a good machine into a bad one. Look before you redesign.
6. Add damping to the base. Filling the sections is the cheapest structural change.
7. Change the base material. Cast iron or mineral cast, with the tooling cost that implies.
8. Redesign the structure. Deeper section, better ribbing, different support points.

Items one to five cost almost nothing and resolve a large share of chatter complaints. Items six to eight are real engineering with real money attached and should follow measurement, not precede it.

Designing a new base with this in mind

If the base is still on the drawing board you can do better than fixing it later:

• Run modal FEA early and look at mode shapes, not only frequencies.
• Map your machine’s excitation frequencies across the whole operating range and place the first mode clear of them with margin. Targets above 80 to 100 Hz are commonly quoted for machine tool structures.
• Design for damping from the start if the machine is finish-oriented. Retrofitting damping is harder than building it in.
• Decide the support scheme at design time. It is part of the structure, and letting the installer decide it means letting the installer decide your dynamics.
• Budget a tap test on the first unit. The measurement that validates your model costs less than the meeting where everyone argues about why the machine chatters.

See also: the natural frequency reference for base design.

Common questions

← All guides