Why clamping-force evidence comes first
A CNC fixture does not fail because it looks cheap. It fails because the part moves. On a three-axis mill that shows up as a chatter mark on one face; on a five-axis job it shows up as a positional error that only appears in the fourth or fifth operation, after most of the workpiece value has already been cut. In both cases the root cause is usually the same: the clamping system could not repeatedly deliver the force the cutting process assumed it would.
That makes clamping force a better entry point for supplier evaluation than catalogue specifications such as jaw width, opening range or material. Jaw geometry is visible and easy to copy; force delivery, force verification and force degradation over a production year are where competent fixture builders separate themselves from assemblers who buy standard vises and plate them.
This guide sets out the questions to ask, the documents to request and the tests to witness before you place a fixture order. It is written for buyers sourcing machined workholding — dedicated fixtures, modular tombstone setups, hydraulic or pneumatic clamping systems — where part position accuracy has a direct cost when it drifts.
Start with the process, not the product
Before a supplier can demonstrate credible clamping capability, they need to understand what the fixture has to survive. A supplier who quotes a fixture without asking about cutting forces, cycle time and workpiece rigidity is describing a product, not solving a problem. Send the following in your RFQ and see whether the response engages with it:
- Material and stock form — alloy, hardness range and whether parts arrive as castings, forgings or billet.
- The heaviest cutting operation — tool diameter, radial and axial depth of cut, and whether the operation is roughing or finishing.
- Spindle and machine interface — taper type and whether the machine is three-, four- or five-axis, because multi-axis setups change the direction of the cutting force relative to the clamping direction.
- Where accuracy is consumed — the tolerance band that actually matters downstream, and which operation it is generated in.
- Production volume and changeover pattern — a fixture for 50 parts a year and one for 50,000 parts a year need different wear and maintenance assumptions.
An answer that returns these points as a written clamping and locating concept is a stronger signal than any certificate, because it shows the supplier can reason about force distribution rather than reprint a datasheet.
Ask for a force budget, not a force number
"Clamping force: 40 kN" tells you very little on its own. What matters is whether that number was derived, and whether the supplier can show how it survives the trip from the hydraulic unit to the workpiece. Ask for a short force budget covering:
- Required holding force derived from the heaviest cut, with the friction coefficient at the contact faces stated and the assumed safety margin given explicitly.
- Losses between source and clamp — pressure drop across hoses, rotary unions and manifolds; internal leakage in cylinders; and deflection in the fixture body that absorbs force before it reaches the part.
- Force distribution across multiple clamps on the same circuit, since the clamp nearest the pump typically sees more pressure than the far end unless the design compensates.
- Thin-wall and distortion cases — where the supplier explains how force is reduced or spread rather than simply applied harder.
You are not looking for a perfect calculation. You are looking for evidence that someone did one. A supplier who can walk you through the friction assumption and the safety margin is telling you how they will behave when a part fails in your shop.
Request the verification method, not just the guarantee
A guarantee that the fixture "holds to the required force" is not verifiable at your receiving dock. Ask instead how the force is confirmed before each cycle. Mature designs use one or more of the following, and the supplier should name which one they propose and why:
- Pressure sensing in the hydraulic or pneumatic circuit, with the sensor positioned so it reflects clamp pressure rather than pump output. Ask where the sensor sits and what a blocked or leaking line does to the reading.
- Force measurement at the clamp using a load cell or a calibrated force gauge during setup and commissioning, rather than continuous in-cycle monitoring.
- Position- or travel-based confirmation, where the clamp only signals "closed" once the jaw has reached the end of its stroke against the part.
- Interlock into the machine control, so a cycle cannot start when the confirmation signal is absent. This is the difference between a measurement and a safeguard.
For each method, ask for the calibration interval, the drift observed in service, and what the operator is expected to do when a reading falls out of band. A fixture with a sensor and no defined response to an out-of-band reading gives a false sense of control.
Separate locating accuracy from clamping repeatability
Buyers routinely merge these two numbers in a single line of a specification, then argue about it at first article. Keep them apart in the evaluation:
- Locating accuracy is a property of the fixture's datum surfaces — pins, rest pads, stops — relative to the machine and the part drawing. It is largely fixed at build and confirmed by measurement.
- Clamping repeatability is a property of the clamp's behaviour over thousands of cycles — jaw wear, seal wear, chip ingress, thermal growth. It degrades, and the rate of degradation is what you are buying with a premium fixture.
Where the supplier cites machine tool test standards as the basis for their accuracy claims, ask what was actually tested and on what machine. Standards such as ISO 230-1 cover geometric accuracy of machine tools under no-load or quasi-static conditions, and ISO 10791-7 addresses cutting accuracy of machining centres through a finished test piece. Neither is a fixture acceptance standard: a test piece that passes on the machine tells you the machine and the process combination can meet a tolerance, not that your fixture will keep doing so after two years of production. Treat such references as context, not as evidence about the fixture itself.
Witness a first-article run, not just a CMM report
A dimensional report on a single part proves the fixture can make one good part once. Ask to witness, in person or by recorded video with the machine display visible, a short run that includes the conditions the fixture will actually meet:
- At least 20–30 consecutive parts, with the key feature measured on every part or at a stated sampling interval, and the spread reported rather than the best result.
- One deliberate disturbance — a change of operator, a stop and restart, a mid-run tool change — to see whether the result survives a normal disruption.
- A loading trial with the operator who will actually run the job, if ergonomics affect how firmly the part is seated before clamping.
- A force verification at the start and end of the run, so you can see the drift within a single shift.
Record what you witnessed and attach it to the approval. That record becomes the baseline for the maintenance discussion later.
Documents worth requesting before you commit
The exact document set depends on the fixture type, but a credible supplier of machined workholding should be able to supply most of the following without a long delay. Treat prompt, organised delivery as evidence in itself.
- Assembly drawing with clamp locations, datum scheme and the locating principle explained (for example, how many degrees of freedom each element constrains).
- Bill of materials identifying purchased clamping elements by brand and model, so you can source spares independently.
- Force budget and the assumptions behind it.
- Commissioning checklist covering force verification, interlock test and first-part measurement.
- Maintenance schedule stating inspection intervals for jaws, seals, pins and rest pads, with the wear limits that trigger replacement.
- Spare parts list with lead times for the wear items, not only for the fixture body.
- Change record showing how design revisions are identified on the drawing and on the physical fixture.
Red flags that survive a good presentation
Most weak suppliers can produce a polished brochure. These signals are harder to manufacture:
- Force numbers quoted without any stated assumption about friction, safety margin or pressure losses.
- No answer to "what happens when the pressure drops mid-cycle?", or an answer that relies on the operator noticing.
- Purchased clamping components with brand and model concealed, which quietly locks you into the supplier for spares.
- Accuracy claims traced only to the machine tool's own test report.
- Maintenance described as "as required" with no wear limits or intervals.
- First article offered as a single part with no run data and no witness arrangement.
- Reluctance to let you speak to an existing customer running a comparable part, rather than a showroom reference.
Any one of these may have a reasonable explanation. Three or more together, and the burden of proof should shift firmly to the supplier.
A workable evaluation sequence
Run the evaluation in this order and stop early when the evidence does not appear:
- Process review — does the supplier engage with cutting forces, cycle time and part rigidity, or just quote a catalogue item?
- Concept and force budget — is the clamping concept explained and the force derived?
- Verification method — is force confirmed, with a defined response when it is out of band, and is it interlocked?
- Documentation — drawings, bills of materials, maintenance limits and spares.
- First article and run data — spread across a run, under a disturbance, with force checked at both ends.
- Commercials — spares lead times, warranty scope covering wear behaviour rather than only workmanship, and the supplier's response commitment when a fixture drifts in service.
The aim is not to find a supplier who promises the highest clamping force. It is to find one who can show, with numbers and records, that the force they promised will still be there in operation 10,000 — and who has thought about what happens on the day it is not.