Why Ten Good Parts Prove Nothing

Ten parts in a bag prove the toolmaker coaxed ten parts out of the mold. They do not prove cavity 4 holds tolerance on a humid Tuesday. Injection molding counts as a special process under FDA rules. Its output cannot be checked by looking at finished parts alone. Weld strength, molecular weight, and hidden dimensions sit inside the part. The process itself must be validated, not just the samples it makes.

A buyer who accepts T1 samples as proof takes on the risk. The mold may run well for a day and drift for a month. Barrel temperature moves, resin lots change, and a worn heater shifts the window. Validation exists to catch that drift before parts ship to a hospital. It is not paperwork for its own sake.

IQ: Is the Equipment Built and Installed to Spec?

Installation qualification checks static hardware before any validation plastic is shot. It confirms the mold, press, dryer, and auxiliaries match their specifications. No molding happens in this stage. That is the point. A fault found here costs a document change. The same fault found in PQ costs a re-run.

Mold IQ covers steel certificates for cavity and core, including heat-treat records for grades such as H13 and 420 stainless. Cooling circuits get a pressure test at 6 to 8 bar for 30 minutes with zero pressure drop. Hot-runner heater resistance and Type J thermocouple calibration go on record. Each cavity gets its own engraved ID.

Press and auxiliary IQ covers calibration certificates for load cells, position transducers, and cavity pressure sensors. The desiccant dryer must hold a dew point of minus 30 degrees Celsius or lower. Thermolator flow must stay turbulent, with a Reynolds number above 4,000. Cleanroom particle counts confirm the room meets its ISO class. Our drying guide for hygroscopic resins covers the same discipline for PA, PC, and PET.

Cleanroom class is part of IQ, and buyers now watch it closely. Demgy Chicago converted its ISO Class 8 clean room to Class 7 in October 2026 to serve a medical customer. Plastics News reported the upgrade from the company’s Elk Grove Village plant. A missing steel certificate at IQ stops the whole program. Ask for the certs with the quote, not after T1.

OQ: Where Are the Edges of the Process Window?

Operational qualification maps the boundaries of the process window. It starts with scientific molding studies, then runs a design of experiments at the window corners. A real OQ proves parts pass spec at the worst-case high and low settings. Those corner values then become the hard alarm limits on the press.

The studies come first. A viscosity curve plots apparent viscosity against fill speed, from 10 to 200 millimeters per second. The flat shear-thinning plateau sets the fill velocity. A pressure-drop study keeps maximum injection pressure at or below 80 percent of the press limit. Cavity balance must reach 95 to 98 percent mass balance across short shots. A gate-seal study weighs parts across one-second hold-time steps until weight plateaus. It then locks hold time 1.0 to 1.5 seconds past gate freeze. Our scientific molding guide walks through each study in full.

Then comes the corner DOE. Engineers vary melt temperature by plus or minus 5 to 10 degrees. Mold temperature moves by plus or minus 5 to 8 degrees. They also vary holding pressure and cooling time by plus or minus 10 to 15 percent each. One corner pairs high hold pressure with a cold mold and cold melt. The other pairs low hold pressure with a hot mold and hot melt. Parts from the nominal center, the high corner, and the low corner must all pass 100 percent of critical-to-quality specs. Shrinkage moves with temperature and pressure, so the corner trials exist to prove the steel still holds size at the edges.

An OQ that molds 30 parts at the center setpoint is not an OQ. It is a repeat of T1 with better paperwork. Ask the supplier for the corner data before you accept the OQ report.

PQ: Does the Process Repeat Across Shifts and Lots?

Performance qualification proves the locked nominal process runs the same way under real production conditions. It runs three separate production lots, each two to eight hours long. The runs often cross shifts or resin lots. Each run starts from a cold press.

Sampling runs at 30 parts per cavity or more, pulled at regular intervals across each run. A CMM measures them with a fixture validated by Gage R&R. Every critical dimension must reach a Cpk of 1.33 or higher. Safety-critical features need 1.67 or higher. A Cpk of 1.33 means about 63 defective parts per million. A Cpk of 1.67 means about 0.6 per million.

PQ is where a process earns the right to run without an engineer watching it. If PQ fails, the fix goes back to OQ or to the steel. No amount of sorting saves a process that cannot hold its window.

FAI vs PPAP Level 3 vs IQ/OQ/PQ

Buyers use three qualification tiers, and the tier must match the risk. First article inspection suits commercial parts. PPAP Level 3 suits automotive and high-volume industrial parts. IQ/OQ/PQ suits medical and safety-critical parts. The table below compares them on the points buyers actually argue about.

The cost and timing figures come from the IQ/OQ/PQ and PPAP Level 3 guide published by Plastics Technology Alliance. Pick the tier that matches the risk. A syringe barrel needs PQ. A diagnostic housing may only need PPAP Level 3.

The Multi-Cavity Trap: Never Pool Cavities for Cpk

Multi-cavity molds hide a statistical trap. A supplier can measure 8 shots from a 4-cavity mold and pool all 32 numbers into one Cpk column. That number is invalid. Between-cavity steel variation is a fixed offset, not random noise. Pooling hides the cavity running at the edge of tolerance.

The correct method measures 30 consecutive shots separately for each cavity. Each cavity must meet Cpk 1.33 on its own. Cavity-to-cavity mean centering should fall inside the middle 50 percent of the tolerance band. Our mold flow simulation guide shows how cavity balance gets predicted before steel is cut, which is where this problem starts.

One pooled number can hide a cavity running at the edge. Ask for the per-cavity table, not the summary line.

USP Class VI Is a Grade Test, Not a Device Approval

USP Class VI confuses buyers more than any other medical label. It is a materials-level biological reactivity test under USP chapter 88. It applies to one resin grade and one lot. It never applies to a polymer family, and it is not a finished-device certification.

The finished device falls under ISO 10993, the biological evaluation framework. Material evidence feeds into that evaluation. It never replaces it. A certificate that says “medical grade polycarbonate” names no grade and proves nothing. Name the grade on the drawing. Ask for the certificate against the lot that ran your parts.

Sterilization belongs in the same conversation. Steam, ethylene oxide, and gamma each punish different resins. PC crazes after repeated autoclave cycles. PA6 and acetal embrittle under gamma. Decide the sterilization method before the material, not after. A grade change after the tool is cut can change shrinkage enough to force steel work.

The QMSR Change: What Changed in February 2026

The FDA Quality Management System Regulation, or QMSR, took effect on February 2, 2026. It amends 21 CFR Part 820 and incorporates ISO 13485:2016 by reference. The old Quality System Regulation subparts are gone. Design controls now sit in ISO 13485 Clause 7.3. Purchasing controls sit in Clause 7.4. CAPA sits in Clause 8.5.

Inspections changed with the rule. FDA retired the Quality System Inspection Technique, known as QSIT, on the effective date. Inspectors now work under Compliance Program 7382.850. The FDA QMSR page lists the final rule, the inspection program, and the mapping resources.

If your supplier’s procedures still cite section 820.30 for design controls, they cite text that no longer exists. Ask when their quality manual was last revised. A supplier that updated its manual in 2025 planned for this change. A supplier that never heard of QMSR has not.

What Medical Mold Validation Costs and How Long It Takes

Validation costs scale with the tier. FAI runs $400 to $1,800 for a 1 to 4 cavity tool, often bundled into the tooling quote. PPAP Level 3 runs $2,500 to $8,500, driven by CTQ count and fixture complexity. Full medical IQ/OQ/PQ runs $8,000 to $35,000 or more for the same cavity count. That figure covers protocol writing, resin, DOE runs, and 90 to 300 CMM layouts.

Timing follows the same ladder. PPAP Level 3 takes 1.5 to 3 weeks after final steel grooming. IQ/OQ/PQ takes 4 to 8 weeks. Protocol writing and Gage R&R fixtures take 1 to 2 weeks. IQ, scientific molding, and OQ corner trials take 1 to 2 weeks. The three PQ lots, CMM work, and final report take 2 to 4 weeks.

Adding validation after the quote is where the money leaks. A buyer who asks for PPAP Level 3 or IQ/OQ/PQ after the supplier quoted a bare tool faces extra charges. The gap runs $3,500 to $25,000 or more in unbudgeted engineering and metrology work. The cheapest validation is the one written into the first quote. The most expensive is the one added after the mold is cut.

CTQ discipline keeps the bill down. Designate 3 to 8 critical-to-quality features on the drawing: mating latches, bearing bores, seal grooves, mounting hole pitch. Measuring a 30-piece capability study on every cosmetic rib wastes $5,000 to $15,000 in CMM time. The rest of the drawing gets a 3 to 5 piece full layout.

RFQ Checklist: Write These Lines Into the RFQ

Send these lines with the drawing. The answers separate a shop that has run medical validation from one that has only read about it.

• Qualification tier required: FAI, PPAP Level 3, or full IQ/OQ/PQ, with the standard named.
• CTQ list: 3 to 8 features flagged on the drawing, with the Cpk target per feature.
• Gage R&R requirement for every CTQ fixture, with the acceptance limit stated.
• Per-cavity Cpk reporting: 30 consecutive shots per cavity, no pooled figures.
• OQ scope: scientific molding studies plus high, nominal, and low corner trials.
• PQ scope: three consecutive lots, run length, and shift or resin-lot spread.
• Deliverables: validation protocol, IQ/OQ/PQ reports, and CMM data per cavity.
• Resin grade named on the drawing, with USP Class VI certificate per lot.
• Cleanroom class required for molding and packing, stated as ISO Class 7 or 8.
• Change control: no material, tool, or process change without written approval.

A supplier that prices validation in the quote has done it before. A supplier that discovers it later has not.

When OQ or PQ turns up a defect, the cause usually sits in the process window, not the steel. The free defect troubleshooting guide on the Topworks toolkit page walks through 17 common defects and their likely causes.

Frequently Asked Questions