Switching from traditional machining to additive manufacturing for high-stakes production parts isn't trivial. I'm a quality compliance manager at a mid-sized precision engineering firm that supplies components for defense and aerospace systems. I review roughly 200+ unique part configurations every year before they reach customers. If you're evaluating Velo3D's Sapphire printers for mission-critical applications—especially given their track record with Anduril and SpaceX—here's the practical checklist I'd use.
This checklist assumes you've already done the high-level vendor evaluation (cost per part, lead times, etc.). We're going deeper. Below are three steps that go beyond the basic datasheet review—and one step most procurement checklists miss entirely.
Step 1: The 'No-Support' Geometry Audit
Velo3D's key differentiator is their ability to print complex geometries without traditional support structures. Sounds fantastic on paper. But I've learned to verify this capability against your specific parts—not just their marketing demos.
What to ask:
- Show me a part with overhangs greater than 45 degrees from the build plate, printed without supports, and passed X-ray/CT inspection. (Their demo parts are often optimized—ask for a third-party example.)
- Provide the surface finish data (Ra, Rz) on the unsupported down-facing surfaces vs. supported ones. The difference can be more than 50% in some systems.
- Request a 'challenge part'—preferably one of your own designs that you currently cannot make with traditional methods or need support structures for—and have them print it.
(Honestly, this is the only way to be sure. Vendor specs are one thing; your geometry on their machine is another.)
It took me four years and about 150 vendor evaluations to understand that capability is geometry-dependent, not a blanket 'yes/no' checkbox.
Step 2: Material Traceability & Process Certification (Not Just Chemistry)
For defense and aerospace work, material certification isn't optional. Most metal AM vendors provide chemical composition certs. But that's table stakes. The real quality question is process consistency.
What to verify:
- Powder bed density traceability: Ask how they verify density distribution across the build plate. Velo3D's 'Assure' quality control system is built for this, but ask for specific data on a recent production run (batch size: 10+ parts).
- Mechanical property stats: Request tensile test results (yield, tensile, elongation) from multiple locations within the same build plate and across different builds. The variance tells you more than the average. A vendor with a Cpk (process capability index) above 1.33 for key properties is serious about production.
- Heat treatment protocol: Post-processing is part of the system. Ask for their standard stress-relief and HIP (Hot Isostatic Pressing) parameters for your material (e.g., Ti-6Al-4V or Inconel 718). Compare these to industry standards (e.g., AMS 4999 for Ti-6Al-4V).
"In our Q1 2024 quality audit, we found that over 30% of first-time additive manufacturing vendors failed to provide meaningful process capability data—only chemistry certs."
Step 3: The 'Hidden' Checklist Item—Software & Calibration Governance
This is the step most people miss, and it's where a lot of quality issues start.
What to ask:
- How is the build file (the instructions for the printer) version-controlled? Can I see the log of changes between revision 1 and revision 2 of a successful part?
- What is the calibration frequency for the recoater blade, the laser power, and the optical sensors? Ask to see the last 6 months of calibration records for the specific machine that would run your parts.
- What is their protocol if the build is interrupted (e.g., power loss, oxygen spike)? Do they restart, scrap, or document the defect? The answer—and the documentation—reveals a lot.
The numbers said go with a different vendor that had slightly lower per-part cost. My gut said to dig deeper on their software governance. Went with my gut. Later learned they had a known bug in their build processor that affected layer alignment for certain geometries—something they fixed in an update but hadn't disclosed to new customers.
Watch Out for These Common Misconceptions
- "AM is always 'good enough' for prototyping." This was true 10 years ago when the technology was more experimental. Today, Velo3D's machines are used for flight-critical parts (supplier to SpaceX, Anduril). Treat their output with the same rigor you would a machined part from a certified shop.
- "Velo3D printers don't need support structures, so we can skip DFM." Not true. While they reduce support needs significantly, design for additive manufacturing (DFAM) is still critical. Thermal stress, powder removal, and surface finish requirements all need design consideration.
- "One test print is enough to validate the process." You need at least 3 to 5 builds with consistent results to establish process capability. Single-point validation is statistically meaningless for production.
A quick note on pricing: Velo3D Sapphire systems are not commodity items. As of mid-2024, a fully configured system with installation and training is typically in the $700k–$1M+ range (verify current quotes—pricing varies with configuration and service contract). The value proposition is not 'cheap'—it's 'capability you cannot get elsewhere.' If cost is your primary driver, this might not be the right system.