Pre-Acceptance Checklist for a Velo3D Sapphire 3D Printer: What I Check Before I Sign Off

Who this checklist is for

I'm a quality and brand compliance manager at an additive manufacturing company. I review every machine qualification report before it reaches the customer—roughly 30 to 40 metal AM systems a year. In 2025, I've rejected 9% of first deliveries because the documentation did not match the agreed acceptance criteria. That number should be zero.

This checklist is for teams that are about to accept delivery of a Velo3D Sapphire 3D printer. It's not a general comparison of every metal printer on the market. I'm not a photonics engineer, so I won't pretend to compare it to a photonics laser welding machine. If you're looking for a wood lathe tool rest holder, you're in the wrong place. And if your team is still asking “do i press the brake when starting a car,” you need operator training before you need capital equipment.

There are six steps. Follow them in order. Skip step four at your own risk.

Step 1: Define the part family before you define the machine

Start with the parts, not the brochure. The Velo3D Sapphire line is known for support-free complex geometries, which makes it valuable for aerospace, defense, and energy components with internal channels, lattice structures, or occluded features. But if your first part is a simple bracket with a drilled hole, a less expensive laser powder bed fusion system may be the right answer.

Before you write the purchase order, name the first three parts you plan to qualify. For each part, write down:

• Material and material specification
• Maximum build height and footprint
• Smallest internal channel or wall thickness
• Certification requirement or quality standard
• The specific dimension that kept you awake last night

That last line is not a joke. In my experience, most machine selection problems are actually part geometry problems. If you can't articulate the dimension that matters most, the acceptance test won't either.

This gets into metallurgy territory, which isn't my specialty. I can't tell you which alloy phase structure you need. But I can tell you that if the material pedigree is not defined before the PO is signed, you'll be trying to define it after the machine is sitting in your cleanroom.

Step 2: Force the site survey before you finalize the price

Almost every delayed install I've seen has the same root cause: the site visit happened too late, or it happened but nobody wrote down the conclusions. A Velo3D Sapphire 3D printer is not a desktop unit. It needs power, argon, compressed air, exhaust, network access, and floor space that meets the supplier's spec.

One example. I remember an install where the electrical connection was fine on paper, but the voltage drop under full load was outside the printer's tolerance. The machine passed the visual inspection. The first test build did not. No wait—the first build actually started fine. The failure came during a later titanium build, and the investigation took two weeks. A site survey with a load test would have caught it on day one.

Checklist for this step:

• Site survey completed on site, not by video call
• All utility points measured under load
• Argon supply and purity confirmed
• Ventilation/exhaust path documented
• Operator access and emergency stop locations reviewed

If the vendor says “our standard system comes with a remote site survey,” ask for the exceptions in writing. That is not a red flag by itself. But it is a risk to price into the project plan.

Step 3: Write the acceptance criteria before the machine ships

The phrase “manufacturer's standard acceptance” is the most expensive sentence in a purchase order. It sounds safe, and it means nothing until something fails.

Define what a passing acceptance build looks like. Add it to the PO as an attachment, or better, write it into the contract. At minimum:

• Test alloy and coupon geometry
• Minimum material density, tested by an agreed method
• Surface finish requirement on critical features
• Dimensional tolerance per ASME Y14.5-2018
• Destructive testing plan and who pays for it
• Sensor log and build report deliverables

I'm not a materials engineer, so I can't tell you which density value is acceptable for your application. That depends on your part and your customer. But I can tell you that “good enough” is not a measurable criterion.

Also check the powder specification. Reference ISO/ASTM 52907 for metal powder characterization methods, and require the material certificate to accompany the delivery. If a powder lot doesn't have a certificate, it is not acceptable. Period.

Step 4: Put powder traceability in the plan (this is the one people skip)

Most buyers spend all their time talking about lasers, build volume, and software. The powder handling plan gets almost no attention until the first contamination issue.

Metal AM quality starts with powder. A Velo3D Sapphire 3D printer is only as repeatable as the material you feed it. Your plan should include:

• Powder lot number and certificate on arrival
• Particle size distribution and flowability data
• Storage conditions and maximum moisture exposure
• Sieving and reuse procedure
• Labeling standard that prevents cross-contamination
• Sign-off for each lot before it goes into the machine

We didn't have a formal powder traceability process at one point. Cost us a two-day production stop when a mislabeled Inconel lot was found at the sieve station. It didn't reach the build, but we couldn't prove that until we'd quarantined and tested. The machine was fine. Our process was not.

I don't have hard data on how often powder traceability issues cause metal AM failures. But based on the audits I've seen, documentation gaps are the top reason a machine acceptance report gets kicked back. Don't let that be your report.

Step 5: Buy the timeline certainty your program actually needs

I know this sounds like a pitch for service contracts. It's not. It's the opposite. I've rejected machines, and I've seen what a response-time gap costs a production schedule.

When you buy a machine, you're also buying a service window. If the acceptance test fails on a Tuesday and the next available service engineer is “as soon as we can,” your cleanroom is now a very expensive storage closet. Ask for a written response-time SLA. It will cost more. It should cost more. Certainty has a price.

In March 2024, I watched a customer wait 11 days for a requalification visit because they had chosen a lower-cost installation package without a guaranteed response window. The machine sat idle under a tarp, the operators were paid to stand around, and the program missed a milestone. The extra expedite fee they had tried to save would have covered that delay many times over.

Rush fees are not about speed. They are about certainty. If your program deadline is worth anything, budget for the guaranteed response window.

Step 6: Define “done” before the truck arrives

A machine is not accepted when it is unloaded. It's accepted when the evidence says it is ready to run. In my system, “done” means:

• Installation report signed off by an authorized engineer
• Acceptance build data reviewed and archived
• Powder lot documentation complete
• Operator training completed on interlocks and startup sequence
• Maintenance plan and spare parts list on site
• Sensor logs exported and backed up

I still kick myself for accepting a machine with a missing gas purity certificate. I told myself we would follow up later. Later turned into six months, and the first serious titanium build scraped a full plate. If I had made that certificate a handover requirement, the gas supplier would have fixed the line before we ever produced a part.

One more thing: put the handover criteria into your own quality system, not in someone's email inbox. The delivery ticket might say “received.” Your quality record should say “not yet accepted” until every condition is met.

This gets into warranty and contract territory, and that's not my expertise. Have your legal team review the acceptance language before you sign, not after a dispute starts.

Three common mistakes I see in machine acceptance

1. Signing the delivery ticket before the acceptance build is completed. The freight company is not your quality department.
2. Using “manufacturer's standard acceptance” without writing down what that means. If you can't measure it, it doesn't exist.
3. Making the laser the center of the story and ignoring the powder, the process, and the people. The machine is one component of the process.

And if a training plan includes the question “do i press the brake when starting a car,” take it as a signal that operator readiness needs more attention. The Sapphire has safety interlocks, not brake pedals. But someone on your team should be able to explain how those interlocks are verified before the machine is powered on.

The bottom line

A Velo3D Sapphire 3D printer can produce parts that are difficult or impossible with conventional methods. But “capable” is not the same as “qualified.” The difference is in the acceptance plan, the traceability records, and the people who will run the system.

Follow the six steps, write down the criteria, and don't accept a machine that doesn't meet them. I learned the hard way so you don't have to.