A Machining Supplier Transition Without Quality Gaps
Moving a machined part to a new shop sounds simple on paper: send the drawing, get parts back, swap them in. In practice, a machining supplier transition is where quality gaps show up. A tolerance that the old shop held through a tuned fixture, a specific tool path, and an operator's habits can drift the moment a different machine, coolant, or inspection routine enters the picture. Parts that pass a quick check can still fail at assembly months later. Here is how to move a program without giving your customer a reason to ask questions.
Table of Contents
Quick Takeaways
Key Insight | Explanation |
|---|---|
Freeze the baseline before you send anything out | Pull drawings, revision history, and inspection data from the last several good lots. That record is your definition of "correct." |
Write down what the drawing does not say | Deburr method, datum setup, fixture orientation, and surface finish process often live in the old shop's habits, not in the print. |
Treat the first article as a hypothesis | A clean first article shows one setup worked once. It does not prove the process is stable across lots. |
Run an overlap instead of a hard cutover | Ship both sources for a defined window and compare parts side by side on the same gauges and the same inspection plan. |
Ask the new shop for its inspection plan in writing | Control plans, gauge lists, calibration status, and measurement uncertainty tell you more than a sample report does. |
Keep the old supplier qualified until exit criteria are met | Releasing the incumbent early removes your fallback at the exact moment a new process is most likely to surprise you. |
Start With a Program Inventory, Not a Quote Request
Most failed transitions begin with a quote request sent too early. The buyer forwards a drawing PDF, asks for pricing, and assumes the drawing contains everything the part needs. It rarely does. Before you contact a single new shop, build an inventory of the program as it actually runs today, including every document and habit that keeps the part in tolerance.
Pull the Real Revision History
Start with the revision history, not the current drawing. Look for engineering change notices, deviation approvals, and customer concessions that changed the part after the original release. In practice, the most common surprise is a note that was added after a field issue, such as a radius callout or a thread gauge change, and never made it back into the main drawing. If the new shop quotes from the current revision alone, it will build a part that is technically compliant and functionally different from the one in service.
Also confirm which revision is actually on the shop floor. Travelers, programs, and inspection plans sometimes sit on different revisions for months. Compare the printed travelers with the drawing archive and flag every mismatch before you lock the scope.
Separate Drawing Requirements From Shop Habits
A drawing defines what the finished part must be. It rarely defines how the old shop got there. Tool selection, clamping strategy, the order of operations, and the deburr method can all affect dimensional stability, especially on thin walls, deep bores, and features that move after heat treatment. Write these down as process notes. You are not asking the new supplier to copy them exactly, but you need to know which habits are load-bearing.
Pro tip: Ask the incumbent shop's machinist or inspector for a one-page list of the three features they worry about most. Those features almost always become the failure points in a transition, and the list costs nothing to get.
Capture the Baseline Your Current Parts Already Prove
Your current supplier has already produced the evidence you need, even if nobody has organized it. Recent inspection reports, control charts, scrap logs, and customer return data together describe what good looks like for this part. A transition is safe only when the new supplier can reproduce that baseline, and you cannot judge reproduction without the baseline in hand.
Inspection Data From Recent Lots
Pull inspection data from the last several production lots, not just the first article report from years ago. Look at where the dimensions cluster, not only whether they passed. A feature that sits at the upper edge of its tolerance band for every lot is a warning sign, because a new process with slightly different thermal behavior will push it out. Note features with Cpk values close to the minimum your customer accepts, and treat those as the first items to verify in the new shop.
If your current supplier does not track dimensional trends, do it yourself from the reports you receive. A simple spreadsheet with lot numbers, feature values, and gauge IDs is enough to show which dimensions are drifting. That data becomes the reference for every comparison you make later.
Setups, Fixtures, and Tool Paths
The setup is part of the baseline. Document the fixture design, datum references, workholding pressure where it matters, and the sequence of operations. Photograph the setup if the shop allows it. Tool paths and offsets are harder to transfer, because CAM output rarely travels cleanly between shops with different machines, controllers, and post processors. Expect the new shop to rewrite the program, and plan your first-article review around the features that depend most on the setup.
Also capture cycle-time and tool-life notes. A shop that runs a feature with a worn insert at a slightly lower feed rate may hold the same tolerance on a different machine only after the feeds and speeds are re-optimized. Those tradeoffs should be visible before the first part is cut.
Qualify the New Source With a Controlled First Article
Qualification is where a transition earns or loses its credibility. A controlled first article, inspected against the full drawing and the baseline you captured, gives you the first real evidence that the new supplier can make the part. Vague sign-offs based on a few spot checks are how quality gaps reach the customer.
A first article that passes shows that one setup worked once. It does not show that the process will hold across the next several lots, the next tool change, or the next operator.
What a First Article Report Actually Proves
A first article report proves that the new shop can produce a part that meets the drawing on the day it was made. It does not prove repeatability, tool life behavior, or how the process responds to a shift change. Read the report with that limit in mind. Check that every dimension on the drawing is measured, that the measurement method matches what the customer will use, and that gauge IDs and calibration dates are listed. A report that measures only the features the shop considers important is incomplete.
Request the raw measurement data, not only the summary pass or fail. Raw data shows whether a dimension passed by a narrow margin, and whether the distribution of several sample parts is centered or leaning toward a limit. Those details are invisible on a one-page summary.
PPAP and Documentation Expectations
If your program falls under PPAP, define the submission level before you start, not after the first parts arrive. Ask the new supplier which elements it can provide, including the control plan, process flow, dimensional results, material certifications, and capability studies. Be specific about the measurement system analysis you expect for critical characteristics. In our experience, the documentation gap is rarely about missing paperwork. It is about the new shop having a different understanding of which characteristics are critical, and that mismatch only shows up when someone compares the two lists line by line.
Run a Parallel Period Before You Release Volume
A parallel period means both suppliers produce parts for the same program over a defined window, and the parts are compared under the same inspection conditions. It is the only reliable way to see how the new process behaves under production conditions rather than in a controlled first article run. It also gives your team time to find problems while the incumbent can still cover demand.
Setting Overlap Length and Sample Size
Overlap length depends on how often the part runs and how sensitive its critical features are. A part that ships weekly may need several lots from each source before you have a meaningful comparison. A high-volume part with stable features may need less. The key is to define the exit condition in writing before the overlap starts, such as a specified number of consecutive lots within tolerance, with no trend in critical features. Without that rule, the overlap tends to end when someone gets tired of waiting.
Sample size matters as much as duration. Measuring two parts per lot will not reveal a feature that drifts over a full shift. For critical dimensions, measure a sample from the beginning, middle, and end of each lot, and record tool changes and any machine stops. Those notes often explain a scatter pattern that looked random in the data.
Side by Side Inspection on the Floor
The cleanest comparisons use the same gauges, the same inspector, and the same fixture where possible. If parts from two suppliers are inspected on different CMMs or with different probe configurations, a difference in readings may reflect the measurement setup rather than the parts. Agree on the measurement method for each critical feature, and confirm that every gauge is in calibration. Where a feature depends on a fixture, consider inspecting in a common fixture to remove that variable.
Pro tip: Mark every parallel-run part with its supplier, lot number, and machine ID at the point of receipt. Parts that arrive without this information are almost impossible to trace back when a dimension starts to move.
Compare Transition Methods Before You Commit
Not every program needs the same transition approach. The right method depends on part risk, volume, and how much the new process differs from the old one. The table below compares three approaches that are common in practice, along with the situations where each one holds up and where it breaks down.
Transition Method | Best Suited For | Main Risk |
|---|---|---|
Full first article and hard cutover | Low-risk parts with simple geometry and few critical features | Problems that appear only after several lots go unnoticed until the customer reports them |
Parallel run with defined overlap | Parts with critical features, tight tolerances, or a history of drift | Overlap can drag on without a written exit condition, and cost rises while both sources run |
Phased cutover by lot or feature group | Large programs where a single failure would affect many assemblies | Mixed-source inventory can confuse traceability if lot records are not kept carefully |
Where Each Method Breaks Down
The hard cutover is fast and cheap, which is why it is tempting. Its weakness is that it relies on the first article and a few early lots to represent the whole process. For a part with a critical bore or a thin flange that moves after clamping, that sample is not enough. Use a hard cutover only when the part's risk truly is low and the feature set is simple.
The parallel run is the strongest method for high-risk parts, but it is expensive and requires discipline. Both suppliers must be kept active, and the buyer must actually compare the data. The phased cutover works well for large programs, but it adds traceability work. If lot numbers are not tied to the correct source on every purchase order and receiving record, you will eventually ship a mixed lot without knowing it.
Cutover, Exit Criteria, and Ongoing Monitoring
The cutover is the moment you stop relying on the old source. It should happen only after the exit criteria you set at the start of the overlap are met. Many quality gaps occur in the weeks after cutover, not during qualification, because production pressure reduces attention and the old supplier's safety net is gone.
Defining Exit Criteria for the Old Supplier
Exit criteria should be written and signed by both quality and purchasing, not just the engineer who ran the program. A useful set includes the number of consecutive lots within tolerance, the absence of a trend in critical features, closure of any open corrective actions, and confirmation that the new supplier's documentation package is complete. Once those conditions are met, release the incumbent in a controlled way. Keep the old supplier's approved status active for a defined period, so you can return to it if the new source fails.
Be careful about what you tell the incumbent. Some shops reduce their own process discipline once they know they are losing the work. If the old supplier must supply during the overlap, keep the same inspection requirements in place until the final lot is accepted.
Watching the First Months of Production
After cutover, keep the inspection frequency higher than normal for at least several lots. Track the same critical features you watched during the overlap, and compare them with the baseline you captured earlier. A gradual shift in a single feature often appears before a rejection does. Review tool change logs and operator notes, because a new shop's tool life may differ from what your baseline assumed. A supplier that is performing well will welcome this attention, and a supplier that resists it has told you something useful about the relationship.
Frequently Asked Questions
How long should a machining supplier transition take?
For a simple part with few critical features, a transition can be completed in a few weeks from first article to cutover. Programs with tight tolerances, multiple setups, or PPAP requirements often take several months, mostly because the overlap period needs enough production lots to show real stability. Set the timeline around the exit criteria, not around a calendar date.
Do I need a new first article report if the drawing has not changed?
Yes. A drawing that has not changed does not mean the process has not changed. A different machine, fixture, or measurement method can shift results even when the print stays the same. A new first article report establishes the baseline for the new source and gives you a documented reference for future comparisons.
What documents should I request from a new machining supplier?
At minimum, ask for a first article inspection report with raw measurement data, a control plan covering critical characteristics, a gauge list with calibration status, material certifications, and any capability studies for features that drive risk. If the part falls under PPAP, confirm the submission level in writing before the first parts are made.
Should I keep the old supplier active during the transition?
Keep the old supplier approved and active until the exit criteria are met and the new source has shown stable performance over several lots. Releasing the incumbent too early removes your fallback at the point where a new process is most likely to reveal a problem. Once you release it, keep its approval status available for a defined period.
What is the most common mistake in a supplier transition?
The most common mistake is treating the first article as proof of process stability. A first article shows that one setup produced acceptable parts once. The real risk appears across lots, tool changes, and shifts, which is why a defined parallel run with written exit criteria matters more than a clean first inspection report.
Have you moved a precision machined part between suppliers, and which feature gave you the most trouble during qualification? Share your experience in the comments so other buyers and machinists can learn from it.
Comments