In this week's newsletter, Aaron Moncur has a conversation with Edward Jaeck, President of Sine18 LLC, a manufacturing and quality strategy consultant with over two decades of experience at Intel, Medtronic, and Lowell Inc. Currently serving as the Midwest Sales Manager at Zygo.
Your drawings are wrong. They're wrong and your GD&T is lacking until you're a senior level GD&T certified expert with five years doing it and you know how to run the CMM and inspection and you know how the statistics play.
In this episode:
Why almost every engineering drawing contains errors that won't surface until supplier validation is already underway
How metrology matching closes the gap between OEM and supplier measurement systems and why dock-to-stock falls apart without it
Why design of experiments maps an entire variable range in the same number of shots that trial-and-error wastes on guessing
How critical feature confirmation eliminates IQ/OQ/PQ cycles on features that never influence output
Bonus Content:
Managing up - A skill that should be taught to every engineer
S5E34 Edward Jaeck | Minitab, Statistical Analysis, & Supply Chain Management
Edward Jaeck spent over two decades inside Intel and Medtronic before launching Sine18 LLC, where he now works with manufacturers and medical device companies to stop losing time on validation work that could have been avoided earlier in the design phase. In this episode, he walks through the uncomfortable reality that most engineering drawings have errors: not because engineers are careless, but because real GD&T fluency is rarer than the industry acknowledges. He covers how strategic planning cascades from high-level objectives down to measurable KPIs, why metrology matching is the missing link between supplier and OEM measurement systems, and a method called critical feature confirmation that can eliminate entire IQ/OQ/PQ cycles by proving up front that a tolerance doesn't affect output. For anyone working in medical device, semiconductor, or precision manufacturing, this episode carries real tactical value.
>If YouTube isn’t your thing, check out this episode and all of our past episodes on Apple, Spotify, and all the rest.
Stop. Before you automate that manual process step-by-step, read this.
In the 1890s, inventors tried to ease the transition from horses to automobiles by building "horseless carriages" - mechanical contraptions with articulated metal legs, rein-based steering, and fake horse heads housing lamps. They were solving the wrong problem. The answer wasn't replicating a horse. It was rethinking transportation entirely.
Modern automation projects make the same mistake. Your manual process evolved around human capabilities. Operators compensate for variation, use visual judgment, make micro-adjustments. Automating these workarounds directly creates complex, unreliable systems.
Pipeline Design & Engineering’s focus? Solving for the optimal outcome. Clear, reliable, effective.
Don't automate the workaround. Solve the actual problem.
Evaluating automation? Pipeline has solved this puzzle 100+ times.
Critical Feature Confirmation: Eliminating Validation Cycles Before They Start
The most expensive validation work is the work no one questions. When a drawing has 20 features and a medical device program requires IQ/OQ/PQ on all of them, engineering teams march through the list - three lots, full measurement, gauge R&R - because that's the protocol. The assumption is that every toleranced feature matters. Edward Jaacks pushes back on that assumption directly.
If you build that logically as a skew set and you test them and there's no statistical difference with those outputs - those tolerances that you ranged - none of them matter.
The concept draws from Taguchi's robustness work. Build skew samples that deliberately span the full spec range - parts at lower spec, parts at upper spec, parts distributed in between - and run them through the output function. If varying a dimension across its entire tolerance produces no statistical effect on performance, the feature is not critical. It cannot be critical, because even at worst case it doesn't move the needle.
How could they be critical features when even at their worst case they don't make a difference to the output function? It's robust. We're in the linear zone. Stop talking to me, these are not critical features. We're moving on.
This matters enormously in medical device, where validation cycles are long, resource-intensive, and often queued behind other programs. A feature that survives critical feature confirmation, meaning it produces no signal in the output, can be removed from the formal IQ/OQ/PQ scope entirely. No gauge R&R required. No three lots of parts measured across all those features. The savings aren't incremental; they're structural.
That in medical device saves a tremendous amount of time because you don't have to do a gauge R&R on it and you don't have to do an IQ/OQ/PQ on that feature and you don't have to build three lots of parts and measure all those features; which, is a lot of times a waste of time.
What makes this defensible in a regulated environment is that the statistical argument is explicit. The engineer isn't skipping validation out of convenience, they're generating data that demonstrates the feature doesn't warrant it. That's a fundamentally different posture with the FDA than simply omitting work from the plan.
Managing up - A skill that should be taught to every engineer
Early in my career, a technically sharp R&D manager was put in charge of pilot manufacturing operations he'd never run before. Within weeks, our team was dealing with unrealistic timelines, shifting priorities, and yield targets that had no connection to what the floor could actually deliver. The problem had nothing to do with technical competency on either side. The gap was in communication, and I was the one who had to close it. Once I started owning the expectation-setting and reporting on intermediate milestones instead of final outcomes, the ambiguity disappeared almost immediately. That experience reframed how I think about what engineering effectiveness actually requires. Managing up isn't a soft skill or a political maneuver, it's a technical competency that determines whether your work has any real impact.
To read the full article, visit the full article on The Wave.
When Your Machine Shop Says No
Last year, a customer couldn't find a shop willing to make their part. Difficult material, tight tolerances, small quantity.
We tested our internal supply chain. Made the parts ourselves. Sent them as a surprise.
They worked perfectly. That customer has ordered hundreds more since.
Now available: custom machined parts with Pipeline engineering oversight, pricing that outperforms domestic shops, and the same supply chain we trust for our own work.
Want to learn more? Contact us at the link below: