Part 1 of a 3-part series on onboarding new customers and their RIM programs.
Quick Answer: Reverse engineering a RIM part means capturing the geometry of an existing physical sample, comparing it against any drawings you have, and producing a clean CAD model that a new tool can be built from.
Table of Contents
A sourcing manager opens a project folder on a Tuesday morning. Inside it: a 30-year-old RIM enclosure that ships on a flagship product line, a 2D drawing that doesn't match the part in hand, and an email confirming the original supplier shut down last quarter. Demand for next year is 1,800 units. There is no CAD model anywhere in the building.
This scenario shows up more often than most procurement teams expect. Legacy programs outlive their original suppliers. Drawings get lost in acquisitions. Engineering teams move on. The part keeps shipping, and one day it stops, because nobody can reproduce it without help.
This is where laser scanning, metrology, and reverse engineering earn their keep. For sourcing teams onboarding a new RIM partner under these conditions, the path back into production runs through digital capture. Here's what that work looks like.
Reverse engineering a RIM part means capturing the geometry of an existing physical sample, comparing it against any drawings you have, and producing a clean CAD model that a new tool can be built from. The deliverable is a digital file your supplier can quote, machine, and validate against.
The work is harder than it sounds.
A 20-year-old RIM enclosure has been through real life. The plastic has aged. Wall thicknesses may not match the original drawing. The original supplier may have made undocumented mold modifications across three production runs. The sample in your hand may be the latest revision, or it may be one nobody kept records for. A useful reverse engineering process accounts for all of that.
Laser scanning uses a coordinated array of measurement points to capture a part's surface as a dense point cloud. Software reconstructs surfaces, edges, and features from that cloud and exports a CAD-ready model. For RIM parts in particular, scanning has practical advantages over traditional touch-probe CMM measurement.
Large enclosures and structural components are slow and expensive to capture point-by-point. Scanning gathers thousands of data points in the time it takes to manually probe a few dozen. Heat maps then show, in color, how the physical part compares against any drawing, master sample, or earlier scan. Where the part is dimensionally on-spec, the heat map stays neutral. Where it has drifted, the deviation flags up immediately.
This is how a reverse engineering team separates two things that look the same but aren't. The first is the part as the original engineer intended. The second is the part as it has been produced for the last fifteen years. Both matter. The first tells you the design intent. The second tells you what the customer's assembly accepts in production today.
This is where engineering judgment lives.
A part scanned twenty years after molding will not match its original CAD perfectly. Material moves over time. Mold modifications accumulate. The version mounting on a product today is the version that needs to be reproduced, even if it has drifted from the original drawing.
Good reverse engineering does not treat the scan as gospel. It treats the scan as data. The engineering team compares the scan against any available drawings, against multiple sample parts when possible, and against the assembly the part has to fit. The deliverable is a CAD model that captures the part as it currently functions, with judgment applied to features that should be tightened, relaxed, or restored to original intent.
For sourcing teams, this is the place to push back. A reverse engineering deliverable that is just a digital copy of one aging sample is not enough. The model needs to produce parts the customer's assembly will accept in volume, on the next mold, for the life of the program.
Sometimes the right answer is "not all of it."
Capturing the geometry of a 40-year-old part doesn't mean reproducing every feature. Some features matter and have to be exact: the surfaces that touch the mating assembly, the mounting points, the overall dimensional envelope. Others are artifacts of an older manufacturing process that no longer applies. Treating both categories the same way produces a worse part than the legacy one and costs more to make.
A recent example. We reverse engineered Engine Bay Service Platforms for the U.S. Navy's AV-8B Harrier, an aircraft that has been flying for over forty years. The original platforms were foam-filled fiberglass. They had aged, broken in places, and had no surviving CAD data. The critical geometry, meaning the surfaces where a technician's foot lands on the airframe during engine maintenance, had to be reproduced exactly. Everything else was open for redesign.
We scanned the original parts, isolated the critical contact surfaces, and built a new design around them. The new platforms run in RIM with uniform wall thickness. They are lighter than the originals, more durable in service, and produced at lower cost than the legacy fiberglass parts. The Navy got a component that fits the same airframe, performs better, and ships from a current production process the supply chain can support for the next forty years.
This is the more useful frame for reverse engineering on legacy programs. The question is rarely "can you copy this old part." The question is "which features do we have to keep, and which can we improve?" For a sourcing team running a legacy program, that distinction is the difference between a like-for-like replacement and a generation-better one, often at the same or lower cost.
For a typical legacy RIM program coming in without CAD, the sequence runs roughly this way. Scanning and CAD reconstruction takes one to three weeks depending on part complexity and how many revisions are present. Mold design follows scanning. RIM tooling fabrication runs four to six weeks for most aluminum tools, based on Exothermic's standard programs and consistent with broader RIM industry experience. First-article inspection happens within two weeks of tool completion.
For a sourcing manager managing the program, the relevant number is total lead time. From "we don't have CAD" to "we have parts in our hands," eight to twelve weeks is realistic for a moderately complex enclosure. That math matters more when production is already running thin. The cost of a stalled assembly line, measured in customer commitments and missed shipments, dwarfs the cost of the scanning work that keeps it running.
The cleanest reverse engineering work starts with the cleanest inputs. A sourcing team supporting a reverse engineering project should plan to provide a representative sample part (ideally two or three, so the engineering team can identify any variation), any drawings that exist (even if they're known to be incomplete), the assembly or mating components the part has to fit, and any historical inspection or quality records that show what the part has been measured against in the past.
Design engineers on the OEM side often have the cleanest version of this data sitting on their drives. Bringing them into the kickoff call shortens the timeline measurably. Design firms acting as intermediaries between OEM and supplier (the kind that source and specify RIM work for their clients) typically already have this material packaged together, which is one reason projects routed through them move faster than projects routed cold.
Yes. A complete reverse engineering process can produce a quotable CAD model from physical samples alone, though available drawings shorten the work.
That is the goal of the heat-map comparison. The deliverable is a CAD model engineered to produce parts the existing assembly accepts.
Multiple sample parts let the engineering team identify the controlling revision, which is the version the next tool reproduces.
Scanning resolution is typically more than sufficient for RIM part tolerances, which are generally less tight than precision-machined metal components. Final tool tuning during sampling closes any remaining gap.
This is Part 1 of a three-part series on what onboarding looks like when a sourcing team brings a RIM program into a new manufacturer. Part 2 covers what happens when a working mold gets transferred from another shop. Part 3 looks at the legacy mold that is wearing out and the decision between refurbishment and replacement.
If your team is sitting on a part with no CAD and a production deadline, an engineering consultation is the fastest way to find out whether reverse engineering is the right path. Exothermic has been making RIM parts since 1971. Let's talk about your part.