Above: A logarithmic scale of price per part per process.
There is a tendency to think that length of development time is the most important aspect of a start-up medical device program (besides safety and efficacy, of course). As a result, addressing the cost of components is often pushed off “until later." While this can be a good strategy, it often relies on the cost per part dropping when volume increases. For the most part, the idea of volume cost savings is simple: buy more and save more. This is not only true at your local supermarket, but for medical device components as well. However, the specific numbers in cost savings can prompt a pivotal discussion when determining whether a new product release is going to be a commercial success.
In six years of medical device design consulting, I’ve noticed the discussion usually goes like this: What is the actual cost of goods (COGs) of the last prototype we built? Wow, that’s high. Is it representative of the final COGs price? What is the estimated final COGs for the pre-production units (PPUs)? What about the first volume build? When can we turn a profit?
While each product development program is wonderfully unique, some very general cost-per-part guidelines are true for all electro-mechanical medical device components. This guide, and the associated discussion after, are intended to provide a very general idea of volume pricing for a given process. Estimates can be refined fairly easily by asking vendors for projected pricing estimates once later-stage prototyping begins. Note that labor and part handling are not considered in any way.
Above: A comparison of volume pricing possibilities per method.
Sheet metal parts are made first with shears, cutters, and punches to create flat patterns. Then brakes are used to create bends. Low volume is often done manually or semi-manually using a generic tool set. This contributes to a high set-up cost. As with most processes, the set-up cost is part development (one-time cost) with a recurring cost every time the part is run. As the volume increases, computer numerical control machines (CNC) can be added to automate some steps in production. In very high quantities, blanks, forms, or custom-machined tools can be used to further drive down price. Easy cost-saving measures also include using standard punches available at your vendor (they often have similar tools), inserts, and colors (matte black is always a safe bet as your vendor probably has it already). When making one or two parts, consider 3D printing them instead. For very high volumes, sheet metal parts can be redesigned to use a more cost-effective process such as injection molded plastics.
Machined parts take some time for set-up and problem solving in order to make the first one. Most machined parts are made on a CNC machine. Very small volumes of parts can be made manually in less time with simpler set-up and planning. When a higher volume oriented process is not possible, reducing the number of set-ups that require manual intervention on the CNC or reducing the total amount of machine time will both cut costs. Tolerance plays a large factor in costs. Precision parts require more passes, slower machining, increased inspection points, and often lower yield. Switching to a cheaper or easier-to-machine raw material also saves cost. Even setting up an extrusion in higher volumes can pay off as well.
Injection molding prices are almost completely driven by volume. Injection mold tools are negative impressions made out of hardened steels or stainless steels. Prototype tools, despite their name, are typically good for up to 5,000-10,000 uses and are usually made from a lower quality steel (or even aluminium). In theory, they produce lower-quality parts that can have more defects toward the end of the tool life. Production tools cost an order of magnitude more (or two!) but can produce up to a million higher-quality parts. Longer runs of parts allow the set-up to be amortized and can be more cost effective. Multi-cavity tools create more than one part at once. They can drive costs even lower. For lower-volume production, the costs of the tooling can be reduced through family tools—a single tool that can produce different parts at the same time. While they do not halve the tooling cost, they can reduce it significantly.
Urethane casting is the process by which a positive is made (often 3D printed), then a negative tool is cast out of a compliant material like silicone, and the final part is cast using rigid urethane. These negative tools usually tear or are damaged after 10 uses or so and have to be remade. It’s really only a cost-effective method in the 5-50 unit range. Below five, it may be cheaper to 3D print or machine the part. Above 50, injection molding or other processes might be the best option. Urethane cast parts have some unique design attributes (draft is good but not always required, walls are quite thick, etc.), which mean they may be perfect for certain geometries. However, they may require design changes between processes or may not be suitable for some designs.
3D printing costs are mostly driven by materials cost and machine time. The latter is a slightly bigger factor than in machining, unless exotic materials such as metals are used. 3D printing typically requires post processing on each part—stripping the support elements and sanding down any protrusions. 3D printing isn’t a great COGs choice for large parts or higher volumes. Other processes should be considered to replace it before production begins; however, 3D printing is very fast and can make parts that no other process can make. It can be a very valuable choice for early or complex development. Because 3D printing is additive, there aren’t significant economies of scale. Typically, cost structures are solely based on material and time on the machine with a one-time handling fee.
PCBAs (printed circuit board assemblies) are made in a very automated process, first creating the appropriate layers, then machining out the form, and finally soldering on all the components from bulk reels and mass soldering. The components themselves are a major cost factor, but often their prices scale well when buying in bulk. PCBA prices scale very well for these reasons. To reduce PCBA prices, avoid more involved features such as coated vias, through-hole components, and high-priced components (though they may actually be cheaper in small volumes). When volumes reach a very high level, consider application-specific integrated circuits to replace the PCBA (or a portion thereof) entirely.
Custom cables typically require a crimped connector on the ends of an off-the-shelf raw-cable. They are cut to length, and the ends are crimped on and then labeled and tested. This is a fairly manual process, so the setup cost is relatively low. Overall, cables like this are fairly cheap and may not make up a significant portion of your overall COGs. Prices can still be reduced by using cheaper components and simpler designs. Major cost reduction can only be achieved by eliminating cables or using custom, multi-wire cables with dedicated lines. These include a fully custom extrusion that has a very high tooling cost compared to the cost of the finished cable.
Don’t forget that COGs are driven by process as much as they are by raw material. In lower volumes, process costs can dominate. In higher volumes, as the process time is reduced, raw material plays a larger role. A well-designed device with parts designed for higher-volume processes will reduce in cost as volumes increase. Changes can always be made at any time to target a reduction in COGs, but major changes to parts in medical devices will likely need to be re-verified. This adds to the investment required if a change is done after product release.
Utilizing higher-volume processes in the detailed prototype phase (after early product definition but well before pre-production design) may result in higher device costs up front, but those parts will see an effortless drop in per-part cost in higher volume without any changes to the design needed. This is a trade-off: in the plus column, this means no re-engineering, re-testing, or letters to file required, but in the negative column, you may be requiring more capital than required. The best advice is to make a conscious, informed decision when developing your unit cost per milestone plan to achieve the right targets for your unique product.