The lifecycle stage of a new product dictates what types of components you can and should design into that iteration. Sometimes you need speed, other times you need quality. We will discuss 3 lifecycle stages and how these impact where and how you source components.
A brand new hardware product starts its lifecycle as a prototype. No engineer starts by building 1000 of a product without testing it in small increments first. There are various resources for obtaining parts to add to a single quantity board.
In-house component stock
If your company already makes products and has parts on hand, this is a great place to source components. The obvious benefit is using already-paid-for components. Reusing components on future products also generates volume discounts on individual MPNs.
The lab supply
Larger design labs will often have a stash of components available for prototyping. These are wonderful resources for fast initial builds. If something goes wrong on the first revision, changing to a new component does not require waiting for UPS or FedEx.
Lab inventory of the in-house components mentioned above makes prototyping faster. Passive component kits, containing standard resistor/capacitor values, allows swapping to a different value as needed. “Good enough” generic parts allow testing concepts without hunting down specific part numbers.
Many semiconductor vendors will provide samples of components to engineers for free in low quantities. This is not charity. The vendor sales team wants to come in and learn about what you are working on. Learning about your other needs means they can attempt to sell you other chips and solutions. Depending on the vendor, the lead time for these parts can be unreliable. You will also need to prove that you are working on something profitable in the future.
This is the most common and reliable method for obtaining parts. Advancing the state of technology on your product requires access to new components and technology. The best way to do this is to buy parts from distributors that will sell you newly introduced components in quantities less than 10. Obtaining these parts is as fast as you are willing to pay for; overnight shipping is an expensive but useful method.
The above methods are a convenient way to get a functional prototype or proof of concept. These parts are not always suitable when moving to a pilot build. A pilot build, defined here as a run of 5 to 50 units, attempts to shake out possible production problems. It is a test run for your eventual higher volume run. By this definition, some products stay in pilot mode for a long time. Some never move past low volume production. This means that all components in that product can be sourced from an online distributor channel (Digikey, Mouser, Newark). Here are how components need to change for a pilot build.
Each component needs a specific MPN
Using resistor or capacitor kits is no longer acceptable. At this stage, it’s important to move to a Bill of Material (BOM) that has all Manufacturing Part Numbers (MPNs) listed. This forces analysis of proper specification on each individual component. Will a 5% resistor have sufficient tolerance? It might have been OK for a prototype, but you may need a higher quality part. If so, is that part available in the supply chain?
Cost optimizing components
When components are not as important as the baseline function, go for lowest cost. For example, using a 2N3904 instead of a more highly spec’d NPN transistor. Doing so during the pilot build will test supplier reliability and specification needs. Listing functional equivalents provides an even more robust supply chain.
Higher spec’d parts
Most online distributors only stock one “flavor” of component. If an op amp has 3 different speed grades, you may only be able to buy the middle speed grade component. Distributors want to cover the widest audience of potential component buyers. Engineers should analyze the available components and check if the specs are sufficient. If not, move to a higher spec’d part. If the component is already over spec’d, move to a lower spec’d, lower cost part. In all cases, ensure your specific “flavor” is available from distribution.
Reanalyze in-house components
This goes in both directions. In one direction, analyze moving from an in-house component to a brand new component. It’s possible that your in-house components are close to obsolescence, or superseded by a new lower cost part. Either could point to a transition to that part.
In the other direction, you should question why you designed in a completely new component at all. Will one of the in-house components work just as well? The cost savings for buying a higher quantity of parts often can offset the cost of analyzing and testing for compatibility.
Sourcing directly from distribution for anything more than 5–50 units is tricky business. In some cases, the part is popular enough to be in stock on a regular basis. For a less common component, sampling 100 or so parts from a friendly sales rep will get you out of a jam. Or you will still be able to order from a distributor and not be required to order an entire reel. But you may incur a lead time for ordering a non-standard part number.
Moving from the pilot stage to the production stage requires more care and analysis. It’s common practice to bring in manufacturing engineers and purchasing agents. The design engineers started with “a prototype or pilot that works”. The production pros make “a final product that works andproduced in large quantities”. Companies, especially small ones, that fail to do this will run into component sourcing headaches.
All components ordered should be on a reel if possible. A tube or tray also works, depending on CM requirements and component type.
Though usually still purchased via distribution, these parts will come directly from the factory. The 8 week (or more) lead time is due to processing time of silicon, which starts when the order is placed.
The largest risk is lead time. A finalized Pilot Production BOM should immediately start quoting production quantities. All scheduling should pay special attention to lead time. Any part delay to the assembler will directly impact the delivery date of your final product. All parts must be on hand to complete a build. Unfortunately for the design engineer, your fate is sealed by this point. The parts on your BOM are now at the mercy of the vendors you have chosen and their ability to deliver parts. Any attempts to change the BOM will result in further possible issues or delays.
High volume manufacturing of a product has serious impact in time and product quality. A prototype is a fundamentally separate entity from the final production unit. The transition deserves serious analysis from the design engineer and any manufacturing personnel/partners. Understanding how components must change over time will result in better products.
We will write about how to pull these decisions forward in the design process in future articles.