<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Michael Corr on Medium]]></title>
        <description><![CDATA[Stories by Michael Corr on Medium]]></description>
        <link>https://medium.com/@michael.corr?source=rss-a440d04f6820------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*-sJO-Z2HKO68_RF15y8vCA.jpeg</url>
            <title>Stories by Michael Corr on Medium</title>
            <link>https://medium.com/@michael.corr?source=rss-a440d04f6820------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 08:11:53 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@michael.corr/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Why your next hardware project should look like a software project]]></title>
            <link>https://medium.com/duro-labs/why-your-next-hardware-project-should-look-like-a-software-project-576c08d7c1da?source=rss-a440d04f6820------2</link>
            <guid isPermaLink="false">https://medium.com/p/576c08d7c1da</guid>
            <category><![CDATA[data-management]]></category>
            <category><![CDATA[plm]]></category>
            <category><![CDATA[hardware]]></category>
            <dc:creator><![CDATA[Michael Corr]]></dc:creator>
            <pubDate>Tue, 03 Nov 2020 00:00:00 GMT</pubDate>
            <atom:updated>2020-11-04T17:37:30.910Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Fw1B5L5RwYWCu-ZE" /></figure><p>Innovation doesn’t come from magic. It comes from a well structured environment that enables and encourages innovation to prosper, freeing people to take risks and try new things. We’ve seen an incredible amount of innovation in software development tools and processes in the past 15 years — in fact, several orders of magnitude greater than any 15 year period, prior. So, why haven’t tools to design and manufacture hardware seen this same explosion in innovation?</p><p>Because hardware teams are bogged down by data administration and inefficiencies. Their native development environments are complicated and rife with intertwined manual processes.</p><h3>Leading versus Lagging.</h3><p>The very first line of code any modern software engineer writes is “git init”. Meaning, software teams lead with revision control and set up their environment <strong><em>before</em></strong> they write any application code.</p><p>Conversely, hardware engineers <strong><em>lag</em></strong> with revision control. Electrical and mechanical engineers more frequently dive right into CAD first and only once they’ve reached a critical point to capture do they encapsulate their work and set up formal revision management tools.</p><h3>Why is that?</h3><p><a href="https://en.wikipedia.org/wiki/Computer-aided_design">CAD</a> and <a href="https://en.wikipedia.org/wiki/Bill_of_materials">BOM</a> files are literally the equivalent of <a href="https://en.wikipedia.org/wiki/Source_code">source code</a> for hardware products. There’s no lesser value in revisioning and centralizing these files than that of software files.</p><p>Do hardware engineers just not understand the value of revision management or is it because the tools are simply too difficult to set up? I’m positing that it’s the latter.</p><h3>This has got to change.</h3><p>The data management tools for CAD, BOM, and Supply Chain data have to be simple, low cost, and ubiquitous. Let’s look at <a href="https://git-scm.com/">git</a>, the dominant revision control technology used for software text files. It’s not only free, comes by default with every Macbook, but quite literally only requires a single simple command line string to set up in less than 30 seconds. It can’t get any easier than that.</p><p>The real impact of git is that it lets developers take risks to try new things with minimal penalty if they fail. Git provides a safety net that developers can depend on to always show them exactly what changed, when it changed, and how it changed — empowering them to quickly undo anything that went wrong. This is quite liberating for a developer and leads to greater degrees of innovation.</p><p>We need to achieve this same ubiquity and simplicity of a comparable tool to manage hardware files. This is a prerequisite before the hardware industry can really accelerate productivity and empower hardware engineers to innovate as quickly as software teams can.</p><p>Currently available revision management tools for hardware files fall into a product category called PDM, or <a href="https://en.wikipedia.org/wiki/Product_data_management">Product Data Management</a>. They are sold as optional add-ons, usually at a substantial cost. Further, they are sufficiently complex to install and configure such that it takes expertise or devoted time to set up. As a result, hardware engineers procrastinate and don’t set up these tools until they absolutely have to, if at all.</p><h3>So why hasn’t this been solved already?</h3><p>Partly due to culture, but mostly because hardware files are predominantly large binary CAD files, often in proprietary file formats. This makes it difficult to create a universal tool that can easily read and compare files as efficiently as git does for text files.</p><p>Today, CAD files require proprietary tools, usually sold by the application provider, to do “diffs” and “merges” to clearly convey exactly what’s changed in the files. <a href="https://www.durolabs.co/blog/2020/9/10/plm-based-part-numbering-system">When a hardware company uses several CAD applications across teams</a> it results in added costs for tools that are all incompatible with each other. Hence, more data administration and inefficiencies.</p><h3>This is a big problem.</h3><p>If hardware files don’t evolve and enable this same liberation, we will never grow and reach the same levels of innovation and efficiency that hardware teams deserve. Users should be demanding more from their CAD providers to raise the bar of expectations. There shouldn’t be so many barriers intentionally put in place to administer CAD data.</p><h3>How Duro is contributing to the solution.</h3><p>While CAD data is the linchpin to truly release the shackles restricting innovation, there are several other lower hanging fruits that we can start with. At Duro, we strive to make it as simple as possible to centralize, revision, and share the CAD, BOM, and supply chain data. Too many software tools are unnecessarily complicated and require weeks to months of custom configurations before hardware teams can even begin to see a return on their investment. Too much flexibility and variance has a measurable impact on adoption. This is a major contributing factor to the unnecessary overhead of data administration plaguing the hardware industry, today.</p><p>Data management tools should just work out of the box for the majority. Duro’s <a href="https://www.durolabs.co/products/#plm-professional">PLM Professional</a> package does exactly that. Hardware teams can be up and running in an afternoon with a complete centralized cloud data repository, customer part number (CPN) scheme, and change order workflow that are compatible with any supplier or manufacturer’s processes. In fact, Duro’s PLM software <a href="https://www.durolabs.co/blog/2020/9/23/duro-plm-scores-big-in-g2s-fall-2020-report">received several awards</a> for its simplicity and efficiency for on-boarding new teams.</p><p>Are you <a href="https://www.durolabs.co/schedule-demo?utm_content=db-2020-09&amp;utm_campaign=thoughtleader&amp;utm_source=medium">ready to innovate</a>?</p><p><em>Originally published at </em><a href="https://www.durolabs.co/blog/2020/10/27/why-your-next-hardware-project-should-look-like-a-software-project-lt6c6"><em>https://www.durolabs.co</em></a><em> on November 3, 2020.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=576c08d7c1da" width="1" height="1" alt=""><hr><p><a href="https://medium.com/duro-labs/why-your-next-hardware-project-should-look-like-a-software-project-576c08d7c1da">Why your next hardware project should look like a software project</a> was originally published in <a href="https://medium.com/duro-labs">Duro Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Got PRD? The Who and Why of developing a Product Requirements Document]]></title>
            <link>https://medium.com/duro-labs/got-prd-the-who-and-why-of-developing-a-product-requirements-document-d2144f536462?source=rss-a440d04f6820------2</link>
            <guid isPermaLink="false">https://medium.com/p/d2144f536462</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[hardware]]></category>
            <category><![CDATA[manufacturing-processes]]></category>
            <dc:creator><![CDATA[Michael Corr]]></dc:creator>
            <pubDate>Fri, 09 Aug 2019 00:00:00 GMT</pubDate>
            <atom:updated>2019-09-05T04:31:33.464Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zRS8aeDDT6kOB1pPZwj6Eg.png" /></figure><p>Time and time again, I see engineering teams approach me all bright eyed and bushy tailed with a great idea and a slick prototype, asking for help to bring their product to mass production. These are the teams I enjoy working with the most — those that understand there’s a significant effort to get a product ready for manufacturing. My first request is always for a copy of their PRD (Product Requirements Document), so I can figure out how to best help them. This is when their bright eyes turn into blank stares.</p><p>So many engineering teams will already have developed a prototype without any written list of specification requirements. This is a certain path to wasted time and capital — both of which, most startups don’t have a lot of.</p><h3>The Why</h3><p>I can’t stress enough how important it is to have a PRD for your product. Without a ratified PRD, too many details become ambiguous, forgotten, or simply misunderstood. For those familiar with the Software Agile methodology, I refer you to the requirement of well written User Stories. A PRD has similar goals.</p><p>I posit these questions to you:</p><ul><li>Without a written and approved list of specifications, how does your engineering team know what to develop?</li><li>Without a comprehensive list of features, how does your sales team know what to tell customers?</li><li>Without forecasted volumes and costs, how can a manufacturing team optimize the production line?</li></ul><p>I’ve seen engineering teams talk about their vision for the product, then start prototyping it only to figure out several features were not developed how they were originally envisioned. The team misunderstood the intention of the feature, because it was not written down with metrics to evaluate against.</p><h3>The Who</h3><p>Which brings me to my next topic: Who are the key stakeholders for the PRD? While the engineering team will likely become the most familiar with the requirements over time, several other key people are critical to get involved as well.</p><ol><li><strong>Product Manager</strong> — The PM is actually the primary author of the PRD. It’s their responsibility to put pen to paper the final vision of the product, incorporating input from all of the teams: engineering, sales, marketing, management, and even customers. While the leads of each of these departments have specific contribution responsibilities, it’s the PM who will aggregate their inputs and make the final decisions.</li><li><strong>Engineering Manager</strong> — The engineering team is responsible for being the reality check. Can the features outlined in the PRD be developed without defying the laws of physics, in the timeline requested, and for the targeted costs? Sometimes a small R&amp;D effort is required to answer these questions first, before the PRD can be finalized. That’s ok. It’s better to use a small amount of resources early to answer these questions rather than wait until it’s too late or too costly to recover from.</li><li><strong>Sales Manager</strong> — The sales team needs to determine if customers actually want the features listed in the PRD and if any fine tuning is required to ensure an optimized product-market-fit. The sales team is also responsible for determining how much customers are willing to pay for the product, as defined.</li><li><strong>QA Manager</strong> — The QA team needs to see early versions of the PRD so they can ensure they have the right resources to test it, and estimate any associated costs to procure them. These peripheral costs are often neglected when calculating the actual cost to develop, test, and manufacture a product. Test equipment can get expensive quickly!</li><li><strong>CEO</strong> — The CEO needs to review the product roadmap outlined in the PRD to make sure it aligns with the company vision and consistent with the business model.</li></ol><p>None of these disparate teams should work in isolation. Developing a PRD is a company wide effort, which involves several conversations and debates. Believe me, people’s negotiating skills will definitely be tested when the engineering team reports back that the features are too expensive or will take too long to build, and the sales team refuses to raise the price for the customer or remove a feature, and the CEO mandates the product be on store shelves within 6 months.</p><p>This Engineering-Management-Sales team triad will need to roll up their sleeves and work together, iterating on the specifications until all three teams are happy with the outcome. The Product Manager is the ultimate tie breaker and records the final decisions.</p><h3>What’s in a PRD?</h3><p>So, what’s in a PRD? Well, remember, the PRD should be detailed and exhaustive enough such that every team can find everything they need to complete their responsibilities. More specifically, the PRD enumerates all the features a product will have, to what performance metrics they must comply to, how many units will be sold, at what price, with what landed cost, and on what timeline.</p><p>In Part 2 of this series, I will describe these specifics in more detail.</p><p><em>Originally published at </em><a href="http://www.durolabs.co/blog/2019/8/7/got-prd-the-who-and-why-of-developing-a-product-requirements-document"><em>https://www.durolabs.co</em></a><em> on September 3, 2019.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d2144f536462" width="1" height="1" alt=""><hr><p><a href="https://medium.com/duro-labs/got-prd-the-who-and-why-of-developing-a-product-requirements-document-d2144f536462">Got PRD? The Who and Why of developing a Product Requirements Document</a> was originally published in <a href="https://medium.com/duro-labs">Duro Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Art of a Good Custom Part Number (CPN) Scheme]]></title>
            <link>https://medium.com/duro-labs/the-art-of-a-good-custom-part-number-cpn-scheme-8a66afb102f4?source=rss-a440d04f6820------2</link>
            <guid isPermaLink="false">https://medium.com/p/8a66afb102f4</guid>
            <category><![CDATA[hardware]]></category>
            <category><![CDATA[bom]]></category>
            <category><![CDATA[manufacturing]]></category>
            <category><![CDATA[cpns]]></category>
            <dc:creator><![CDATA[Michael Corr]]></dc:creator>
            <pubDate>Wed, 07 Aug 2019 00:00:00 GMT</pubDate>
            <atom:updated>2019-08-13T21:24:47.014Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/562/1*k0qLc73z4mOt4C15tCFFCA.png" /></figure><blockquote>The only important thing is to make sure there’s no room for interpretation.</blockquote><p>There is no single correct part numbering scheme for your hardware parts, but there are many wrong ones.</p><p>When building your hardware products and using a BOM (Bill of Materials) to define the discrete components, a Customer Part Number (CPN) is required for each individual line item. A more human readable name is also used, but these can be abstract and open to interpretation. So, fixed, alphanumeric strings with consistent formats is much more reliable, and the default reference used by professional hardware manufacturing teams.</p><p>Every hardware company is entitled to design their own CPN scheme that best fits the requirements and characteristics of their own products and processes — but it’s crucial to think it through properly and design one that won’t lead to mistakes in production.</p><p>The most important concept to keep in mind when designing your scheme is that:</p><blockquote>There must not be any room for subjectivity or interpretation. The rules for assigning a CPN to a new part must be clear and concise.</blockquote><h3>Terminology</h3><p>Let’s start with some basic terminology. There are two extremes when picking a CPN scheme: Intelligent and Non-Intelligent, and then variations in between.</p><p><strong>Non-intelligent</strong> CPNs are simple, usually 5–10 digit numeric values that are sequentially incremented and assigned for each new part as it’s added to your company’s part library — regardless of what the part is or does.</p><p><strong>Intelligent</strong> CPNs are complex alphanumeric strings where each character has a significance and meaning to it, usually tied to the parametrics and usage of the part.</p><p>Non-Intelligent CPNs are simple and fast to implement, and require no thought process when assigning new values. They are effectively limitless and infinitely expandable as your product evolves and new parts and part categories are added to your library. However, as a result of this flexibility, they are very hard to “read” and know what the part may be, without cross-referencing some other piece of information. This additional work can lead to delays and mistakes.</p><p>Intelligent CPNs on the other hand, usually can be deciphered from simply reading the CPN value itself (following a ruleset that explains the meaning of each character), such that one will know exactly what the part is just from the written CPN value. This provides a lot of efficiency in the procurement and assembly processes. However, it requires a lot of upfront design work and foreshadowing so you don’t design a scheme that ultimately paints you into a corner and doesn’t allow for a unique CPN value for some new part or part category added to your library. The worst thing to do is to modify the CPN scheme part way through production to accommodate a new part category that wasn’t thought of at the beginning.</p><p>The best solution is a hybrid of the two — where there is a short (3 or 4 digit) prefix code which defines the part category, with a longer (5 to 7 digit) counter suffix that is incremented with each new instance of the category. This is often referred to as a <strong>Semi-Intelligent</strong> CPN scheme. This design makes it easy to add new prefixes as new categories are added to your product in the future, but the granularity is coarse enough that it’s fast and simple to assign new CPN values when new parts are added to your library.</p><h3>The Chosen One</h3><p>My favorite scheme is a <strong>{3}-{5}</strong> format, with a 3-digit numeric category prefix, a 5 digit numeric suffix counter, and a hyphen between the two to create an easy visual separation to enhance readability. This allows for 999 unique categories and 99,999 unique parts per category. I challenge any team to run out of unique part numbers using this scheme!</p><h3>Do’s and Don’ts</h3><p>When designing your CPN scheme, there are several key things to remember:</p><ul><li>Do use consistent fixed length numeric CPN schemes</li><li>Don’t use alphabetic characters</li><li>Don’t use a mix of coarse and fine granularity in your category list</li><li>Don’t use variants</li><li>Don’t use leading 0&#39;s</li></ul><h3>Categories</h3><p>When selecting your list of categories, the only decision to make is how coarse or fine you want to make them. For example, you could simply have only two categories:</p><ul><li>Purchased</li><li>Custom</li></ul><p>For some teams this works just fine. Or you can go the other extreme and list out every possible category imaginable</p><ul><li>Resistor</li><li>Diode</li><li>LED</li><li>Capacitor</li><li>Screw</li><li>Bold</li><li>Nut</li><li>Washer</li><li>etc..</li></ul><p>The only important thing is to make sure there’s no room for interpretation, where two different people could reasonably have picked two different categories for the same physical part. This happens when category lists include a mix of some generic (coarse) terms and some specific (fine) ones.</p><h3>Variants</h3><p>I’m probably going to get some pushback on this next topic, since I still see many engineering teams implementing it — even though I think it’s a horrible idea. I’ve seen teams use a 3rd field in their CPN scheme to define variants of a part or assembly. It’s often represented as a 2 or 3 digit suffix, which defaults to 00 for the first instance of a part. When a “variant” of a part is added to the library, the base CPN value is kept the same, but the variant field is incremented. It’s used to show there is some form of relationship between two parts, but not enough for them to be considered exactly the same.</p><p>Why is this bad? Because it leaves room for interpretation. Two people may reasonably disagree on whether two parts are completely unique or just variants of each other. This can happen when the rules for what constitutes a variant aren’t clear.</p><p>Parts are defined by their parametrics. Anytime two parts have different parametrics, they are not considered the same, in a strict literal sense. But, all parametrics are not assigned equal priority by all people. So, engineers will result in subjectively selecting which parametrics constitute two parts being different and which parametrics are considered equivalent.</p><p>When you have a variants field in your CPN scheme, it exacerbates the problem, by empowering someone to make a subjective decision on the matter, which ultimately leads to inconsistencies in your parts library. It always happens: one group of parts were considered unique enough to warrant their own CPN values, but another group of parts were not. In isolation, these individual decisions may have made sense, but objectively, two different rules were ultimately used to make these decisions. Anyone not involved with these decisions (i.e. a 3rd party Contract Manufacturer or supplier) may get confused by the inconsistencies and result in making a costly mistake in production. Just don’t do it. The value of “grouping” like parts into variants is far outweighed by the risk of making a mistake in production.</p><p>What CPN schemes have you used that you’ve found efficient and easy to enforce?</p><p><em>Originally published at </em><a href="https://www.durolabs.co/blog/2019/8/7/the-art-of-a-good-custom-part-number-cpn-scheme"><em>https://www.durolabs.co</em></a><em> on August 7, 2019.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8a66afb102f4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/duro-labs/the-art-of-a-good-custom-part-number-cpn-scheme-8a66afb102f4">The Art of a Good Custom Part Number (CPN) Scheme</a> was originally published in <a href="https://medium.com/duro-labs">Duro Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Small design tweaks have a large manufacturing impact]]></title>
            <link>https://medium.com/@michael.corr/small-design-tweaks-have-a-large-manufacturing-impact-ea3520f53714?source=rss-a440d04f6820------2</link>
            <guid isPermaLink="false">https://medium.com/p/ea3520f53714</guid>
            <category><![CDATA[smt]]></category>
            <category><![CDATA[dfm]]></category>
            <category><![CDATA[manufacturing]]></category>
            <category><![CDATA[hardware]]></category>
            <dc:creator><![CDATA[Michael Corr]]></dc:creator>
            <pubDate>Wed, 08 Feb 2017 16:12:00 GMT</pubDate>
            <atom:updated>2017-02-13T16:24:40.951Z</atom:updated>
            <content:encoded><![CDATA[<h4>Easy DFM changes help to account for natural component and assembly variance</h4><p><a href="https://www.durolabs.co/blog/2016/10/18/a-tale-of-two-tests">In a recent post on the Duro Labs blog</a>, I discussed the natural variance that occurs in manufacturing. Variance can be found in both the raw parts and the processes used to assemble them. In order to have a more mature product design, it’s imperative to understand where these variances come from and how to account for them in your designs. If you choose to ignore this fact, you greatly increase the risk of low manufacturing yields and high product failure rates.</p><p>An example is the +/- 0.1 mm dimension tolerance specified on a 5 mm diameter screw. This means the individual screws will have actual diameters anywhere between 4.9 mm and 5.1 mm. If the clearance hole for the screw doesn’t account for the increased screw diameter, you will most certainly have issues assembling the parts for some percentage of your products. Repeated over a large number of operations, the production yield will fall which will lead to higher production costs to achieve the same finished goods quantity.</p><p>These variances are unavoidable, but can be accounted for in your product design. The first step is to understand and accept Gaussian variance as a fact of life. Then take the steps to improve DfM (Design For Manufacturing) and DfA (Design for Assembly). This will lead to a more professional and higher quality product design.</p><p>To manage variance, one extreme approach is to simply specify tighter tolerances from your parts suppliers. In the screw example above, this could mean requiring a +/- 0.01 mm (or smaller) tolerance. While this would ensure a higher yield in your product assembly, it would also result in higher prices to you since the part manufacturer will need to invest more resources to ensure they can meet these stricter specifications. And guess who has to pay for these resources? A better approach is to simply account for the published tolerance ranges of your parts and processes when designing your product from the beginning.</p><h3>How Do You Account For Variance?</h3><p>Mechanical parts are easier to control, test, and design for variance. Part design tricks can hide mechanical imperfections caused by variance in their fabrication. The first rule is to recognize that all mechanical parts will have dimension variability. So be sure the junction points of these parts have enough play to accept this.</p><p>Electronic components have variance as well, but can be more subtle in their impact to your production yield. Like mechanical components, electrical resistors, capacitors, inductors and other passives have published tolerance ratings in their datasheets.</p><p>Start your design with lower tolerance parts (5% to 10%). These will be the lowest cost and widely available parts. Only narrow the tolerance (0.1% to 1.0%) if you’re not achieving the required feature performance. Even then, only select the individual parts which require tighter tolerances. Your circuit can meet your product specifications without tightening tolerance on all components. Work through an error budget to determine the critical components in your design to expedite and validate this process.</p><h3>Electrical Part Assembly Variance</h3><p>The more subtle variances in your electrical circuit designs are from the PCB assembly process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/714/1*6UQ22fIfPHcfP-tv6REbbQ.jpeg" /><figcaption>A component being placed and soldered by hand</figcaption></figure><h4>Hand Soldering Electrical Components</h4><p>The most obvious risk in assembly comes from human variance. If your circuit design and component selection requires manual solder assembly, you are at risk. In this case, the variance between assembled instances of your product becomes inevitable and is not advised. Even with rigorous assembly and inspection policies, humans are inherently inconsistent and it’s not worth the risk. Remedy this problem by avoiding PCBA designs that require manual solder assembly. Furthermore, hand soldering can result in poor soldering joints which can lead to big, big problems down the road including costly product recalls.</p><h4>Automated Soldering of Through-Hole Components</h4><p>The automated assembly process for thru-hole electrical components is called wave solder assembly. While wave soldering is better than hand soldering, it still has a notable amount of variance in its process. Though the physical application of solder is automated, the parts are still inserted on the PCB by hand. So, these small variations in placing parts into their designated footprint holes introduces variance and risk.</p><p>If your circuit inevitably requires wave soldering, there are best practices to follow. One of these is to provide sufficient clearance around the thru-hole part leads. This allows the use of a wave pallet to mask out any peripheral components not meant to come in contact with the wave solder. Understanding this requirement in the design stage will save time when moving to assembly. However, the wave solder process is messy and has poor resolution. So still try to avoid using components that require this process whenever possible.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/729/1*GsvuY_TKwrhKFagnVpkQDg.jpeg" /><figcaption>Preparing Parts for Wave Soldering, Using a Wave Palette</figcaption></figure><h4>SMT Assembly: The best approach</h4><p>The best electrical component assembly process to use is reflow soldering with SMT components. This is due to the high degree of automation used throughout the entire workflow. Solder paste application, component pick-and-place and final reflow oven assembly are optimized for minimal human involvement. Your contract manufacturer (CM) may decide to insert a person for inspection or handling between these assembly stages for quality control purposes, but the equipment itself does not require a human “in the loop”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_rDpvWxuQJ-eaIEMpwtuFQ.jpeg" /><figcaption>Automated Pick-And-Place SMT Machine</figcaption></figure><p>Automation inherently reduces variance and risk — but it doesn’t alleviate it completely. For example, a solder paste printer introduces variance in the amount of applied solder paste between PCBs. While the CM is responsible for minimizing this variance, good footprint design techniques will lessen the constraints.</p><p>This starts with ensuring SMD footprint pads are the appropriate size. Dimension specifications can be found in the part manufacturer’s datasheets. Start with these published recommendations but confirm with your PCB assembler that the pad sizes are conducive to the equipment they specifically are using to assemble your boards. PCB assemblers regularly request modified footprints to improve solder-ability and yield. Moving this discussion into the design stage of your product minimizes schedule delays when moving a product into higher scale manufacturing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NlPzCEpNf0rNVQlB2I6Wfg.jpeg" /><figcaption>Inspecting PCBAs After Reflow Oven Assembly</figcaption></figure><p>Like through hole components, SMD footprint pads on the PCB should have adequate clearance. This minimizes the risk of electrical shorts between pads. It also increases access to the pins for the heat source in the reflow oven — resulting in a cleaner and less variant solder joint.</p><p>Industry best practices will provide guidelines on exactly how much clearance you should provide between footprint pads. But again, always talk to your assembler to see what they recommend as they are the ones doing the actual assembly and you want to set them up for success.</p><h3>Size Does Matter</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-fLWvvvUeK5BGktswXmlEw.jpeg" /><figcaption>Various SMD Packages</figcaption></figure><p>SMD parts come in multiple package options for the same electrical values. This supports different power ratings and allows for different sized designs. Many PCB assembly suppliers will brag about how small of a component they can reflow, but you can ignore this as it just adds risk in my opinion. I start all circuit designs with 0402 packages. It is a very common size and widely supported by SMT equipment. It is also still big enough to replace by hand, if necessary, during development or rework.</p><p>With reference to the spacing risks I addressed above, the 0402 package is still small enough to allow good clearance between parts. Only when a component has a higher wattage requirement will I increase to a larger package (0603 or 0805). Going smaller to 0201 packages greatly increases the risk for poor solder joints and is not worth it. The only reason to use any package smaller than 0402 is if real estate space on the PCB requires it, such as in small products like wearables. In this case, I strongly advise you to work with your assembler. Spending time with them to optimize footprint pad size and component clearance will ensure a high yield assembly.</p><h3>A discussion with your assembler is an investment in your future product</h3><p>The best DfX advice I can give is for you to talk to your assembler. Yes, there are industry guidelines and best practices for getting started. But at the end of the day, it is your assembly supplier that will be doing the actual work. Your designs will have the greatest success if they comply with your assembler’s specific equipment and processes. Talking to them early in your product’s design cycle allows your engineers to incorporate their guidelines before it’s too late or too costly to remedy.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ea3520f53714" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The 6 Stages of a Successful Product Lifecycle]]></title>
            <link>https://medium.com/duro-labs/the-6-stages-of-a-successful-product-lifecycle-f34736f16d77?source=rss-a440d04f6820------2</link>
            <guid isPermaLink="false">https://medium.com/p/f34736f16d77</guid>
            <category><![CDATA[evt]]></category>
            <category><![CDATA[plm]]></category>
            <category><![CDATA[hardware]]></category>
            <category><![CDATA[dvt]]></category>
            <category><![CDATA[manufacturing-processes]]></category>
            <dc:creator><![CDATA[Michael Corr]]></dc:creator>
            <pubDate>Tue, 03 Jan 2017 00:00:00 GMT</pubDate>
            <atom:updated>2020-03-09T18:40:08.092Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qVy-FuSnXmqJdv3E5g5SCg.png" /></figure><p>Any hardware product has discrete stages in its lifecycle from concept to completion, or more colloquially: <em>from Cradle to Grave</em>. Each of these stages has a purpose with its own set of goals and exit criteria. If not administered properly you greatly increase the risk of a less than successful product launch plagued with delays and budget overruns. In the worst case, you completely miss the market window and your chances of any success!</p><blockquote>Did I scare you, yet?<br>Good. Now let’s start at the beginning…</blockquote><p>Once a product has been conceived, whether it was from a late night fueled by alcohol and bar room banter or a strike of lightening inspiration leading to a quick sketch on a napkin, the Product Lifecycle begins.</p><h3>The 6 Stages of Any Hardware Product Lifecycle</h3><ol><li>PRD — Product Requirements Documentation</li><li>EVT — Engineering Validation and Testing</li><li>DVT — Design Validation and Testing</li><li>PVT — Production Validation and Testing</li><li>MP — Mass Production</li><li>EOL — End of Life</li></ol><h3>Let There Be Light</h3><p>The first step is to write down the requirements in a formal document called the PRD (Product Requirements Document). Why is this important? Because without requirements or performance specifications, you won’t ever know if you’re done developing. What size does the product need to be? What temperatures does it have to operate in? How fast does it have to move? How bright do the LEDs need to be? What age is the intended user? How long must the battery last? You get the idea. Without these specifications formally laid out, you run the risk of going down a rabbit hole and spending too much time and money on something that is not required.</p><h3>A Proper PRD Includes</h3><ul><li>The complete list of features to be included in the product</li><li>The specific performance metrics each feature must meet</li><li>Estimated production volumes</li><li>Target costs</li><li>Target product release timelines</li><li>A product roadmap</li></ul><p>The Product Manager is responsible for leading and authoring this document, but they must solicit input and final sign-off from all the other team leads: Engineering, Sales, QA, Executives, and Marketing. Once a revision of the PRD has been approved by all these team leads, it becomes your dogma and all subsequent product decisions must conform to what is written in the PRD. It is acceptable to ratify the PRD and create subsequent revisions as new information becomes available that impacts your earlier decisions, but each new revision must be reviewed and approved by the team leads, as well.</p><h3>Prototyping</h3><p>Next comes the fun part, more formally known as EVT. In this stage the engineering team uses whatever means necessary to implement each of the features outlined in the PRD. The goal of this stage is to prove that it is humanly possible, without defying the laws of physics, to design and build an instance of the product that meets the functional requirements in the PRD. But keep in mind that the components selected and the processes used to assemble them <em>don’t necessarily have to be those that will be used in the final version of the product</em>. Duct tape and hot glue are ok at this stage, as long as they allow you to build something that functionally works and passes your specification metrics.</p><p>The primary goal of Engineering Validation and Testing is to <em>identify any and all risks of the requirements outlined in the PRD and find ways to remove or greatly mitigate them</em>. This may take several iterations and millions of dollars in research and development, but a product doesn’t leave the EVT stage until all the <strong>functional requirements</strong> and performance metrics have been met.</p><h3>Designing</h3><p>In the DVT stage, the goal is to work your way towards the final look and feel of the product. This is when you start selecting materials and mechanical designs that meet the final form, fit, and aesthetic requirements outlined in the PRD. Materials and components selected in these revisions are candidates for the final production version. If possible, instances of your product built in the DVT stage can be shown to potential customers to start soliciting feedback and testing <strong>product-market-fit</strong>. This is actually valuable to do at this stage before you begin investing significant capital dollars for the materials, tools, and processes required to start manufacturing.</p><p>A product can exit the DVT stage once there is an instance built that meets all <strong>functional and aesthetic requirements </strong>in the PRD while using candidate designs and components for the final production. This version may in fact be very close to your final product implementation and be indistinguishable to the untrained eye, but possibly created with lower volume manufacturing and assembly processes.</p><h3>Pre-Production</h3><p>Now it’s time to relinquish control and engage in serious conversations with your suppliers and manufacturers. This is the PVT stage where you are confirming your DVT product can in fact be produced at the volumes and target costs listed in the PRD. Since more often than not you will be outsourcing the actual manufacturing and assembly of your product, you must engage with these suppliers to solicit feedback of your design to make sure you can achieve the production goals.</p><p>Hopefully this only requires a few small changes to your DVT model, like replacing a part with a cheaper pin-compatible version or slightly changing the draft angle on a custom mechanical part so it will eject more reliably from the injection-molding tool. But, in some worst case scenarios you may have to go through major overhauls or design changes to make your product ready for mass production. For this reason, it is best to actually begin talking to your suppliers well before you enter the PVT stage, so you can help ensure it will be successful by incorporating some of these DFX (Design for Manufacturing, Design for Assembling, Design for Testing) principles upstream in the product lifecycle.</p><p>The other goal of the PVT stage is to begin solidifying the production processes and setting up the supply chain. A great way to do this is to simply build a pre-production run of your product using the <em>actual</em> supply chain and manufacturing assets. This is the best way to see where the weaknesses are in your production ecosystem and determine if you’re ready to start increasing the volume and crank it up to 11. You will be working very closely with your manufacturer at this point — often on-site. So, don’t expect to be getting anything else done in the office.</p><p>To exit PVT, the manufacturer must sign-off that the supply chain is setup. This means parts can be ordered at acceptable lead times and pricing and the assembly and test procedures are repeatable and reliable. You must also sign-off that the performance and quality of the final products meets your standards. Don’t wait until Mass Production to check this — it will be too late.</p><h3>Products for the Masses</h3><p>The scariest yet most exhilarating stage is the Mass Production of your product. Of course “Mass” is a relative term and specific to your product and market on exactly how large your MP sizes are. But, when a product reaches MP it means you are producing and selling customer ready versions of your product at some meaningful quantity. Again, depending on your product and market on how big and how often each production build is, but more often than not, a product is produced in several batches over time with each batch increasing in volume size.</p><p>At this stage, the bulk of the responsibility is on the manufacturer to procure the parts, assemble, and test them to meet your production demands. But, you still have responsibilities to ensure the production quality and yields stay high while costs start to decline. You must support the manufacturer as necessary to achieve these goals. This may include working with procurement teams to build up a strong supply chain to keep costs and lead times low, or working with the manufacturing teams to continuously find ways to improve efficiency, increase yield, and lower production costs.</p><h3>C’est La Vie</h3><p>A product may come to its end of life (EOL) for various reasons. The best case is you have developed its successor, which is better, faster, cheaper, stronger, smaller, whatever. And so the original product is stopped or ramped down in its production. The worst case is a major flaw was discovered and the production line needs to be halted immediately and all existing products possibly recalled. The point is, one doesn’t simply stop production of a product when its time has come. There are several peripheral tasks that need to be implemented once the decision is made to EOL a product.</p><p>First, what is to be done with the existing completed products, also known as Finished Goods? Are they to be recalled and destroyed? Can the sales team continue to sell any remaining products in Finished Goods Inventory (FGI) but not order any new ones to be produced? The same questions must be answered for any remaining raw materials and components that haven’t been assembled or consumed yet. Can they be returned to their original suppliers for a refund? Can they be saved for use in other products? Do they need to be disposed of? The answers to these questions are referred to as: <strong>Product Disposition</strong>.</p><p>The point is, there is often a lot of capital invested in the finished goods and raw materials, and it’s important to leverage them as much as possible either by converting them back into liquid capital or moving them into another product revenue stream. Unfortunately, there will always be some loss where goods must be disposed of at a cost, but hopefully through good planning during the EOL stage this loss can be minimized.</p><h3>The Devil is in the Details</h3><p>The goal of this article is to articulate the fact that there are subtleties to the lifecycle of a product, and it’s imperative they be taken seriously. A good Operations Manager will clearly define the Entry and Exit criteria for each of the discrete stages, minimizing “waste” and increasing productivity.</p><p>Too often we get caught up in the excitement of developing a product and bringing it to market and don’t set ourselves up properly for success. Only the most successful companies know to take a breath, clear their heads, and give due attention to all the peripheral tasks associated with Product Lifecycle Management.</p><p><em>Originally published at </em><a href="https://www.durolabs.co/blog/2016/12/16/the-6-stages-of-a-successful-product-lifecycle"><em>https://www.durolabs.co</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f34736f16d77" width="1" height="1" alt=""><hr><p><a href="https://medium.com/duro-labs/the-6-stages-of-a-successful-product-lifecycle-f34736f16d77">The 6 Stages of a Successful Product Lifecycle</a> was originally published in <a href="https://medium.com/duro-labs">Duro Labs</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>