In the current era where we have so many reference designs from semiconductor suppliers and open source projects, product engineering is no more a secret sauce of big and established companies. Many individuals/startups that are low on capital but high on innovation, can bring out novel ideas to life. These organizations solve the day-to-day problems and add synergistic value to existing systems. With the abundant material online, they are able to quickly get a proof of concept (POC) of their idea. And with more accessible ways of raising funds through crowd funding websites, they start the journey of scaling functional PoCs to robust shippable devices. However, in reality, very few organizations are successful in scaling their products, from PoCs to mass manufacturing. Possible reasons for failure are many, prominent among them is “in-sufficient thought to product engineering”.
In this article, we synthesize our years of experience in all stages of product engineering over a multitude of high-volume products.
Ideation to Proof of Concept
The ideation phase is one that seeds the whole product development and it kick-starts the phase of proof of concept (PoC) development of the product. It is difficult to build a custom solution that fulfills the needs of the idea. The most obvious approach is to start from something already available and build a quick PoC. In our opinion, it is worth to leverage on the off the shelf solutions and customize them to bring them as close to the concept as possible. But, going in this direction is also a challenge. With the current technology state, there are many off the shelf solutions, that will look like a very good choice. Hence, the selection of one among the available choices becomes a tough proposition. Choosing the right solution will be the key success. The solution starts from the CPU module. Choosing the most common module used in the mass market is an obvious choice as it brings open forum support handy when needed and time to market looks very short with this choice.
One of the easiest pitfalls during this phase could be to choose a semiconductor component that is the end of life [EOL] and land in a tough situation. Hence it is important to consider the following:
- Identify which industry does the product idea belongs to - consumer or automotive or industrial. Select a semiconductor-based on the required longevity for each category.
- Semiconductor product support team is there and it is not an end of life product. During scaling for manufacturing, this becomes a key for getting any production-related anomalies. As simple as the package is giving low yield in the assembly line.
- Proof of concept realization should be achieved with minimum possible investment and at the same time, the scalability option also should be there.
Human Resources: Partnering vs Self Staffing
To build a PoC or a product, human resources is a key success factor. Whether it is during PoC or after successful POC, human resource plays a huge role. Getting the right team to scale the POC to mass manufacturing and shipping is a challenge. How can one solve the problem of human resource/building the team? How to influence the talent? Is it worth to build the team or outsource the execution and manage? This indirectly impacts the budget in the long term, but choosing wrong options can decrease the probability of delivering the product on time. Hence this has to be managed cleverly.
For a startup, getting the right talent is neither so easy nor so judicious due to various reasons. Hence, it is always recommended to partner with consulting and product engineering services firms who have a credible history of helping startups to scale. Hiring consultants might look expensive, but it is a good way to start. Consultant experience will help in learning dos and don’ts for transforming working POC to mass production.
How to choose the right partner?
As discussed above, a lot of organizations prefer working with partners/consultants while building the whole team. But, how does one choose the right partner? What are the things that need to be considered for selecting the ideal partner for you? Some of the key things to consider include:
- Has the product engineering services company deployed products on mass scale?
- Has the company been successful in engineering products in your specific target industry?
- Does the company have working experience with large manufacturing companies and supply chain contacts?
- Most important is the experience and commitment of the leadership team
Getting answers to such questions helps in identifying the company or consultant.
Partner says “Re-design the product”. Is that right? How to evaluate?
As part of the engagement, many a time, a consultant’s first recommendation will be to redesign the POC. This is one of the tough decisions that need to be taken. Redesigning means losing time and re-doing everything from scratch. This is the critical phase of product development where the report provided by the partner needs to be evaluated. But how? One of the good approaches is to go with following categorization of the feedback and evaluate.
- Design for Manufacturability & Design for Testability – Electrical, Software and Mechanical
- Software System
Based on our experience, following common mistakes/pitfalls can happen under each of the mentioned bucket. Consultant feedback and recommendations could be possible to map to this list. Hence, evaluation of the consultant can be weighed based on few of the aspects covered here and partner as needed.
Design for Manufacturability & Testability
Being focused on proving the concept, this is one of the foremost factors, which could have been missed. What exactly does this mean? And why does it take the highest priority? The reason for it to take the highest precedence is that it directly influences the cost of per product production and squeeze the possible margins. This is not a one-dimension problem, but has multi-dimensions on all aspects of the engineering of the product. Starting from as simple from software and going to electrical and mechanical engineering. It directly impacts the yield rate of the product which can further hamper the possible profit. Let’s see, what could be simple pitfalls that we could land into, during each of the product engineering phases.
Electrical Engineering starts from the component selection to complete design.
Component selection is the most critical aspect of product engineering. The selected components need to support the right voltage rating and have the right physical properties for the design to be reliable for over the product environmental conditions. This becomes very critical when working for an industrial or medical product. Another important aspect is End of Life (EOL). This becomes very critical as ODM will not be interested in working with the EOL component fearing no support from supplier and it may become a huge liability. This is one of the most common issues faced by many startups. Reason being that, unless the product is available in the mass market and matured for a good time, you will not have the right collaterals and support for developing on own. But this point has to be evaluated very closely and decisions need to be taken as soon as a decision to mass scale is decided.
PCB Gerber is very important for manufacturing after having the design done. IPC is the standard body that defines the component land pattern for assembling and most of the manufacturing equipment which are build are compliant to this. There are several IPC standards that govern this and depending on whether the product is for consumer or industrial, IPC standards vary. The overall motive is to reduce the cost of assembling & verification of the PCB unit. Few of the common pitfalls, which escape during PoC stage:
- PCB Stack - PCB consists of various layers called core and prepreg, which on sandwiching we get the PCB. This directly influences the EMI/EMC performance of the design and cost of the PCB. If the PCB stack is not complaint to ODM or for best EMI performance, the design can fail in certification.  
- PCB Gerber generated is not complaint to recommended land patterns at different layers. The ratio of solder mask vs solder paste is one of the common mistakes.
- Mixing up of the height components which will not facilitate in a single flow of reflow and wave soldering. Requiring both side wave soldering and requirement to perform hand soldering.
- Missing electrical test points for verifying the design in an automated test jig. Test pad dimensions and positioning. This increases the test jig cost manual testing
Selection of software configuration is greatly influenced by the hardware design and component selection. We will discuss more on this in the software system section, but will cover a few aspects which are relevant for manufacturing and testing. Software goes hand in hand with electrical systems. One needs to ensure that the device is easily configurable for different SKUs (Stock Keeping Unit) and have tracking methods able to track from PCBA testing to final product shipping. Provision for a serial number, identification of SKU and an interface for production logging and tracking are a must. Sometimes it is worth having two different software versions where one facilitates in production and another for shipping configuration. Automation of capturing the information and configuring the device is key to avoid manual errors.
The major pitfall that happens is that the test suite for the system will be drawn from a designer’s perspective rather than from the user perspective.
If the unit has to undergo different work stations for testing and quality check, ensuring all tests have passed should be trackable. This is very much needed to ensure that the complex product passes the final quality test and there is a way to verify at each stage. Sometimes this is missed as most of the time it would have been done manually during POC and automation becomes a cumbersome process.
As the saying goes “First impression is the best impression”. Mechanical engineering is hence the key in product engineering. For any product development, concept ID (Industrial Design) is the first design activity that starts. A smart mechanical engineer transforms the design to ID assembling parts and directs the electrical design which indirectly dictates the performance capability of the software. Hence, mechanical engineering plays an important role in the whole product designing. It can make the product or break the product.
In the age of miniaturization expectations and computing units capable to scale to small form factors, starts the tug of war for the realization of ID and electrical engineering problems. During POC, ID design realization takes precedence over anything else and that creates a lot of pitfalls.
Post this the splitting of the PCBA assembly happens. During this instance, as in POC we focus on ID, there is every possibility that cabling and component placement goes for a toss without much simulation and analysis. One of the classical example is the placement of a Wi-Fi antenna behind a metal sub-part assembly. That completely degrades the Wi-Fi performance it is a common pitfall for product failure . With more boards splitting, cabling and interfacing sub-boards become critical. This is a major pitfall where it can end up in selection of low-cost contacts that work very well on bench and when it is assembled into the mechanical, contacts cannot hold the contact or fail certification tests for EMI and EMC due to long cables. IO interface exposure for servicing and examining the product in the production line. This is the often missed part in the PoC stage.
On mechanical engineering, for meeting the ID, non-conventional mechanical parts for assembly could have been selected. This poses a big problem for scaling by directly influencing the production time, which increases the cost of the per-unit production. At the same instant, producing the component and material also could be an issue that increases the sourcing prices. It could be also a situation where the mechanism poses reliability in usage and assembly. It can come from as simple as cable assembly of PCBA. Cable assembly and disassembly becomes a major concern from service and maintenance point of few.
Software SystemSelecting a software configuration for the product is very critical for the maintenance and scalability of the product. During the PoC stage, for the constrained time and resources, it would have been chosen to go ahead with open source and off the shelf components. Later as time evolves, the component may not be maintained by the open-source community, and scalability would become a problem. If you consider any software, there are several layers that would be required to match, for the software to be functional. These span Hardware Abstraction Layer (HAL/BSP), Operating System (OS) and application components. Selecting a semiconductor whose BSP and OS doesn’t get evolved as time progress, will kill the program to a greater extent. But the inclination for such a selection would be use of this component for mass market and open support forum. At times, once the component reaches EoL, semiconductor vendors may have to stop support. This is very common to happen with the gestation of OS version getting as low as to yearly. This could change the landscape of the software configuration that is compatible for the product and may land in non-compatible software configuration.
Certification is very crucial for the commercial deployment of the product. It requires support from all engineering domains. One of the common mistakes would be to choose a software stack that has not been certified and also have a situation where having a certified logo is the mandate for shipping the product. This is a big problem. Have a PCBA layout, which will influence the product passing the EMI/EMC certification. ID/Mechanical design can have reliability issues in meeting the targeted market segment. As simple as, ID doesn’t meet the drop test requirements for a consumer grade product. The material selected doesn’t suite the thermal profile of the product and poses hazard and become a blocker for shipping the product. These are few pitfalls in certifying the product, but at every stage of the program, it requires us to monitor the same. It is very crucial to select the HW which has the right software support to realize the product and satisfy thermal needs for the ID which is needed.
The journey of product engineering, from concept to manufacturing, looks simple. But there are many aspects that need to be considered to be successful. At any instant of the time, human resources play an important role and choosing a wise team or partner is the foundation for success of the product to ship. From perspective of certification, it would be wise to do some pre-certification tests and keep a tab on the passing capability of the product for regulatory bodies.