How to Survive the R&D Process

Great ideas for innovative industrial printing projects often don’t make it to production, a sad and often expensive outcome that can be avoided.

In the world of printed functional products (including appliances, consumer electronics, transportation equipment/controls, medical devices, sensors, and printed electronics of all types), only a small fraction of new product designs ever make it to production. All too often, game-changing – even life-changing – product designs are shelved, never to be seen for reasons that range from poor production feasibility to excessive time to market.

The reality is, most new products never make it out of R&D, and it’s not because the design or product type is not in demand or feasible to produce. More often, jobs are greenlighted too soon, by people who don’t ask the questions that would enable a different outcome.

Here is a typical scenario common to many of these canceled designs, especially with consumer and medical electronics: A manufacturer wants to upgrade or introduce a variant of an existing product with a new function, shape, appearance, materials, or packaging. Because the project only involves updating an existing product, it’s assumed at the quote stage that everything should proceed smoothly. When it doesn’t, everyone begins making assumptions (usually incorrect ones) as to what went wrong.

The question that should always be (but too seldom is) asked at the quote stage: “Has the product design advanced far enough to begin R&D printing?” Early in the design of a product, such printing should come after an initial trial-and-error process of establishing the “proof of concept” (PoC). In PoC, virtually any tool or method, including hand printing, can be used to find the best device configuration for the materials to be used. This step is oriented to the functionality of the design. In the next stage, R&D printing, the research should verify that the PoC design is robust, and the development process of printing sample batches is designed to check whether the equipment and printing process fit the requirements of the job.

At least, that’s how it should be. Unfortunately, modern design cycles are compressed, high-pressure affairs. In order to save time, the early design research testing phase is often combined with development and sample printing. Once preproduction parts are on hand, they may be pushed into clinical trials, focus groups, and test programs too soon. Data gathered during these sample trials can change the design before it’s actually a finished concept.

There is an added risk when companies use sample production runs as process qualification pieces in advance of the final design actually being approved. This is a gamble that can either save the expense of several sample print runs or increase the cost of the R&D process by orders of magnitude.

The common theme running in the background of this compressed cycle is that some designs are not ready to progress from PoC into development printing, much less into process qualification runs. This unreadiness can be caused by a functional design issue or a PoC process that was simply rushed or incomplete.

So why are working printed designs deemed to be unsuitable for production? Two reasons are commonly cited. The first is “Excessive product cost per unit and/or excessive time or cost for R&D.” (While time is money in the world of R&D, there are actually two different problems here. Excessive time may just mean missing the customer window for product rollout.) The second is “Cost of retooling existing equipment, or need to invest in new processes to work with new substrates, inks, or tolerances.” Again, there are a number of different problems encompassed here that tend to be used interchangeably.

Collectively, these two failure modes may also be reported as:

• Slow production speed
• Incompatibility with existing equipment
• High QC/inspection costs
• Unstable production process prone to a high volume of defects

In the project post-mortem, this all tends to get grouped under the heading of “high cost per unit.” It’s rare that the root cause is assigned to flaws in the basic product design (since the product actually functions) or the printing and downstream production processes (since, after all, we know how to print). It’s just assumed that the product is too costly in one manner or another.

The two basic reasons that manufacturers give for canceling designs are pretty broad, and can hide a lot of contributing factors. Let’s dig a little deeper and take a look at them.

Excessive Cost/Time for R&D
Typically, a request for quote (RFQ) starts as a question: “Do we have the capacity, or budget, to do this profitably?” It would be better to phrase this assessment of the project like this: “Is this project within our capability? Do we have the knowledge of the new materials and processes that the project entails? If not, is the opportunity lucrative enough for us to retool for it?”

The risk at the quote stage is that the job may not appear to be unusual. Unless new print tolerances, substrates, inks, or device functions are clearly involved, the job may not look like something that’s bleeding-edge new. The product design may be very similar to others the printer has worked with before. But being similar does not means it’s the same.

It’s a cliché, but it’s true: Assumption is the mother of all failures. Mistaken assumptions about the degree of difficulty a project involves are commonly made both at the RFQ and in the PoC stage by the designer making the first device mockups. The designer compares the substrate, inks, form, quantity/time/scale equation, device stack-up, and other job specifications to previously printed parts. These first correlations can drive or support the initial answer to the customer that the company can indeed make the part, even if that is still unknown. An important question to ask at the RFQ stage is whether the material set and form factor fit the capabilities of the production line. But this is something that only a process design engineer or process manager can accurately determine, and these experienced people are rarely involved at the RFQ stage.

Process design engineers know that the function of a printed part may be altered by your normal printing method when the job involves a totally different type of device. Just being similar in material use or shape doesn’t mean the device will have the same function. What may be normal print tolerance for texture, deposit, edge resolution, or porosity within your plant may lead to failure with a new design. In other words, the materials and form of the job may fit the existing production, but will that process change the function of the part? Only an experienced process engineer will know.

From the viewpoint of the material and sub-component form factor, many printed electronic devices may seem almost the same. They may all have traces, use the same conductive inks, and be printed on the same substrate type. However, end-use functions can drive considerations that make them different to manufacture on the same equipment.

For example, for a membrane switch, the key considerations may be trace resistance, width, pitch, and lifespan. In an industrial control product, the important issues may be vibration resistance, weather sealing, and flexibility. In a printed medical device, the most critical functions may be telemetry signal frequency or the ability to interface with other liquids or chemicals. Whatever the most important considerations, the design engineer should keep them front-of-mind when crafting the PoC part to demonstrate these functions. The “R” part of R&D will then find out what the production process will do to this functionality.

A variant of this risk can happen when production staff produces parts at the sample and qualifying print-run stages without knowing that the finished component has abnormal sensitivities to characteristics they would consider normal for similar parts. Often, they won’t know this because a design engineer did not test for it; the process went straight into development.

When these types of unknown requirements cause the test batch devices to fail, that’s when the label of “excessive cost for R&D” is usually applied. The device then gets kicked back to the designer for revision, which is too far backwards in the process. Had the right questions been asked and answered early on, the problems could’ve been addressed through simple process revisions to improve production efficiency or reduce printed defects in an otherwise good design.

Instead, these revisions push the product out of development and back into design, when it should be in research. Time and money is spent on redesigning the devices and specifying different raw materials, which are big changes – ones of sufficient magnitude that will alter the production equation yet again.

Excessive Cost of Retooling Existing Processes
This label is sometimes expressed as excessive cost per unit or unfeasibility, and it involves much of the same missing data that causes jobs to be rejected as having excessive R&D expense. Many of the same fundamental issues need to be addressed, though this time at an administrative level. It all circles back to whether or not the company should be taking the job in the first place. These three questions must be asked (and the answers should already be known):

1. If the PoC printing/construction stage indicates a process or material that is not compatible with your existing printing and downstream equipment, or is unknown to you, then why are you taking this job?

2. Do you know what your exact imaging capabilities (resolution, speed, and size) are? At a finer level, can you establish what chemical, curing, pre- and post-treatment, and other capabilities you have (and don’t have)? If you answered “yes,” is that based on real production data or on the theory that your design and production team will innovate under pressure (the philosophy that they will always get it done, somehow)? If you answered “no,” is that based on data from your facility establishing your production capabilities, or was it taken from the equipment manufacturer’s datasheets? Can a new technique, accessory, or chemistry augment those capabilities to expand what you are capable of doing?

3. Has a cost-benefit analysis been done to assess whether you actually need to retool physically versus simply developing a new technique (different screens, substrate handling/prep, etc.) or outsourcing segments of the process?

The reason that many companies have difficulty answering these questions in the short space of time allowed for quotation and acceptance of a project is that they haven’t gathered enough data on their own capabilities. They may know what they can do with materials and devices from their past and present, but that may not help them when a job involves new raw materials or unfamiliar device functionality.

There is a good business case for what might be called “speculative R&D” to help quantify these unknowns. Budgeting the time to do testing and trial designs for new product types and materials can give you information that will pay off down the road. It’s the foundation of continuous process improvement that can expand your future capabilities.

Finally, as the process flowcharts in this article suggest, you can also prevent “death by R&D” by getting the right members of your team involved at the right time. When an RFQ for a new project comes in, your designer may be involved as the contact point, but don’t wait until the project has already been accepted to get your process engineer into the decision. Those assets must be interlinked. Lack of collaboration between them predetermine a project’s failure before the job ever gets to production, and after a lot of money has been spent on R&D.

Read more from Screen Printing's October/November 2017 issue.

View more from this Screen Printing issue