Poll: Are You Interested in element14 Offering Reference Design Activities?

I would have to say that it strongly depends ...

The RoadTest program already offers up EVKs which are usually a reference design of sorts, but the EVKs themselves are not designed to be integrated into other equipment for reasons of compliance, potential changes in future revisions and legal implications. These can already require quite a bit of time to properly evaluate, especially if things go wrong which has happened on occasion. However, the manufacturers which are serious often already have the documents available, albeit usually under a sign-up access or by NDA. Whether the data for these manufacturers could be "republished" on element14 seems doubtful.

A middle ground could be EVKs for modules where a sample of the modules are provided for own designs in RoadTests, but also where the modules can be populated at home with a minimum of equipment (e.g. a standard rework hot-air station), but even that (and the necessary consumables) may not be readily available to the majority. These could probably fit Design Challenges more, as it is very much design, but reduce the risk as the difficult part of the design has already been (mostly) done. If there are "build-it-yourself" EVKs that are a bit like a kit, then I'm not sure how much one will learn from it. I do like building kits as a therapeutic exercise, but as the design is already set, I suppose as long as you have sufficient construction skills, it should work just fine. But maybe there is a middle-ground here with "mix-and-match" EVKs where the platform may be one part, and the individual sensors or parts could be interchangeable depending on the application - in that case, selecting the appropriate part for an application would make it easier to evaluate in a given application, rather than just evaluating the part we are given by shoehorning it into a non-ideal scenario.

To have a reference design activity of the sort where the design data is made available for the user to modify, integrate into their own PCB design, get manufactured could be interesting but seems to be an order of magnitude more difficult. A barrier could be the simple choice of PCB EDA software as many reference designs are made in the "industry standard" Altium which is often out of reach for hobbyists. Another barrier could be the design complexity and expertise required especially if the design involves any high-speed digital buses, RF impedance control or high-power/high-dissipation components. A failure in the design could prove to limit the ability to evaluate the part entirely, or cause the part to perform sub-optimally. The time-frame necessary could easily blow out since the boards will need to be manufactured - for some component reference designs which are intended to maximise performance, this could mean manufacturing using costlier 4/6-layer processes and some components (e.g. BGA) may not be amenable to populating "at home" either. Not to mention the supply chain issues that are at play at this moment in time. It would be a very limited number of people who may be able to take something like this from start to finish - but how will element14 support the activity? Would their own contract manufacturing team use their contacts to aid in design/manufacture, or would participants expect to "go-it-alone"? How many people would be able to take it up given the time investment? What happens if they don't get it right on the first try?

I think there is a lot of potential risk here - but it is definitely a good "acid test" for whether a part is easy to use in a design, the relative strengths of a given part and whether the documentation is sufficient. I can't say I would necessarily be game to do it as I feel that my design skills may not be quite up-to-par, but even then, it's often not only the design but the debugging that comes afterward to salvage a partially successful design.

- Gough

The learning opportunity can also be part of the appeal. The opportunity to gain new knowledge and share that with the community.

I see 3 types of EVM (evaluation module) boards:

1. Must have, board can be used as if for most purposes.

2. Capability Testing boards, designed to be hooked up to test equipment for things like power draw, waveforms of just torture testing. Not usable as part of something else due to form factor (PCB with lots of blank space), price or design choices.

3. Complex projects, boards with many different related chips from the same company. These can be huge, many are design files only, some even need custom parts ordered to be built.

For #2, I just quickly design my own board, test points aren't as needed when you have a set of PCBite micro probes. Would like to actually test the parameters of a part with the manufacturer's EVM, I've just had a pressing need to.

#3 at the moment are problematic at the moment, they aren't usually made in large numbers, so can be hard to find, or the parts are just unavailable to make them.

One I'm really interested in, has become a major project just to see if it's possible to get the parts, it's not. The temptation is always there to modify the reference design slightly, and open a whole cupboard of canned worms.

I broke this into two comments because the first time I commented the whole comment disappeared. I needed to see if it could be saved.

One issue with reference designs is that although everyone uses them, not many members would have a use for a specific reference design at the same time. There is often a significant learning curve which implies a pretty big commitment to use the part in a design. I suspect it would take extra incentive to stir up widespread interest in participating in a heavy learning or building exercise around a specific reference design, but there may be a few examples where the reference design is useful or has wide appeal, that would work..

If members were incentivized to blog about reference designs, I am sure many members would be interested in reading about them.

Reference designs are often pivotal when learning how to design with a complex new component. It seems like there should be a way to leverage this type of resource to increase the forum's scope and member interest.