An intro to Flex PCBs

Where a traditional electronic PCB is solid, flex PCBs are electronic circuits and components mounted on a (normally bright orange) ultra flexible film that can be folded with a lot of freedom and give. The current generation of wearables are dependent on this technology to fit all the electronics it needs into something that can be worn. For the last few years I’ve seen more and more manufacturers offer them as an option, and really they’re not much more difficult to design for.

As it looks like I have a project coming up that will use a flex PCB design, as a reminder to myself and as a guide to anyone who’s new to this, I thought I’d write a quick how too guide.

Uses

The most common use I’ve seen of flex PCBs are for connectors between devices. If anyone has used a Raspberry Pi Zero camera, then you’ve likely seen one.

Pi Zero and camera

This is simply a set of traces placed on the film that then connects two components together. By doing this you can create connectors for your custom components when an off the shelf flexible cable wouldn’t be suitable.

As with a normal PCB, electronic components can also be placed on them allowing you to create intricate ultra flexible designs. The example below is from a new experimental wearable health sensor that can be wrapped directly around someones arm to get readings. With a traditional PCB, this would be close to impossible to put together.

Flex PCB worn on a persons arm

If you have a wearble IoT device such as a smartwatch, you’ve likely already worn one.

Advantages / Disadvantages

The obvious advantage is that they’re more flexible, allowing you to create unusual designs and create products that would be otherwise impossible. They’re also thinner for when space is a factor.

And the obvious disadvantages is that they’re more expensive. They don’t benefit from the economies of scale that you get from a traditional PCB, which mixed with the more difficult manufacturing process increases the cost of creating your designs. On top of that their fragility makes them harder to assemble or repair if need be.

Considerations

When placing components on a flex PCB design, stiffiners are recommended to be placed parallel to where the components will be place. The picture on the left below shows stiffiners placed on the underside of a camera design. These reduce the flexibility in specific areas.

Although not strictly necessary, they make it much easier to place the components when manufacturing and reduce the likelihood of the contacts for the parts being damaged when the PCB flexes. In my experience there isn’t a standard for setting where these should be placed on a design, so when ordering I would speak to the manufacturer to ensure they know where to place them.

Also the more components placed on a design the less flexible it will become. This is especially pertinent for double sided designs which can lead to a lot of the advantages being lost.

Although this isn’t unique to flex PCBs, if you’re using one for a camera connector or a high speed sensor, a long connector can absorb noise from the rest of the circuit. There are a couple of techniques that can help with this issue when they come up, I want to write a blog post on this in the near future.

If you’re interested in a much more detailed guide that goes into details on the materials and manufacturing processes behind the tech, you can read one from Altium here.

Good luck with your design, I hope you consider making a product using one in the future. If you want some advice or help putting something together, feel free to let me know.

Featured image credit: Furniturewalla et al, doi: 10.1038/s41378-018-0019-0.

Optimising the 3D print for your prototype

For a lot of cases, it’s a good idea to put together a 3D print for your prototype as a quick way to validate how it looks. After putting together enough failed prints and wasted enough plastic with reattempts, I want to give some of the hard lessons I’ve learnt.

Follow the Y H T rule

Hq6yptx.png

When designing your print you can speed up the results and make your print look better by following the Y H T rule. When designing your print, you can reduce the amount of supports you need in your model by putting your overhangs in a Y shape. The smaller the angle the better, but 45 degrees is a good rule of thumb.

If that isn’t possible, making a H where the inside is hollow can minimise the amount of support needed depending on the material in use. Keeping the bridge between the two pillars to less then 30mm should be safe enough to bridge the connection and shouldn’t disrupt the print too much if you remove supports all together.

However a T shape will almost certainly require a lot of support or fail completely. If your design has this shape to it, you might want to consider the next rule…

Orient your model correctly

PrintOrientation.png

When your getting your print ready, think about the orientation it will have when you print it. Putting the majority of the printing surface on the bed is a quick and easy way to reduce the printing time. If you’re not sure, experiment with the printing angles and see the results from your cutting software.

Keep the quality low

At some point I want to put together a full blog post on this, but almost every time you’re better off with a fast, low quality print rather then a slower, higher quality one.

Like it or not, there’s normally something wrong with either your prototype or your design requiring a reprint. You’re better off learning that sooner rather then later. If nothing else, unless its an elaborate design there isn’t enough of a difference in how it looks to justify the effort.

When you’re finally sure of the design and that it all works as it should, then you can turn the quality up.

And finally, cheat!

A quick and easy way to find and fix issues is to have something else do it. By using MakePrintable you can upload your model and get free feedback showing any obvious issues.

Scaling up your Raspberry Pi prototype

Lets say you’ve been working on your Raspberry Pi based prototype for a while, and you’re starting to think about what you’d need to do to bring it to market. It can feel like a big jump, but you have options available to you if you want to turn your hobby project into an actual product.

The first thing that’s important to say is that I would not recommend using the Pi Zero for a commercial product.

Raspberry-Pi-Zero-Overhead-1-1748x1080

It’s tempting because of it’s small size, low cost, and very good value. The Pi Zero is an absolutely brilliant prototyping board for that reason and it’s a good idea to keep a couple around. The problem is that it is very difficult to get hold of them in scale, you’d struggle to get hold of more then 100 of them at a time due to their low production runs. On top of that they’re made at a loss, which means that there’s no economy of scale to a bulk order. From talking to suppliers that work with the technology, there’s also no guarantee that the design won’t change in the future.

So what about the Raspberry Pi B?

Raspberry-Pi-Model-B-overhead-1-1540x1080
Compared to the Pi Zero this design has the advantage of being available in much higher numbers and being a stable design. In particular Farnell offers the design in boxes of 150 in bulk at a reasonable discount. There are two big caveats to hold in mind:
1) Although it’s definitely available in bigger numbers then the Pi Zero, you’re still likely to struggle if you want to buy more then 1000.
2) It uses SD cards to handle its memory. These can be expensive compared to embedded flash memory, and has a higher risk of failure, especially if it is constantly being written and read to. If it powers off while being accessed, it can lead to memory corruption causing it to break down.
I don’t think there’s anything necessarily wrong with using this part in your design with these in mind, but don’t forget that it was designed for hobby coding and prototyping. These can very quickly become a showstopper for your business.

It’s worth also pointing out that there are now a few competitors to the space, such as the OrangePi that may be worth looking into in case you need a variant of the part that’s a different dimension or has a different spec.

61hSn-fER3L._SX466_

The part that the Raspberry Pi Foundation recommends for this however is the Compute Module.

Compute-Module-1-1920x1075

It’s essentially the bare minimum of what makes up a Raspberry Pi, enclosed in a small RAM chip sized board. It doesn’t have any external ports or embedded WiFi, but does come with an internal 4GB ram flash chip.
It’s available at scale, and has been tested to industrial conditions, and can better benefit from economies of scale compared to the alternatives. The downside is that it has no external ports or interfaces built into it, requiring you to design your own PCB that can work with it.
If you’re interested in designing with one, I would recommend that you purchase a development board and use that as your reference for the design and platform to prototype the code while you wait for it to be delivered.

How to get your invention in front of people

So you’ve completed your prototype, be it hardware or software, and now you need to do something with it. What do you do?

I was in this exact position about a year ago when I’d completed my first prototype of the Owlett. I’d built a device to help a family member, I knew it could help others, and I wasn’t really sure where to start to get the idea out there.

What I ended up doing is looking for other entrepreneurs trying to do something similar, searching via meetup and Google to see what’s out there. I was lucky enough to find the Health Foundry, a Waterloo based digital health tech collective that helped me to make some connections with experts and charities that helped me to get feedback.

Meetup.com can be a valuable source for groups, especially if you live near a big city. Even if you can’t find a group for your specific field, there’ll likely be more general entrepreneurial groups that you’ll be able to discuss this with.

And if you have a local makerspace, find out if they have an open day or an expo night that you could take the idea to. This will let you talk to experts who will be able to comment a bit more on the technology you’re using, and give you some advice if you need a better idea on where to go next with the engineering.

With that, hopefully you’ll be able to get the validation, connections, and feedback you’ll need to turn it into a real business.

MVP Hardware or Phone App?

A frequent question when I talk to people about the things I build is “Why didn’t you make it an app”, and the answer is always “It wouldn’t work as one”.

A lesson I had to learn from a previous hackathon is that if you can make it using a phone app, you should. Kotlin (for Android) and Swift (for iPhone) are both mature and well supported languages. Whatever features and functionality you want to use, it’ll be there with a ton of examples to help you quickly get setup. And then when you release it, you’ll have easy access to a massive store front and millions of users.

Even if the idea is that you’ll eventually make a hardware version, if you can make it as an app then I would recommend it to validate the concept. There’s no reason why you couldn’t put together a quick app to show to your potential customers, then work on the hardware and build in their feedback as you go.

And presuming you already have a phone, there’s no extra cost for you. Not that Arduinos and Raspberry Pis are expensive, but not only can it not compete with free it can’t compete with the number of features that are present in a smart phone which is perfect for rapid prototyping when you aren’t sure what new features you might need.

That said, there are three big reasons that I’ve come across which makes a phone app prototype unfeasible:

  1. Scale – Say you’re working on a set of simple IoT sensors that are designed to be placed in the corner of each room. You could make one using a phone app, but making a dozen phone prototypes can get costly, especially compared to a tiny Arduino Nano or a Pi Zero.
  2. Features – A phone can do much more then any prototyping board, but there a couple of things it wouldn’t be able to do. I’ve recently been working on a guitar tuner that requires a direct audio input from an electric bass. I would never have gotten that to work without some specialist hardware for the task.
    I’d like to throw ‘the form factor is wrong’ into this as well. If you want to make something that’s wrist mounted or works as a pair of headphones, then a phone wouldn’t be practical.
  3. The users can’t use a phone – This comes directly from the medical prototypes that I’ve worked on. If you’re marketing to the elderly, the blind, or another group that would struggle to use a mobile then coming up with something embedded that can be simple and made to purpose is the best option. The device I’m working on at the moment had to be made into an embedded device so it could be used by the blind, necessitating something with close to no interface required to work.

Hold that in mind, make sure you give it some thought, and don’t forget that whatever you choose there’s no wasted time when you build prototypes. You’ll always learn something.