Documenting Design Part I: WheelAir

I'm going to start this post on a tangent: 

It really doesn't seem like it, but it's been nearly 4 years since we launched STEP 3D.

We started the company as a 3D printing hub or service bureau, but as the months started ticking past, we found more and more of our clients asking if we had the capability to conduct design and development work on their behalf. Well it just so happened that Dickon and I are both qualified product design engineers, and we were really keen to start pursuing this as it became apparent that there was an unfulfilled market need.

Fast forward to a few years later, when in October 2016 we had a brand refresh and updated the list of services we offer to include the range of product development services we'd until then been conducting on the side.

It was about the same time that good luck (or our exceptional Business Gateway advisor) put us in contact with a startup company called Staels Design

I met Corien Staels, founder of Staels Design, at her office in the one of the buildings at Glasgow University. Corien had spent the last year or so working towards solving a problem that was illustrated to her by her wheelchair-user tutor at University. This problem was one that never would have occurred to me as long as I had uninhibited use of my legs: many wheelchair users struggle with overheating as a result of sitting in their wheelchair for extended periods of time. 

By this point, there were already a number of prototypes created to explore different methods of cooling - from air cooling to water cooling, from blower fans to pressure vessels. However, the results were pretty unimpressive and none of them had led particularly near to an end product.

So the design brief was:

Develop a retro-fit wheelchair backrest cooling device that was effective, subtle, unobtrusive and easy to use

The first couple of months of the project we spent jumping from one frustration to another just trying to find a system that was effective enough to provide noticeable cooling. We tried all manner of combinations of axial and blower fans in different form factors and airflows, in combination with 3D printed ducting to direct the airflow into the target area. None of these worked well enough, mainly because we were trying to squeeze high flow, low pressure air down into a path of much smaller cross section, which resulted in a significant loss in kinetic energy. 

Some of the fan-based cooling embodiments we experimented with

Some of the fan-based cooling embodiments we experimented with

We then moved on to try perforated hoses inlaid to the cushion. We knew that this idea would require low flow, high pressure air, so we got hold of a number of small compressors. Even with the biggest compressor we dared test, the airflow wasn't detectable. We also weren't keen on the idea of mounting a vibrating compressor to the wheelchair. 

At this point we threw caution to the wind and started sitting on some thermoelectric coolers. Also known as Peltier modules, these devices are small plates that conduct heat from side A to side B when voltage is applied across them. We experimented with different intensities of cooling, varying the number and position of the coolers, and using flexible graphite or copper sheets as heat spreaders to try and increase the heat dissipation. 

One of the methods we experimented with was several Peltier modules sandwiched between thermally conductive copper sheets

One of the methods we experimented with was several Peltier modules sandwiched between thermally conductive copper sheets

The thermoelectric system actually managed to provide adequate cooling, but this came at the expense of being fairly power-hungry and created the new problem of heat sinking from the underside of the modules. So back to square one...

What I'm describing is really a drastic over-simplification of the process we went through. I'd like to be able to say we strictly followed a unidirectional design methodology from one stage to the next, but the reality is that one setback with one concept area sent us jumping across to another concept fairly haphazardly until we had a good understanding of the reasons for the 'underperformance' of each idea.  

We moved back to air cooling, using blower fans. We knew that there was only really one problem to solve with this concept:

Minimising obstructions to the air flow path. 

Our way around this problem was to create channels in the backrest with a similar cross section to the output aperture of the blower fan, and ensure that the channel wasn't squashable when the user put pressure against the backrest.

Staels Design had previously done some research to obtain pressure maps of patients leaning against a backrest. From these we could deduce the optimal areas for cooling, which just so happened to be two largely vertical areas either side of the spine. So we created two flexible channels to direct the air from the blower fans up across the backrest, then covered the cushion in a 3D spacer mesh fabric to encourage further dissipation of airflow.

Electronics Prototyping

The electronics system was designed by our electronics & embedded software developer partner, Positronics. Early iterations were quickly mocked up and tested with Arduino boards and breadboards before moving into custom PCB layout.

We ordered in the boards and components and did the board assembly in house. The board is mainly surface mount componentry,  so we needed to screen print the solder paste and place all the components before reflowing them.

One of the prototype boards we assembled and reflowed in-house

One of the prototype boards we assembled and reflowed in-house

Remote Design

When it came to considering the design of the user interface, we prioritised two objectives.

  1. Keep it as simple and easy to use & understand as possible.
  2. The remote must look and feel fantastic

In terms of control given to the user, it was as simple as an on/off function, the ability to select a cooling power level, and to display information about the battery level. On top of that, the system needed to convey some simple information to the user, namely whether the system is on or off, the cooling power level and information relating to the battery level and battery charging status.

We decided that the simplest method would be to provide two momentary buttons; one to cycle through different power states, and another to display the current battery level. In terms of the user interface, we opted for a series of 4 RGB LEDs, giving us the flexibility to use different colours to represent different information (eg. Blue for battery level, white for fan power level etc.).

The client was keen on using high quality materials for the remote body, as this would be where the user engages with the product. They brought us a couple of examples of products with similar aesthetics, in particular the PAX Vaporiser and the Belkin portable battery bank.

To achieve this look, we've designed the remote around a custom aluminium extrusion. The extrusion profile was developed and optimised through numerous 3D prints, testing fit and feel in the hand, as well as functional aspects like ease of button actuation and visibility of the LEDs. The extrusion is cut to size, then have several features machined in, followed by a light bead blast then hard anodising.

The full electronics system for the WheelAir is enclosed inside the remote in order to minimise the size of the air supply unit. The remote is connected to the air supply unit with a short cable, and a micro USB port on the bottom of the remote provides the battery charging.

Foam Design

We knew from the very beginning that the foam cushion needed to be made from high quality viscoelastic foam (memory foam). The other criteria it needed to fulfil included:

  • Provide adequate support for the air channels
  • Adjustable to different backrest heights
  • Comfortable
  • Minimal thickness

We began with a fairly typical cushion structure of low-density (soft) foam sheet laminated to a firmer medium-density memory foam sheet. The denser layer gives the cushion structure and support, and the softer layer provides a more comfortable, conforming cushion for the user's back. Once the channels were cut into the foam, we discovered we needed a third layer of foam to hold the the shape of the cushion.

To tackle the height adjustability problem, we opted for a fold-over cushion, with relief features to encourage folding over the top of the backrest.

A technical drawing of an early stage foam  cushion prototype 

A technical drawing of an early stage foam  cushion prototype 

Comfort was quite a difficult problem to solve - some users noticed that they could feel the firmer air channels pressing against their back. The contrast between the soft foam cushion and the firmer sidewalls of the channels were causing this, so we went to work iterating differences in geometry of the air channels, as well as experimenting with different shore hardness elastomers. 

We created a pouch to hold the cushion and the channels in place. This makes the foam cushion easy to remove and put through the washing machine without worrying about the separate components coming apart.

Cover Design

The cushion cover isn't just a functional component of the product - it provides most of the aesthetic features.

The primary design criteria of the cover were:

  • Dissipate the cooling airflow
  • Moisture-wicking
  • Sexy-looking
  • Weatherproof
  • Machine-washable

The spacer mesh was chosen from a surprisingly big selection. The chosen fabric is the highest quality spacer currently on the market as this was the one we felt provided the best results for the above criteria, combined with overall quality, feel and minimising the thickness of the mesh.

In terms of looks, we tried several textile patterns with panels of different meshes, fabrics, leatherettes etc., but we were unhappy with the outcomes. They felt cluttered and dated - a little bit like a 90s racing game chair. So back to simplistic design - we went for one large main panel of spacer mesh in black, with an artificial yet sustainable leather pocket at the back.

Corien & Euan at a design review meeting

Corien & Euan at a design review meeting

Revisions

As with most 'first-of-their-kind' products in development, we went through many stages of iterations, variants and revisions. While we fired through loads of these in the initial stages, we'd more or less settled on an overall concept and layout for the majority of the design stage.

We carried out two separate stages of user testing, and even after the second stage we still needed to make changes to the design.

The main issue that presented itself to the users was the ease of connecting and disconnecting the air supply unit to the air distribution channels in the cushion. Space in that area is tight, and the channel connectors were just a little bit too small to be easily connected. We also changed the design of the snap clips to be a lot friendlier on the fingers to release.

Assembling prototypes for user testing

Assembling prototypes for user testing

Then when we moved through to the DFMA stage (Design for Manufacture & Assembly), aspects of product assembly and the logistics of the production chain meant careful consideration of certain design decisions that had already been made. Thus far, we'd designed the product for low volume manufacturing, economising wherever possible. As the product gained more and more interest from the industry, the first production run grew from 200 to 500 to 1000 units, with yearly quantities expected to be at least 4000-5000 units.

Now we had the luxury of economies of scale on our side, we could start making adjustments to the design to maximise product quality, reduce unit cost and decrease assembly times.

We had originally designed the air supply unit enclosure to be a clamshell of two parts - both parts being identical with a rotational symmetry about them so that they would clip together. A number of compromises had to be made to make this possible, so when we found out that production quantities were significantly increasing, we re-designed the air supply unit enclosure to be two separate parts, allowing us to make the unit more robust, better looking and allow for improved maintainability. 

Double & triple checking the manufacturing drawings

Double & triple checking the manufacturing drawings

Thoughts on the final product

At this stage, we’re just about to ramp up to production. While there will no doubt be further challenges that crop up as we continue on this journey, this is as good a time as any to reflect on the final design, and to contemplate what we've learned.

It was a pleasure to work on a product that wasn't mainly an exercise in cost reduction. We had the scope to choose higher quality materials, more costly manufacturing processes etc. in order to maximise product quality.

Go into any mobility device distributor, or even just a medical device company in general, and it becomes quickly apparent that aesthetics and desirability of most of the offerings are less than what I would hope for. We're proud to say that WheelAir is the start of a new approach to mobility device design. Staels Design have already been coined "The Dyson of the mobility industry" by the judging team at Scottish Edge, which is obviously a huge compliment!

More importantly though, this is also one of the first projects we've worked on that we were able to see an immediate improvement in the quality of life of the users. 

Jay Anderson, a multi-disciplined athlete and wheelchair user, was one of the users to test the prototype units, and said

“This is the first time I’ve been in the park in the sun without overheating in 20 years. And not only is it adding so much comfort, it is actually attractive!”
WheelAir: It lives

WheelAir: It lives