3D printing or additive manufacturing has been a silent disruptor, mostly propagating under the surface and away from public view, generally viewed as an experimental science of limited real-life use.

 

Then Covid-19 struck. Suddenly additive manufacturers catapulted themselves into the spotlight by producing PPE and medical parts on-location all over the world, long before traditional manufacturers could even tool up. This new technology, it turns out, is not only useful in real life but could be very disruptive. Not because it displaces existing manufacturing methods but because it creates new possibilities that were previously unthinkable.

As with any new technology the early adopters are found in well-funded sectors: the military, motorsport, aerospace, medicine and high-end car manufacturers. Over the last 5 years we've seen 3d printing capability improve and costs decrease, making it accessible lower and lower down the value chain.

One area ripe for disruption is the automotive aftermarket. To understand why, we must examine the strong and weak points of additive manufacturing.

3D printing is slow (but really fast)

 

We've all seen how long it takes to 3d print even a small object. Large objects could take days to complete. That's because 3d printers add material one layer at a time, slowly "growing" a 3d object. The larger the object, the more material has to be cured or laid down, the longer it takes. The cost of time and material per item stays roughly the same irrespective of quantity. A linear quantity/cost graph in other words.

Casting or injection moulding on the other hand is quick. In the time it takes to 3d print one object, it could have been produced hundreds or thousands of times using injection moulding. Furthermore as quantity increases, per-item manufacturing costs decrease for injection moulding. The more parts you need the more cost effective it becomes.


There's a catch though. Injection moulding requires tooling which takes time and a fair bit of money to manufacture. Every injection moulded part requires it's own mould or "tool" and very specialised machinery to apply the tool. Also, every design iteration requires a new tool. The barrier to entry for injection moulding is therefore very high and only becomes viable when high quantities are needed.

Herein lies the strength of additive manufacturing. By the time the first part is injection moulded, which could be weeks after the design has been frozen, the item could have been 3d printed hundreds or thousands of times. It could also have undergone numerous tests and updates during this time. The 3d printer starts afresh every time and simply produces the latest version, turning agile manufacturing from concept to reality.​

Another interesting consideration is the fact that an injection moulding tool can easily cost more than a basic industrial 3d printer. The tool can produce only one part. The 3d printer can produce an infinite variety of parts.

You can 3d print anything (except most things)

One of the party tricks of additive manufacturing is the ability to produce impossible geometries with ease. Objects inside objects inside objects; complete assemblies printed in place; items with complex internal lattice structures. The fact that 3d printers can do the seemingly impossible gave rise to the notion that they can do everything. It's unfortunately not the case.​

3d printers do very well with some geometries, many of which are out of reach of traditional methods. However they struggle with many geometries which are really easy to make using traditional methods. The trick is to understand which manufacturing technology is most applicable to every situation. As 3d printers become more sophisticated and materials become more versatile, the balance is gradually shifting in the direction of 3d printing. We are however very far from the point where everything can be 3d printed and will likely never reach that point.

How does this apply to the automotive aftermarket?

 

It's a simple matter of supply and demand. Popular parts are always in stock because it's worth doing production runs to fill the shelves. However as vehicles get older and demand for certain parts drop below the threshold where production runs are viable, stock runs out. The sheer variety of parts that a manufacturer made through the decades also makes it impossible to keep stock of everything indefinitely. Something has to give and many parts simply become NLA (no longer available). At this point aftermarket manufacturers step in and fill the gap with reproduction parts of varying quality. Often the OEM will do a production run again if the item's value has risen to the point where it becomes viable. That explains why so many original parts for classic cars are eye wateringly expensive despite being fairly simple to manufacture.

Enter additive manufacturing where stock is virtual and any part is always a print away. The same 3d printer can produce a popular part today and an extremely rare one tomorrow and the cost of both are only determined by time and material, not rarity or sentiment. The fact that it takes relatively long to print a part is cancelled out by the fact that no tool change is required between parts and new designs can be printed as soon as they become available. What's more, the 3d printed part has in many cases better material and dimensional properties than the original which may be moulded using old and worn tools.

This adds an important new link to the supply chain. Firstly, many parts which have been NLA for decades can now be produced again economically and to a modern standard. Secondly, many newer parts which are still in stock may never require a traditional production run to be replenished. They may simply be 3d printed on demand once traditional stock runs out. This naturally only applies to a small number of parts because it has to be 3d printable to a very high standard and must comply with safety and other standards. However this is not fiction and Porsche are already doing it with select 959 parts which are simply too expensive to leave on shelves for an indefinite time.

 

Two things will drive this in future:

  1. Lower print cost and higher print quality. Today only a handful of parts may be viably 3d printed in a production environment but the list of viable parts is only going to grow.

  2. As additive manufacturing becomes more widely adopted more parts will be designed for additive from the outset, which means that they are future proof or "additive ready". 

To add another dimension to this, many parts will in future be printed on-demand and on-site at the dealership. No delays, shipping costs or import duties. No old or superseded versions. Some major manufacturers are already investing in this type of infrastructure.

From brittle plastic to durable carbon fibre

 

All of the above, albeit on a small scale, is already taking place in various niche markets. Let's look at a real aftermarket example: a classic Porsche part which has been NLA for years. There is a constant demand for it - not high enough for a traditional production run but definitely high enough for it to be digitised and 3d printed on demand.

 

This is what the example looks like:

It has experienced fire damage at some point and is badly cracked but of all the examples we could get hold of this one is the most complete and dimensionally accurate.

 

Sometimes a good example doesn't exist and then a virtual composite must be created by combining 3d scans from surviving bits. Of great value in those instances are technical drawings or old photographs.

 

In this particular case we have more than enough to work from.

The process starts with a high definition 3d scan. The scanner projects structured light onto the object and interprets the result to create a 3d mesh. It takes multiple readings from all angles until the complete object has been scanned. The result is an accurate virtual representation of the physical object. This is what the scanned object looks like:

The next step is to design the part in CAD from the scan and measured dimensions. It's at this point where inaccuracies in the original are often discovered. In this case non-uniform wall thicknesses and a slightly out-of-round hole. It is unlikely that Porsche produced a sub-standard part and we can only assume that this has been caused by warping over the past five decades. Deformed parts like these are often used to create casting moulds, thereby replicating the inaccuracy in aftermarket parts.

The technique we use takes longer but results in a perfect digital design which is accurately reproduced in the physical world. As additive manufacturing technology develops this part will be printed increasingly stronger, neater and closer to the CAD dimensions in future.

The original was injection moulded in plastic which became brittle over time. Ours is produced is from carbon fibre infused nylon which is more heat resistant than the original, much more durable and wear resistant.

This is what the original and our reproduced part looks like side by side:

This part, along with many others is now in our digital catalogue and printed on demand. No tooling required, no stock kept.

© 2020 by Martin Wiesner