Read all pages from top to bottom this is the order in which they occurred in the project.

---

Open Desk Chair Hack

I chose to hack an open desk chair to see if I could find ways to make it more convenient to disassemble or geared towards being made by anyone in a modular fashion.

Typically an open desk chair is fabricated using a CNC milling machine; however laser cutters are also widely available for use within fab labs and maker-spaces. Taking the cutting files I converted them to lines using illustrator, these could then be imported into Rhinoceros where they could be realigned to fit onto a standard sheet of plywood that would fit within a laser cutter (1 x 2 meters).

It was at this point I realized something significant about furniture in general, the supporting structures of a piece of furniture are often not the components that take up the most material and contribute the most weight. It is the surfaces that contribute largely to the inconvenience and weight of an item of furniture due to being large, flat and often solid pieces of material.

Due to being changed to a laser cutter and the restrictions of cutting power compared to the chair being designed for a 18mm thickness. The parts were cut out of 9mm and glued together with wood glue and clamped. Sanding was done afterwards to remove burnt edges.

Additional Components were made that attached along where the seat is normally secured to allow for string to be threaded through and pulled tight to provide a tensile surface to support the user instead. This reduced weight of the chair whilst still being load bearing. This was carried later into the project in the form of fabric surfaces.

---

Initial Ideas

Initial Ideas following the hack were done through rough sketching and card modelling to get an idea of how a joining system for furniture would function.

The initial ideas sketched and seemed most promising were turned into sketch models to examine their form and function in a basic way.

At this stage snap fit or clipping mechanisms seemed most promising as they would allow for assembly and disassembly of the product without tooling which would increase ease of use of the final concept.

---

3D Printed Clip Test

I wanted to investigate further whether clips were a viable solution for furniture so using solidworks I modelled a clip and had it printed to test:

• Tolerances between the female and male clipping parts
• Durability of the part after repeated clipping and un-clipping
• Force required to press the clip in and at if the clip would release with further pulling force.

This initial clip before printing was iterated on in solidworks to ensure that deformation looked correct for a 50N force, as an average pinching force of a person.

The final iteration before printing had increased the prong length by 10mm and also reduced the width of the prongs down from 20mm to 8mm.

Some considerations that were added were:

• Fillets on the interior corners of the prongs to reduce stress at the base of the flexion.
• Chamfers along all flat faces to ensure that printing maintained accuracy and prevented spread of the filament on the print bed that would produced inaccuracies in the tolerances.
• 0.5 mm tolerances on all sides of parts involved in flexion
• The ledge that the top of the prongs sat on was made to be 90 degrees to prevent the clip from pulling out
• Fillets on the catch at the top of the prong allowed for it to slide in and be pushed out a bit easier.

 

---

Clipping Mechanism Analysis

It was within 100 cycles of doing and undoing the snap fit that one of the prongs broke, although flexion and the force was acceptable I think due to the brittle properties of PLA that this method for securing load bearing furniture would not be suitable due to wear on the plastic over time and the general fragility of prongs long enough to preform effective flexion.

Even with an interior supporting structure that takes the weight. The prongs just being there to keep from a shift from position. I believe the prongs breaking would be a likely occurrence as well as adding complexity into prints due to overhangs and ensuring print accuracy for tolerances on such small parts.

The potential however for internal snap fits to secure larger components printed in multiple parts was something that was considered for complex assembled parts or utilizing space on the print bed.

---

Clamping Prototypes

A rough g-clamp was made using the laser cutter to create the clamp with a small section to insert an M5 hex nut.

Finger joints between the pieces were secured with wood glue and clamped together whilst drying.

A bolt could then be threaded through the slots in the laser cut pieces and utilize the thread from the hex nut to be screwed through and secure an object up against the wall on the opposing side.

In this prototype the head of the bolt was on the interior as it provided a larger surface area to secure the square dowel within.

From this the idea was developed into an attachment that slid over pvc tubing with the holes drilled into it to allow for the bolts to pass through. This would allow for the clamping force to be applied from two sides to secure the dowel in the tube; effectively doubling the securing force.

---

Clamping Analysis

Clamping using bolts was a very strong way of securing the dowel into position and would account for varying diameters and sizes of dowel to be used within the same joining method.

However it did leave groove marks within the wood. This would not be suitable as one of the requirements is that the methods used to construct the furniture were non destructive to the materials used.

A potential solution was the idea of having 3D printed feet that would attach to the threaded end once inserted onto the PVC tube.

This would allow for textured surfaces to increase grip as well as a larger surface area for the force to be applied to hold the dowel in place without damaging the material.

The issue with this is that to be able to easily assemble these parts the user needs to be able to reach inside of the PVC tubing and screw them on.

This is very fiddly and could not be solved without opening the side of the tube, which could allow for the dowel to slip out horizontally if being held in a vertical position.

The use of captive nuts to utilize their threads would allow for prints to be simplified as well as the threading not being worn down over time as would occur with printed threads. And this development was used later in the project.

---

Engineering Fits

Using tolerances between dowels and the connector pieces of the system was the most effective way of allowing for non-destructive fixings by securing the dowel due to the tension and frictional forces between the two pieces.

Complex mechanisms and securing methods would not be required and this also made sense for fast printing and inexpensive fabrication due to less parts.

The issue at this stage was creating a tolerance fit system but also accounting for deviation in diameter of the dowels which could affect how secure they were in the printed components.

It was decided that the type of tolerance zone required would be a push fit as opposed to a force fit. Whilst more secure, a force fit could potentially require tools to remove which might not be available and was not meeting with the specification.

Formlabs. 2020. Understanding 3D Printing Tolerances For Engineering Fits. [online] Available at: <https://formlabs.com/uk/blog/3D-printing-tolerances-for-engineering-fit/>.

---

Understanding Use Scenario

Considerations into the types of furniture and the use case were examined to understand what forces would be required of components made. These were then used to assess structural performance and usability of CAD iterations.

600N of force on each assembled part of the joining system was assessed as a minimum to ensure safety of any seated person.

---

CAD Iterations

Due to restrictions of COVID-19 ideas were explored and assessed through turning the initial card model ideas into CAD which could be assessed for suitability through comparing to ergonomic requirements (could they provide angled sections in all directions) and strength requirements through simulations.

This was one of the first ideas to form frame structures using attachments that tighten around the end of dowels and allow for them to be slotted onto a supporting core on the vertical dowels.

This idea had quite a high complexity and number of parts required, as well as it being a small piece of overhanging plastic that would not be about to withstand heavy loading.

If an attachment method was to be used the support between the connecting pieces needed to be of a more substantial thickness to withstand compressive forces.

---

Using a bayonet twist mechanism was another option to join dowels to a core section that would allow for angled connectors to be added. The bayonet could twist in two directions and was supported by the outer wall of the core as opposed to being hooked over like the slotted iteration.

---

3D printed connectors and cores with oversized threads with a large pitch. This would allow for the parts to be secured with only a few twists but the parts would be interlocked with a much larger wall  thickness comparative to the slotted iteration increasing the strength.

---

Fixing Selection

Using threads and screwing components together was not a suitable solution as there would only be one fixed position the component could be in to be fully secured.

Similarly using any of the bayonet solutions would allow for two directions for any angled pieces, twisting left or right to a locked position.

Whilst better this was not as suitable to a flexible system for furniture as wells as the potential for dowels to twist loose easily during use of the furniture.

---

The connector and node system that was eventually used as it allows for angled pieces to slide into one of four positions which is then secured with the screw. The dowels could then be slid into the piece to assemble the furnishing.

This idea was a combination of both the tolerance fit method as well as the captive screws investigated in the early prototypes before lock down.

It was selected due to the increased strength of the connection between the components, the use of captive nuts and bolts would prevent wear that could occur with 3d printed alternatives and due to the number of variations achievable:

4 directions for each angled connector multiplied by 6 potential directions = 24 for one connector multiplied by 6 potential connectors total on one core = 144 variations to build structures from.

---

Initial Connector Developments

One of the specification points for this project was to allow for the furniture to be dismantled easily by one person. Because of this there was a goal of minimal tools being used so that disassembly could be quick and done by hand.

The first iteration had a slot cut out of the connector piece; this was to allow for diameter of the hole to be smaller than the dowel. The idea being it would allow for dowels with slightly differing diameters through manufacturing error to still be fitted securely.

The printed part would bend out when the dowel was pushed in and hold the dowel under tension. The diameter of the part was 23mm whereas the dowel was 25mm, this would allow for 1mm of flexion on both sides of the part.

On examination though this caused issues primarily when the parts were put under load. Strain would be applied at the base of the splits on either side.

In a horizontal position this was likely to bend downwards under load from above and increase the gap between the part and the dowel. Leading to it being in-secure or break at the base of the split section where Von mise stresses where highest.

A similar issue would occur with vertical dowels if a sideways force was applied, This was most likely to occur in seating with the user leaning on a back rest for example.

Changes

Because of this the fixing for the dowels was changed so that the splits were not part of the connectors.

On consultation with a local 3D printing service the dowel hole in the connector was adjusted to be 25.2mm in diameter.

This would allow for the user to secure the dowel into the connector by the interference between the dowel and the plastic to be enough to prevent from unintentional loosing as well as preventing the issues caused by the flexion under horizontal loads.

Because of this a new requirement of users was to specifically use 25mm dowels and no others.  Whilst limiting in flexibility of material it does increase durability and safety of furniture for the user, an acceptable trade-off.

---

The intersection point for the nodes had a crease where the early CAD development was created by a boolean union of separate straight sections.

However this would be difficult to fabricate for a 3D printer as well as being a weak point when forces are applied.

Changes

Further iterations were created with a 5mm fillet between the straight sections of the part to distribute the stresses throughout the connected section rather than along a crease line.

At this stage flat edges were added to the parts for two reasons as well:

The bottom being flat would allow for better bed adhesion during printing and reduce errors near the bottom of the print.

Secondly it would allow for better stability of plastic components that sit on the floor when constructed in furnishing.

---

Standardized Dimensions

It was at this point that furniture was mocked up in CAD to see if the system would work for aligning different nodes and connectors to construct furniture.

The main issue found was that the overall dimensions of node variants differed so that when connected dowels were misaligned. This is better demonstrated in the illustrations below.

Between these two 3-way connector variants there is a 2cm difference between the center of where the 90 degree sections intersect, leading to a misalignment of two connectors of the same type.

For example a dowel coming out at a 100 degree connector attached to the left hand node would end up being 2cm lower than the 100 degree connector on the opposing side due to this dimensional difference.

Changes

To account for this there needed to be standardized dimensions across all variations to ensure that the linear dimensions between opposing nodes that secure a single dowel were the same:

• The fillet between two 90 degree sections would be 5mm rail to rail ( 2.5mm either side of the middle of the curve)
• The straight section after this would be 1cm in any direction
• The male joining section on the node with the captive hex nut would be 3cm in length
• The female joining section on the connectors would also be 3cm in length on all parts

---

Hard Surface Attachments

Hard Surface Attachments developed to utilize the vertical holes of the node to be secured whereas the connectors used the horizontal ones.

They hook over straight sections of the connector on either side and create a corner support to lay surfaces on without any permanent changes to the surface material.

Initial simulations at the desired 600N force showed that stresses occurred due to inconsistent wall thickness at the hook sections caused by the  90 degree join where the hook could run up to the edge of the rest of the component.

The horizontal section supporting the hard surface deformed, which would cause the surface to shift and would break at the edge between the vertical and horizontal bodies.

---

Surface Development

Initial Ribs patterns were made with diagonal sections to reduce the distance between supports as this was assumed to be a stronger formation.

However due to the shape of the surface this led to inconsistent sizes between the ribs and was overall inefficient use of material to strength.

• Nexpcb.com. n.d. Reinforcing Rib Structure Design For Plastic Parts. [online] Available at: <https://www.nexpcb.com/blog/reinforcing-rib-structure-design-for-plastic-parts>.

This specification for ribs on plastic parts was used to calculate the dimensions of the ribs for the surface attachment in a grid pattern :

A = 0.5 x 5mm (thickness) = 2.5mm Rib Thickness

B = Same as D in this instance

C = 4 x 5mm (thickness) = 20mm (distance between ribs)

D= 4mm under-hang – 1mm = 3mm (rib height)

 

There were still stresses occurring where the horizontal and vertical bodies met however. This was discovered to be occurring primarily on the underside where the ribs met the horizontal body at 90 degree.

To solve this .5mm fillets were added at the intersections of ribs with each other as well as the vertical body of the overhang.

After this simulations showed no displacement of the horizontal surface under a 600N load with PLA.

It should be noted that fillets could also be added to the top of the horizontal surface but this was decided against to allow for the hard surface placed on the attachments to be flush with the plastic components.

 

Final Prototype

When a local printing service resumed operation after lock down a coupler printed with a 30% infill ratio and 4 perimeters was made. (See Manufacturing Considerations)

Notes:

• The .5 tolerances for the join between the coupler and node were good.
• Some hex nuts fell out of the nodes during assembly to the connectors
• A way of removing secure hex nuts was highlighted by screwing in the bolt to the node without the connector piece and pulling. This allowed for separation of the standard components from the printed parts if disposal or recycling was required.
• tolerances for the bolts were good
• there was a slight shift side to side once the connector and node were secured.

 

Changes:

• Slight indentation would mean that horizontal shift could not occur once the parts were joined, as shown below.

• Tolerances for the hex nut slot should be reduced to 0 – 0.1mm.
• Fillets should be added to spread the load occurring on the male section between the connector and node components as this is where breakage is likely to occur.