The Widening Range of Cheap, Printable Materials

I've been using an XYZ Printing Da Vinci 2.0 Duo, which has had obstacles but overall held its value for over a year of regular usage. I'm thankful Dennis introduced me to the brand. The proprietary filament has been predictably limiting, but the build volume has been great. I also planned to use the Duo to help build a more specific printing setup once I've better understood what Touchdown Delivery needs long-term. This material reevaluation has been needed, and I'm looking forward to creating some illustrative images for V8 showing where the plastics below will be incorporated. 


From Most Attractive to Least Useful:

  1.  Polycarbonate Blends - Desirable because of its strength, and UV stable for long enough to fulfill the planned lifetime of the drone. The best choice for structural prints. PC Max Is the popular brand, but it’s relatively expensive, so I’d like to test a single roll first. 
  2. PET/PETG - This is such a common plastic, it’s recyclable in most cities as beverage containers and other packaging. It can also print clear with Glycol-modified additives (the G in PETG) and has an average glass transition temperature at 82 C (180 F). Although PLA has a 1.57% density advantage, the higher structurally operational  temperature range of PETG makes it the better choice for the flight prints.
  3. ASA - Slightly more flexible than ABS, just as inexpensive, and most importantly UV resistant, ASA is a choice material for several exterior prints. It's polystyrene based, very common in many appliances, and easy to transition from printed prototypes to injection molded products. It's great for grips and landing feet. ASA can be printed on most ABS-capable set ups. Color options aren’t great, but the longevity is comparatively more important for many exterior components, such as the servo mounting and handles. It also can be smoothed with acetone. 
  4. Flexible Options (TPU)  - These are specialized Urethanes. The key consideration in choosing which filament is whether the component in question requires flexibility or flexibility and elasticity. For small and mid sized components where softness is desired, (weather sealing, fabric backing, vibration isolating) printing in a low-A rating filament is reliable if the settings are right. I haven’t seen any ultra soft and opaque/clear options. For harder prints (higher A rating) where tendon-like or high usage properties are required, there are generally even more options, even with opaque colors. Some brands I’d like to try are Polyflex by Polymaker, Essentium’s 90A, Makeshaper 85A, Fiberlogy Fiberflex, and Diabase X60 60A
  5. Wood/Metallic/and Textured - These require some tuning for the individual print, and typically use PLA as the base plastic. Hatchbox, Proto Pasta Rustable Iron, Azure Film, are some common brands I’d like to try, particularly the H2O2 reactive rusting, Pine, cork, and bamboo filaments. I’d like to apply these materials to a few interior clasps. Multi-material printing using compatible plastics can allow a clasp to be printed with a structural aspect and a exterior shell aspect using something less structural, like cork. For the first flock of drones, some will have a rubberized neoprene foam interior and some will have a cotton-backed fabric interior. The clamping is the same regardless of the material. Having the capability to print in wooden and metallic textures gives a lot of design possibility for Partners to influence their customers’ delivery experience. 
  6. ABS - I love ABS for the color range the ubiquity provides and the resolution smoothing possibilities it allows with acetone. This can be by brush or vapor. It’s considerably less dense than PLA as well, at around 7.8% difference. Great for early prototyping, but shrinkage makes it so the final prototypes won't be ABS. 
  7. PLA - It has a low thermal threshold which is limiting for structural applications considering what landing on asphalt on a hot day will be like. However, the options for opacity and it is ubiquitous in the 3D printing population make it a great option for prototyping. It has an average glass transition temperature at 63 C (145 F). 

Far from being unworkable, integrating 3D printing into my workflow has been an adventure with a silver lining. Now simple parts are easy to create for prototyping or general purpose, like this wall wiring outlet, view here, or get the print file here


A product like a drone will benefit immensely in the near future from being largely designed for print. This is because the printing will become increasingly less labor intensive as well as higher resolution. In addition, it allows for entirely local production, which not only eliminates pre-production transport costs and some taxes, but is more ethical in that it prevents the need to contract with companies which might have associated labor that is unscrupulous. 

Additionally, by eventually investing in the needed machinery for rectification, Touchdown Delivery can use recovered support material and old parts to supplement on-site production from beads or welding-rod style stock. While I am still playing with the options for specific machines in my Production Budget, I continue to feel optimistic about printing being a major fabrication mode in the first flock of drones.   

Added Somed Core Features V7.3.1

  1. lots in the tablet rest for chargers to connect to devices when not in use.
  2. Bluetooth Pairing button and Dark Mode button.
  3. More fans for cooling. Air circulates in from the bottom back exterior and out the back top interior.
  4. included a CAD of the transformer, although the mounting structure needs a rework in the next version.
  5. Central buttons are easier to reach.
  6. wo Qi charging are now present, built into the rest. I plan on covering them with a neoprene layer.
  7. I added non-scrape padded feet using some hard black rubber. 

 

 

IMG_0557.PNG
IMG_0555.PNG
IMG_0558.PNG
IMG_0556.PNG

V7.3.0 Charging Station and Defined Connector Space

IMG_0462.PNG

I compacted the Charging Station and added an iphone 7 for scale in the tablet tray. I also defined the connector I am designing as fitting within a box defined by 0.75” x 2 x 2 (19.1 mm x 50.8 x 50.8). Once I have gone through an iterative design process I should be able to reduce this driving constraint as well, but for now I’d prefer to have the room and focus on well designed springed/clamping contacts for the DC power.  

IMG_0463.PNG

V7.2.1 Charger Redimensioned

I redimensioned the V7.2.1 charger to incorporate the V7.2 battery unit. I like the result, but feel I’ve learned all that I can from this version, and already have some simpler folds planned for the aluminum case. The V7.3 battery unit is similar to the 4 units shown here, but shorter with one less of the unit LiPo battery.  

IMG_0453.PNG

I’ve never liked using the force-sensitive touch screen on document printers and other appliances. The nicer appliances have capacitive touch screens with a digitizer like a smartphone, but even on these the lessened sensitivity is noticeable and hitting the edge of a button will often not register with the digitizer, a problem exasperated by a small LCD screen. I think its much better to invest in a clean design with LED indicators and minimal switches. If any interface is needed, including a Bluetooth unit and a microprocessor in the design allows for delegating the UI to the Toyuchdown Delivery app. This makes sense from both a power efficiency and user experience standpoint, because the user already has a high quality digitizer and LCD local to their smartphone. 

That being said, I am only going to find a home for the processing once I’ve dealt with the power conversion. Designing a custom power adapter doesn’t sound fun, but will ultimately be a necessity if I want to maximize the efficiency of the charge/discharge cycle. The more drone flights per day there are, the faster the drones get paid off, and the higher the monthly take home for the Partener. For the short term, however, I’m going to find several power adapters I like, install them, see what the sizing is like, and reassess. 

The V7.3.0 Charger will also incorporate the connectors I blogged about a few days ago. I have some designs for these, but would rather just include them later when I post the charger in its entirety. The connector I’m currently designing might stay around for a few versions, but I’d be foolish to think I can solidify a connector design in CAD alone, it will probably take a dozen prints to work out all the stress points. It’s time intensive and detail-driven work, but if the drone has a mvp, it would be the battery connection - a power interruption could cause the vehicle to crash.

This is one of the reasons I’ve designed with two battery units per drone. Like hands and eyes in animals, having duplicates of a critical feature is a survival advantage. One planned recovery mode of the drone involves noticing a one-side power enteruption and performing an unplanned safe descent. And while this would mean a ruined experience with a customer, if I design a robust power connector, a reasonably safe failure rate can be obtained.   

More Charging Design V7.2.1

It’s not finished, but once several rounds of design iteration have gone by it will be further refined. My strategy is to populate, assess, and compress. It is designed around the V7.2 battery parameters, which had 33 Ah of power per side. V7.3 and V8 will have 22 Ah per side. 

I’m once again glad I committed to the size reduction. Even with 36” in total length, this barely fits the slots needed for 8 batteries. I might find a new battery unit to act as the principle unit (I’ve been using combinations of 11 Ah LiPo batteries) so that the width of each slot can be reduced. 

IMG_0420.PNG

Additionally, I left out the connecting mechanics between the battery and the charging unit. This critical component is also Absent from the V7.2 drone design currently up on this site. I hope to devote some significant research to this topic before continuing with the Charge Unit design. The battery/drone/charger connection must handle roughly 1500 - 3000 insertions and disconnections in order to ensure the lifespan of the drone, so it will need to be a robust connection. In addition, it must also handle the already existing requirements of being vibration resistant and high-current. 

Positioning of this connection on the Charging Unit and Drone will be critical and need to be stress tested for thousands of cycles, likely with a pneumatic rig. Additionally, thoughtless placement of the connection in the Charging Unit risks voiding the IP22 rating if dripped water accumulates and bridges the contact.   

The battery swap on the drone is meant to be splash proof - the battery rig is compressed into a soft, silicone sealing strip. However, the sliding latch for this compression hasn’t left my sketchbook yet. These latches will be present in V8. This means the critical battery/drone/charger connection doesn’t need to be fully waterproof, as it is already protected from the elements inside the environmentally controlled electronics bay.  

Once I have designed this connection I can focus on how it is implemented in both the Drone (V8) and Charging Unit (V8). My current thoughts are two banana-style compressing copper inserts which contact inside two corresponding copper tubes. In addition to that, spring pins will contact a simple exposed pad for the cell balancing during charging. At minimum, this will create a starting point for me to begin the design of this small assembly. Updates in the weeks to come. 

Thoughts on Charging

Maximizing the utility of any electric vehicle means careful consideration for the charge and discharge the battery will experience under normal operation.

An electric weed wacker might see a dozen uses a month, whereas the drones I’m currently designing need to see roughly 150 cycles per month to generate the minimum profit necessary to justify their placement at that particular retailer or vendor. 

One way to mitigate this load is to introduce battery swapping, which could allow you to increase the battery:drone ratio to 2:1 and therefor halving the average usage. In V7.2, the one currently on this site, the battery units weigh 9.7 lbs (4.4 kg) each. This was my solution to the bulky battery problem; one 19.4 lbs battery is cumbersome, but two 9.7 lbs batteries are easy to balance with two hands, like buckets of water. The next drone version, V7.3 reduced the battery mass by 33%, and V8, which I’m outlining currently, will maintain this reduction. My hope is to design a battery unit anyone can easily transport and recharge at a station (CAD screenshots coming). This will allow me to stress test the battery and determine that critical ratio of batteries to drones I mentioned. 

The charging station has existed on paper for a long time, but now I’ve decided to make the first design, which I’ll Introduce alongside the V8 drone to create a more complete vision of how this design will interact in a retail and vendor environment. The needs of the recharging station that I’ve outlined so far are fairly basic and straightforward:

  1. Meant for stationary operation, but not impossible to move.
  2. One size, meant for keeping two drones in constant operation.
  3. Meant for indoor operation in a climate controlled environment, but designed for safe and reliable operation up to ~100 Degrees Fahrenheit (38 C).
  4. IP22 rated, a common minimum safety rating for indoor use, which provides regulatory assurance that fingers wont get pinched or zapped and that the device won’t become unsafe if overhead sprinklers drench it. 
  5. LED indicators (no gaudy LCD screens) for reasons of reliability, visibility, and simplicity. 
  6. USB interface for diagnostics.
  7. A standard internal current limiting circuit.  
  8. A phone/tablet rest.

So with two battery units per drone, a good number of charging slots per unit might be 8, which would allow 2 drones to be discharging 4 batteries (40 min flight time + 10 min loading/unloading) for a total of 50 minutes. 

While 50 minutes isn’t enough to get a decent charge on a large LiPo, 100 minutes can get you anywhere between an 80%-100% charge depending on the charger / battery combo. This is on justification for the extra staged slots. While 4 batteries are in the air, 8 more are charging in the desk unit. So when both drones complete their 50 minute circuit, the 4 batteries that have been on charge the longest are up to bat. The 4 expended batteries from the drone take their empty slot.  

This is another reason flight time is so important to the overall drone delivery experience; not only does it enable more miles traveled, but it also enables a series fo short deliveries in a metropolis to be carried out in quick succession without need of a battery swap.

If this sounds futuristically absurd, consider that a vendor in a city might only be 2.1 miles away from a customer. Currently, road-based delivery services inevitably waste time in traffic and at stoplights (even if made autonomous). A 2.1 mile trip might feature 15 major intersections, each with a ¼-⅓ probability of introducing a 30s-120s stop. A capable drone could complete several such trips in succession, all while consuming less energy from the grid than an electric car, and in fact eliminating that hypothetical car from the roadway to begin with. 

A drone properly designed around a deep cycle battery means the user isn’t required to swap batteries every time a landing and reload occurs. And when a battery swap is needed, even in the heat of rush hour, the Recharge Unit will be designed so that its maximum rate of comfortable recharge exceeds the maximum discharge rate of the drones at that particular network node. 

The Recharge Unit will be critical for a pleasant and efficient user experience. I have designs, but don’t want to devote time to an image capture until the V8 drone is ready to be introduced alongside the Recharge Unit. 

Reconsidering Payload Parameters

I’m excited about adding a Subscribe to Updates feature, which will enable weekly emails featuring progress.

I’ve been reconsidering the overall parameters of the payload, in terms of mass and specific dimensions. For the V7.2 drone shown on this site, the payload is 20” long by 14” wide and 8” tall (50.8 cm x 35.56 cm x 20.32 cm). It was also intended for up to 5 lbs (2.27 kg).

In hindsight this was largely arbitrary. I’m not sure why the FAA sets the maximum weight for drones to be 55 lbs (24.95 kg), but it seemed like an attractive design ambition to set 5 lbs as the starting target.

My new line of thinking revolves around what is most cost effective; at some point a drone is so small it loses its utility, or so large it becomes unmanageable for landing in a suburb urban or urban environment. So what payload capacity to design for should be governed by this consideration.

The example payload I’ve used in my math is 2 liters of fruit juice or soda and an additional ½ kg of goods. This is a fairly sizable margin, especially considering ordering 2 liters of soda by drone is an odd use of the technology. A more likely use scenario might be ordering 1 liter of fruit juice, maybe several other sizable items such as 3 apples and 2 peaches (½ kg), and a variety of other lightweight items (¼ kg), for a total of 1.75 kg or 3.86 lbs.

Designing for a smaller payload will further assure me the thrust margins will be sufficient for the drone to perform during the ascent with a maximum payload. It will also bring the cost of a single drone unit down.

This and other factors are being included in V8, which I intend to begin in the near future, but I’m still in the midst of the office set up, and I don’t want to fracture my attention too much.