Charles’ Note: The assignment is completed using the current version of Chibicopter. The one that is built already has many electronic issues – the inability to perform wireless programming is the most critical, and a few traces were mistakenly left out of the Eagle schematic. The Reset button and battery plug are also in ridiculously awkward places. A new revision for Chibicopter is in the works (to be documented on my website separately) which will, sadly, ditch the curvy 3d printed shell in the interest of practical PCB layout and ease of assembly, and the design will move towards a wholly PCB based, one-piece structure. While it means that Chibicopter will join the ranks of the other small single PCB copters, it will make for a more elegant and durable product.
The primary production model for Chibicopter v1 is a partially-assembled kit:
- The board will be partially pre-assembled due to its all surface mount design and use of 0603/SOT23/TQFP parts. Going to a full kit board would limit its audience to only people who want to solder tiny SMT parts.
- The shell will be a pre-made component
- Motors, propellers, motor control modules, and IMU module are all separate parts.
- Battery is a me-fabbed piece that is ready to plug in. At qty 100, it is not economical to commission a custom “pack” with pre-soldered JST-PH connector. It would take me about 20 seconds (this has been timed three times for 3 cells) to solder the two connections and seal it in Kapton tape resulting in a similar quality product to a hobby-class 1S battery. They probably have a 20-something Chinese guy making theirs anyway.
Therefore, the user will have to solder the MCMs and IMU to the board (all thru-hole parts), push fit the propeller onto the motor (x4), then push-fit the motors into the frame (x4). Wiring of motors to the MCMs is also a user exercise, but the MCMs come with connectors for the purpose.
Example Production of 100 units (A nice small first-round number to start with)
Chibicopter’s full BOM: ChibiBOM
Being a very small and simple board at its core, it is reasonably easy to assemble. through MyroPCB.com, the PCB assembly quote was $396.00 with a lead time of 10 days.
Based on the estimate of my time spent soldering batteries, I estimate a 1 hour rosin-filled adventure making 100 battery packs. At the near-poverty-graduate-appointment pay level, the anticipated assembly cost is $12.00 for 100 cells. Seriously. That’s just wrong.
In the ideal streamlined production situation, I would be picking parts out of boxes or reels, putting them into little baggies, and the labeling the little baggies.
- The main frame and mechanics will fit into a 6 x 9 bag ($7.62 / 100ea on McMaster-Carr) with an internal 3 x 5 bag ($2.71 / 100ea).
- I would have to design and print a short printable label for each bag (“1 chibicopter”, perhaps) on some convenient mailing label paper ($12.79/pack) whose convenient overabundance may actually be used, in the end, for mailing labels. Amortized over packing 100 kits, the time of printing may be considered negligible.
- If I am able to fully pack 1 kit every minute, then the kit packing labor overhead is about $20.00 at the NPGA pay level.
An instruction and assembly guide will be put together online (much like Adafruit instructions). In fact, I like their “product label sticker” model, so my label would say something as simple as “1 Chibicopter, see howtoassembleandflyyourchibicopter.etotheipiplusone.net” or something with a less distressing URL. It would likely take several hours of picture-editing and writing to put together this website.
With the above overhead, the total cost rundown is:
- $16,332 for parts and services
- Approx. $50.00 for part-related overhead
- Approx. $32.00 labor overhead after production (Discounting design time, etc.!)
The rollup cost per unit with the above overhead is therefore almost exactly $164. For a tiny quadrotor kit produced in low volume including a massive digital fabrication cost (3d printing), that’s not too bad. The estimate does not include shipping and shipping products (it is assumed that shipping *cost* will be taken up by the buyer, but not, for example, boxes, tape, labels, etc.) nor difficult to characterize overhead such as all the running around I’d have to do.
I would be thrilled to sell the units for $199 – a nice round “shopping number”. At this margin (21%), it would take approximately 500 units for the enterprise to become self-sustainable (i.e. the next production run can be purchased fully using the revenue from the previous runs).
Time to Market
The “lead time” in the parts list is indeed very optimistic. It assumes that 100 (or multiples thereof) of whatever part I need is in stock and ready to ship. While this may be true for a high-volume vendor like Digi-Key, Hobbyking is an example of a vendor whose lead times are indeterminate if the item is out of stock. Some parts have been known to remain out of stock but still listed for 3 or 4 months. Because the parts in question which need to come from HK are generic R/C motors and ESCs potentially produced by several different Chinese manufacturers, it would perhaps be more effective to contact the manufacturer directly and place a bulk order. The amount this would save is not known at the moment.
In the ideal all-in-stock situation, the lead time would be dominated by the 3d printed frame. An order of 100 units could potentially be enough to fill all of Shapeways’ machines for a time, in which case the lead time could actually be very short. If not, then it will take at least 10 business days (their policy). The summary is that the kit could potentially be ready in as little as 2 to 3 weeks.
(The part cost is expected to fall dramatically once I rid myself of the 3d printed shell and add an option for R/C micro receivers – which typically retail at $10 single-unit, unlike XBees.)