Mar 03, 2016 Small Steps towards the Giant Leap!! Hello all, Apologies again for the lack of updates! We are working hard to sort out the little niggles that have to be solved prior to production. We are really learning the 90-90 rule: the first 90% of a project takes 90% of the allotted time, and the last 10% of the project takes the other 90% :-) So here is a list of developments for many of the various elements that go into making the Quadshot: Legalese Much as we would prefer to only have to deal with engineering, the unfortunate fact is that we had to spend time doing a lot of legal paperwork in order to get a project of this scale off the ground (so to speak). For example, in order to avoid losing a large chunk of your donations to taxes, we actually needed a corporate bank account, which required us to incorporate a new company (we are now Transition Robotics, Inc.). The paperwork associated with this took an inordinately long time, but as of today, we do finally have access to the funding which you guys generously provided us! Control Software We originally wanted to have a battery monitor beeper on the production QSs, but we were concerned it wouldn’t be audible when the QS was flying far and fast in horizontal mode. We have programmed instead a "wing waggle" into the QS where once the battery has reached a critical limit the wings will oscillate in the roll axis every few seconds, telling the pilot it is time to land soon. The oscillations are quick in order to be able to differentiate them from any wind induced vehicle motion, but small enough that there is no loss of control. We tested this in all three modes. Chris’ comment after the test flight was, “I had not witnessed it before but was told what to expect and I noticed it easily without losing any sort of control.” Mission accomplished! This feature will be incorporated in your QSs. Electronics Piotr has been revising our electronics and getting them ready for production. He made a a new low profile Aspirin IMU carrier board for Lisa/L, allowing for an easy fit into the avionics box on the QS. The production run for this board is already complete! This is shown below both from a top view and a profile view: Furthermore, Piotr went through two revisions of the Lisa/M design, solving a large number of issues and adding improvements based on feedback we have received from the Paparazzi community as well as lessons we have learnt in our development and testing process: Removed pull-ups on the USB GPIO. Removed pull-ups on the CAN GPIO. Connected USB_VBUS to PA9 (needed by the USB OTG). Removed USB pull-up transistor as USB OTG has a built-in pull-up. Swapped UART1 with UART3 (UART1 was used for GPS and PA9 was its Tx line, to be able to talk to the GPS unit UART3 is a better choice, as UART1 only has an Rx line now it is a better choice for Spektrum RX modules). Removed USART3 Tx GPIO from the GPIO connector and moved to the GPS connector. Added voltage selector jumpers to the RC RX connector; to enable powering of 3v3 or an 5v receivers. Replaced vertical board solution with through hole servo pin headers (easier assembly). Servo connectors are in groups of two; for easier assembly. Servo VBUS is connected together on all four layers; for lower resistance. Moved LED's from under the analog2 connector; to be able to populate LED's and the connector. Moved the RC Rx connector a bit; to prevent crashing with the JTAG plug. Added one additional servo connector; now we have all 8 accessible through the standard servo connectors. Added secondary through hole Picoblade USB connector for easier routing of USB inside an airframe. Lots of silkscreen improvements. Improved routing to allow tear-dropping improving manufacturability. Fixed STM32 F1, F2 and F4 compatibility circuit. Connected existing UART RX pull-ups to the respective connector power pins instead of 3v3. To prevent connecting 5V over IO pin to the 3v3 power rail. Added pull-ups on all UART RX lines to prevent undesired floatation. LED's are connected to 3v3 now. To make sure we don't have an issue with voltage tolerance on the GPIO pins. These new Lisa/M PCBs have already been ordered from our suppliers and are scheduled to be shipped to us on the 5th of December. A prototype of the Lisa/M that will ship with the Espressos is pictured below: The new Lia PCB for the Mochas is also through its design cycle. Since the Lia is based on the Lisa/M, it was critical to iron out any bugs and make any improvements necessary for the Lisa/M prior to commencing the Lia design process. The PCB order will commence production a few days after the OEM Lisa/M PCBs have been received and tested by us. A screen shot of the finished Lia PCB layout is included below: Mechanical Design As you may imagine, designing parts for injection molding requires very careful attention to detail. For starters, every vertical surface in a part must be drafted, i.e., at a small angle, instead of being truly vertical. This ensures that parts can be removed from the two mold halves without getting scored, warped, or broken. A complex shape like our avionics box therefore required a lot of work to make it compliant. Just as an example, here is a list of the iterations we had to go through with the plastic molding company prior to converging on workable designs for the box, lids, and pylon ‘skeletons’: We have also been trying to make the Quadshot easier to pack and transport as well as even more robust to damage on crash landing. To this end, we have created pylons that can be snapped onto the vehicle, allowing it to pack flat during transport. In addition, we hope (and we will continue testing this) that hard landings will cause the pylons to separate, thus absorbing some energy, rather than damaging the entire structure. This necessitated moving to a square spar and changes to the wing and avionics box spar channels. However, we do believe that the effect is well worth the effort. Prototyping & Flight Testing Apart from nailing down potential issues in assembly, we have been making our wiring system more robust and resistant to failure. We have also prototyped and flight tested the removable pylons mentioned above. Here is Chris demonstrating how these work, as well as a test flight segment to ensure that the removable pylons do not detach in flight: The last couple of days have also brought high winds to our Bonny Doon area, resulting in loss of electricity, phones, fax and for some really annoying stretches, the internet; but one huge benefit has been ….. wait for it ….. high wind flight testing! We carried out several flights in modes 1 and 2 with vertical take-offs and landings in average winds as high as 12 mph gusting to 30 mph! The Quadshot remained controllable and finished the flight test series without damage. Since winds at altitude were higher than the stall speed, we even had situations where the QS had negative ground speed while maintaining forward flight! Manufacturing We are pleased to announce that we are going with a US-based company for the injection-molded plastic parts. We hope to pull the trigger on production in the next week, and ‘first shot’ initial parts will take about 2 weeks. Jeff spent a lot of time looking for EPP molding companies within the US for our foam parts, but most are focused on the automotive and packaging industries and not set up to do small (both in size and quantity) production runs. They also usually have longer lead times and much higher costs than offshore manufacturers, so we have decided to use an offshore manufacturer at this time (although we hope to eventually move to a US-based molding company). We found an experienced manufacturer in China that has just opened a new factory with adequate floor space and new machines. As far as we can tell, the labor conditions are also good. We hope to kick off foam production soon as well - once we do it will take them about 40 days to make the tooling and “first shot” parts. As mentioned before, Lisa/M PCBs are due to be shipped to us on the 5th. Lia PCB production will begin as soon as we have tested one of the Lisa/Ms from the new batch. A US manufacturer has been chosen to assemble these boards. Typical assembly times are expected to range from 2-3 weeks after PCBs and components have arrived. Other developments In the spirit of sharing our work here we also decided to submit a paper describing the work we have done in getting the Quadshot design up and running to the 2012 IEEE Aerospace Conference (Mar 3-10, 2012) (http://www.aeroconf.org/). This paper was accepted by the conference and once it has been presented and published, we will make it available to you since we feel that you might enjoy reading about the thought and effort that has gone into this project. As an example of the things we did during the design process, here are some streamline views, generated during the aerodynamic analysis, showing the effect of the winglets on the airframe: Quadshot without winglets Quadshot with winglets Decafs Finally, the Decafs shipped on the 7th of November. If any of you ordered a decaf and have not yet received it, please let us know by messaging us on Kickstarter. Please include your address, qty ordered and size in your message. That's all for now folks .... Thanks again for your continued support and patience. We will keep you updated on further progress and delivery schedules as and when we know more. We hope you had a great Thanksgiving weekend! Do check out the new website and upcoming shop at thequadshot.com and visit our forum for even more regular updates and faqs at http://forum.thequadshot.com/. Best, -tQt: Piotr, Chris, Pranay, and Jeff.
