Jack Crossfire's blog View Details
Archive for October, 2020
Posted by Jack Crossfire | Oct 30, 2020 @ 04:13 PM | 7,521 Views
Given the 3 days required to fabricate, it's never going to be mass produced. The wires should all be wrapping wire. They would have to be soldered on a separate jig, then glued in place.


The hall effect sensors were off center. The screw & the speaker need to change places. It needs a stand to balance on the charger. At least the frequency hopping works & uses a very simple algorithm. The hall effect sensors have to be the lowest mV/G possible.


Got the hall effect sensors as close as possible without removing material. Spacing out the magnets more just increased the deadband. The hall effect sensor fiducials are exactly where they need to be. They could be shifted to allow for hot glue residue. The magnets could be closer to the shaft, but that would require going to a metal shaft. There's still a bee's dick of hot glue residue offsetting the steering sensor. Hot glue residue could be a legitimate reason for adding clearance.

Only a driving test would show if the hall effect sensors are good enough. Getting within 0x08 might be good enough. Gluing the hall effect sensors has been a matter of dropping a blob on the plastic, then pressing in the sensor with the mark 1 lion paw. It cools the hot glue instantly but might be the most accurate way.
Posted by Jack Crossfire | Oct 28, 2020 @ 06:28 PM | 8,921 Views
Obviously a lot more than a few tweeks happened. The mane event was moving the speed buttons to the sides & the power button to the back. The button spacing was dictated by mechanical clearance rather than ergonomics.

This would be a lot easier with tact buttons, but the lion kingdom insisted on something more durable. These MHPS buttons could have gone down to 6mm per side without a major increase in cost, but the lion kingdom wanted to finish. Despite every effort, hot gluing all the hot glued bits is still a buster. What's needed is a fine pointed hot glue gun.

There's some debate about whether the pinky grip should go all the way across. The 2 halves don't line up, making it look awful.

Too bad there isn't any more investment in lion machine interfaces. They all standardized on the game pad, 20 years ago. There were a few voice, gesture startups, 10 years ago.
Posted by Jack Crossfire | Oct 26, 2020 @ 03:26 PM | 6,901 Views
The 1st prototype had no showstoppers, just a few tweeks. The speed control buttons would be easier to reach if they were side by side. This would allow the power button to go in back. The height transition was messier than hoped. Throttle needs to rotate 30deg north. Pinky hook should extend the full height. Speaker hole could be smaller. It's heading to a dogbone.
Posted by Jack Crossfire | Oct 20, 2020 @ 11:39 PM | 6,544 Views
Another well intentioned modeling job ended in a hot glue fiasco. Didn't want to order another battery eliminator & wait 6 weeks for China post, but didn't have any gold contacts either. The way to do it is to solder the contacts on a PC board & model around the PC board rather than try to model the complete wire routing in plastic or just rework an existing battery, as lions did for the 5D.

At least it worked & now every time the camera starts on the battery eliminator, it requires clicking through several screens of menus. At least this battery format was easier to replicate than the laptop battery connector in the T4I.
Posted by Jack Crossfire | Oct 18, 2020 @ 08:07 PM | 6,965 Views
This one was the 1st to have an overlapping clamshell everywhere without a hole. That should give it some minimum waterproofing. Given the joystick is never going to be water tight, the best waterproofing is to have every switch be a custom hall effect sensor mechanism & to put the
electronicals inside a gasket enclosure. It would increase the size. There was no problem with the ergonomic grips only being 1/2 as high as the controller.

Made a new gear with smaller teeth. That allowed a 4.5:1 gear ratio. The matterhacker filament was clearly darker than the novamaker filament.


Finished the test run with a lot of firmware crashes. The joystick was arguably too big. Some more font size tweeking is required. Hot glue doesn't stick to planar PLA. It sticks quite well to PLA laminations. Vertical stuctures have to be made on a planar surface for hot glue to stick to. The battery should be stacked on top of the inductor & the length reduced. The leads on the switch can be cut to reduce the length. The clamshell should have the inner wall on the electronics side to make it easier to hot glue.

This was close enough to be the last prototype for the camera panner. The next remote control is for vehicle control.
Posted by Jack Crossfire | Oct 17, 2020 @ 07:38 PM | 4,738 Views
Bed facing isogrids without support or infill are possible if you can handle sagging. There's enough sagging to get flat tops without infill. They print a lot faster than solid plastic. Helas, there's no way to subdivide the isogrid when facing the bed. Supports require too much cleanup. Modeled supports don't buy any advantage, because the entire surface has to be supported. It would take a dual extruder.

Prototype #5 for this controller had another batch of surprises. Tempting to make it another 5mm thicker in just the battery area to make it 32mm shorter. Thickness is only constrained by the switch. Saying Freecad has parametric models is generous. The spreadsheet program has no insert, remove, differentiation between strings & variables, no way to view the available parameters in the 3D view, a text editor that randomly exits.

Having a temporary screw retain the spring during assembly was gold. Lions 1st saw the technique in building the lunchbox. The MHPS2283 switch is also the best switch of its size. Lions 1st saw those inside a microphone. The momentary version of these switches is not so ideal, since they don't click like tact buttons or boundary switches. The best momentary button may be a hall effect sensor.

This one has modeled supports for countersinking screws in the isogrid. The countersinks are 2mm deep. An overlapping clamshell should make it slightly waterproof. Overlapping clamshells require parametric modeling.
Posted by Jack Crossfire | Oct 14, 2020 @ 01:22 AM | 3,570 Views
There aren't many gootube videos on the subject, but printing small parts requires designing the piece with 3D printing in mind.

For the standard lion settings of .2mm layer height & .4mm line width, z coordinates have to be a multiple of .2mm. Wall thicknesses have to be a multiple of .4. Cura doesn't enforce the multiples for some reason. It does enforce a minimum top & bottom thickness of .4mm.

A fiendish problem lions chased for weeks was ridges in horizontal surfaces. Horizontal pieces have to be at least 1.2mm thick to get a smooth top surface. That creates .4mm of top thickness, .4mm of bottom thickness, & .4mm of infill.


If the piece is 1mm thick, it needs to have .2mm of infill & will have a ridged but bearable top surface. If the piece is .8mm thick or is solid, it'll have no infill & unbearable ridges in the top surface. The printer needs a void between the top & bottom to push in excess filament. Ironing literally pushes filament into the infill.

For pieces over 1.2mm thick, the issue becomes how thick the top & bottom shell can be, since the only force is the top layer against itself instead of against the bottom layer. There is a certain limit beyond which it'll form ridges, probably .8mm.

Parts have to be printed so forces transfer along the filament lines rather than between layers. This results in funky orientations, but resolution in the Z direction is also a lot less than resolution in X & Y. Parts have to...Continue Reading
Posted by Jack Crossfire | Oct 13, 2020 @ 08:14 PM | 3,086 Views
Didn't think a lion would ever be able to replicate the crew dragon isogrid pattern. Didn't even realize it was just an isogrid until starting to model isogrids. Thought it came from aliens because it has inner subdivisions.
It actually feels just like a cheap toy. Most of the world's artistic CAD modeling is devoted to toys. While the isogrid does strengthen the PLA, it's manely artistic. The crew dragon is probably procedurally generated just like this, but with a much more sophisticated model that allows inserting extra doohickeys.
Posted by Jack Crossfire | Oct 08, 2020 @ 03:55 PM | 2,596 Views
The next remote control began life as a series of unmitigated disasters, but completely expected. It takes a lot more prototypes to get mechanics right than it took to get electronicals right.

A full prototype burns $1 of filament but takes 8 hours to print. If it printed 24 hours a day, it would burn $100 of filament per month or 5% of rent. Creepy how the lowest steering wheel layers look like a gas mask.

Things finally started turning the corner when lions abandoned automatic horizontal expansion & manually tweeked the CAD models. Automatic horizontal expansion yielded very erratic results. Shafts ended up too thin. Wide layers ended up too thick.


3D printed text is pretty awful from a .4mm nozzle, but changing nozzles requires changing all the expansion tweeks.
Posted by Jack Crossfire | Oct 05, 2020 @ 12:50 PM | 2,773 Views
The answer is no. You can't buy the ergonomic knobs & sticks in game controllers as standalone parts. They all have to be custom designed. The only standardized parts are hall effect sensors & potentiometers buried inside the controls, which are not ergonomical.

It took 4 prototypes to finally arrive at a spring loaded wheel suitable for steering. It just needs a few tweeks to reduce friction, tweeks to increase the spring tension, size optimization, & a more common spring. The lion kingdom believes with common pen springs, hall effect sensors, & only PLA, a highly water resistant, ergonomic remote control can be developed. A fully waterproof remote control is not affordable.
The only 2 waterproof controls a lion can afford are a steering wheel & throttle slider. The power switch, reverse, speed adjustments all have to be stock buttons.

Helas, it's clear from playing with this that it's not possible to jockey the steering wheel in back while simultaneously jockeying a trigger in front. Boosted boards use their trigger as a safety rather than a throttle. They always press the safety when jockeying the wheel.


Exposing the wheel on the rear is good enough for just a camera panner. The next goal is not a rear wheel but to standardize all the controllers on a side wheel where the steering buttons are.
Posted by Jack Crossfire | Oct 05, 2020 @ 12:30 PM | 2,726 Views
Direct drive panner (9 min 50 sec)





Quite a relief to finish 8 years of trying to solve this problem with a 1st design that gave pretty good results on the 1st try. It definitely would be better with proportional steering, especially for tracking objects. Timelapse mode was overrated. The ship would have to be stationary to give good results or timelapse needs a brushless gimbal.


The remote control buttons need to have the same steering directions as the driving buttons. Being a right or left pawed controller makes no difference in lion perception of what direction should be on top. It needs an enclosure to prevent cargo from interfering with the motor. Needing an enclosure, it might as well receive a gearbox. The remote needs the pinky grip. Holding it like a TV remote for 1 hour is stressful.


Smooth panning is actually really hard for a direct drive motor. The mane problem is cogging. Brushless gimbals have a true position from an encoder which allows them to lead or lag the motor phase, depending on the cogging. Most of them just rely on moving fast to hide the cogging. The lion kingdom's AI camera is smooth by being incredibly expensive & heavy.

Without some kind of encoder, your only option is driving the magnets harder. It takes a cooling fan, metal motor mount, high temperature farsteners, high capacity power supply. A geared stepper motor is what you need, but takes more space. A gearbox would allow it to be powered down...Continue Reading
Posted by Jack Crossfire | Oct 02, 2020 @ 09:27 PM | 2,733 Views
It was finally possible to abandon the stock gopro mounts & make something lighter, which didn't slip, & which farstened to carbon fiber.

TODO:
- round joints on the outside instead of square joints on the inside
- knobs for the wire bits