Dyne's 99-99 Build

[Dyne]

Introduction

In Treadwell's thread, I said I was currently trying to decide where to focus my energies for my builds in the runup to Dragon Con.  And I did decide: I intend to get LD-F1 completed or at least running before the beginning of September, which gives me a bit less than three months.  If I take the remainder of June to nail down most of the design, that leaves July and most of August to build, test, and paint it.  I have done a lot of design already in that build thread, but even so, there's a fair amount left to do.

However...

It's become clear over the last few years that my brain has developed this annoying habit of rebelling when I've decided to focus on something but I've not yet built enough momentum to be "in the zone".  It will suddenly start nagging me with a sense that it's urgent to be doing Other Thing Entirely (be it some other droid I want to build, or a wholly unrelated project) and it often won't stop until I give in or lose steam.

I can't afford to continuously fight the distraction for the next three months -- that will sap my motivation quickly.  So this time, I've decided to try a new tactic.  Rather than risk waffling between LD-F1 and an even bigger build like Treadwell or R2 (and ending up not getting much done on any of them), I am going to try leaning into the distraction, tricking myself by speed running -- for a little while -- a small secondary droid.  Specifically, the one that my brain keeps trying to convince me that I need to think about Right Now ™.


The droid in question is 99-99 (aka Four-Nines) from the Star Wars: Visions episode "The Ninth Jedi".  I've been pondering how to do this droid ever since I saw Visions back in September last year, but apparently the best time to start giving it a go was when I'd just decided that I might be able to get LD-F1 built if I can focus well enough.


Overview

This particular droid consists of a relatively simple ball head that rotates omnidirectionally at the top of a short cylindrical body with a hemisphere (not a roller, just a dome) on the bottom end.  It just hovers, and lacks legs or other features underneath.  



(The head is usually not that far from the body.)

There are four arms spaced at 90 degree intervals around the body, with a small rectangular panel between each pair of arms.  The arms deploy from the body on small tubular stalks, usually only as one diametrically opposed pair at a time.  

That's pretty much the entirety of the droid's appearance, aside from sundry panel lines and the details of the arms.  There are no lights or other protruding or moving parts shown, so if they exist, they are behind panels.

Now, I don't plan to try to make this droid hover like it does in the series**, nor am I going to try to make the head excitedly depart the body -- if anything, I'm hoping to prevent that from happening -- but I do hope to give it some life.


The Arms

From the exterior appearance, each arm consists of an upper "shoulder" piece, a single "arm" section in the middle, and a simple fingerless "hand" at the lower end.  One opposing pair of arms has white coloration in the middle segment with yellow on the shoulder and hand, while the other pair has all three sections painted yellow.

The white pair seems to be used as the droid's actual arms -- mostly for gesticulating -- and each one has a small black wheel-like feature of unknown purpose at the "palm" or its "hand".  My thinking is that this could be some sort of electromagnetic grabber or a "developable mechanism".  See this video by BYU for that idea:





The yellow arms are not seen deployed from the body in detailed closeups like the white ones are, so it's difficult to determine exactly how similar the design is (beyond their external appearance).  They appear to be normally employed only as flight assist devices: They were extended at least once when the droid was descending from a greater height than normal.  The droid also rotated its body in the same shot -- and thus these arms looked like a rotor.

Whether any of that was strictly necessary or even helpful is unknown (maybe 99-99 was just having fun).  The yellow arms are, however, later shown to contain micro thrusters in their "palms" (like a tiny Iron Man).  It's unclear whether the flames were produced from black "wheels" identical to those on the white pair of arms (if so, that would suggest that the white arms can produce thrust as well), or whether there's something else mounted on the yellow arms for this purpose.



As mentioned before, the white arms are attached by stalks that can angle up and down on the tilt (bank) axis as well as back and forth on the vertical (yaw) axis.  They can also rotate up and down on the pitch axis.  The arms themselves have a simple hinge between the "shoulder" section and the "arm", while the "wrist" joint appears to be capable of twisting or potentially even hinging (though it doesn't actually do either).  It is not clear if any of this is the same on the yellow arms or not.  They are not shown being able to move in any way other than angling directly out from the body.  They do appear to be jointed like the white arms, but they don't make use of this in the show.

The droid is capable of firing small blaster bolts from somewhere on the front of its head, akin to the Marksman-H training remote on the Falcon.  


Build Plans

Four-Nines is shown in the episode to be roughly the size of Lah Kara's boot (it is a bit behind her in the image below, as indicated by its shadow).  



Kara's exact size is unknown, but her height feels a bit exaggerated by the art style, so it could be argued that the droid might be a bit taller in proportion to a realistic human of her likely age (though that is also indeterminate).  In any event, I have elected to make it one foot tall, which seems pretty reasonable for a small droid companion.  This scale makes the droid's head a bit smaller than the Marksman (140 mm vs. 150 mm diameter).

99-99's head is a bit like the reverse of BB-8's setup: The head rotates instead of the body.  It needs to be capable of spinning to any orientation, which makes the droid's facial features a minor challenge.  I decided that printing a clear resin "lens" -- similar to what I did with Treadwell's lenses -- with the external spherical surface contour could work (as shown in the pic below).  This also lets the eye surrounds have some of their on-screen dimensionality.



I'm thinking three magnets in the top end of the body for gripping force, along with a set of three rollers ("ball transfer units", as opposed to "ball casters".  The latter tend to have an axle, while the former do not.)  I have a bunch of strong magnets lying around already, leftover from modifications to my Spinmaster Hero droid BB-8.

The head will need to be hollow, and needs to have something along its inner surface for the magnets to pull on.  My initial idea was to put a free-rolling magnetic coupler inside the head, similar to BB-8's "dome magnet mount".  It would work a bit like the hamster portion of a hamster drive, but not powered (just 1-3 magnets on some sort of caster arrangement, to let it roll in any direction as the ball rotates around it).  Or it could be a pyramid with rollers at the vertices and magnets supported on one face.

It then occurred to me that if I can get the gaps correct, it'd be possible to ditch the hamster idea entirely and simply attach ferromagnetic material (like steel washers) or a bunch of small rare earth magnets all over the inside of the head for the body magnet(s) to grab onto.  All the body magnets need to do is pull strongly enough to keep the head on and sufficiently tight against the wheel below for some traction, without pulling so hard that they interfere with the rotation or mar the surface/paint.

Currently, the thickness of the head is 5 mm, and it can't get much thinner if it also remains spherical inside -- this is mostly due to the space needed for the eye recesses.  The thickness of the dish-shaped portion of the body that the head rests in is also 5 mm, with the ball transfer units sunk into the surface enough to create a 1.2 mm gap between the head and the body, plus or minus a bit as the balls roll across panel lines and such.

However, the body magnets need not be the full 11.2 mm distance (5 + 1.2 + 5) from whatever they are grabbing; they can be mounted a lot closer to the surface of the dish (I could go as low as 1 mm from the surface, possibly 0 depending on how much plastic would be left to retain them).

Animation of the head will be provided by a small traction wheel at the very bottom of the dish, powered by one of those 3-6V DC motors with yellow gearboxes that you see in Arduino Robot Kits (TT motors) ... that's the T-shaped part shown in cross section in the below image.  The wheel and its motor would turn on the vertical axis by a servo, changing the wheel's direction as desired.  I plan to give this setup random speed, duration, pause duration, and direction values rather than planning to controlling the head manually or having programmed sequences, but either is certainly possible.

 

(By the way, the rectangular cross section on the right is the rough size of a Turnigy 2200 mAh 2S LIPO)

Due to the nature of the traction wheel setup, the wheel wouldn't be able to easily rotate the head on its yaw axis like an R2 dome ... the only way that could happen is if changing the wheel's direction applied enough friction and torque to drag the head along for the ride, which is conceivable but seems unlikely (and would be a bit counter-productive, as I want the head to have an entirely new spin axis between movements, not shift the existing axis to a new orientation).


Arm animation would add a lot of life to the droid, but it is a lot more challenging to implement, given the body's small size.  It might be more practical to make both pairs pose-able rather than animated, but if I can find a way to fit some mechanism in there, I will.

I figure that, per arm, three servos would be needed (one servo for raise/lower pitch + two servos for R2 holoprojector-like tilt/yaw motion).  Add a fourth if you also wanted to actuate the arm hinge.  You would also need some sort of servo mechanism to extend and retract each pair of arms. 3-4 servos per arm, times four arms, plus a servo for deploying each pair plus 1 for steering the traction wheel is 15 to 19 servos.

Fitting that in a cylinder 110mm diameter x 130 mm long would be a herculean feat, considering I'd also need to fit the recesses for the retracted arms, a battery, motor driver, audio board, speaker, hole for the mounting post, the switches, the microcontroller, and the wiring in there as well.  (There's a bit more space in the tapered lower section ... but then again, the upper "dish" and its hardware also takes away some of the space in the cylinder.)

On top of that, there isn't actually much room beneath the curvature of the "dish" to fit any bulky arm mechanisms.  At best, I'd probably have to have some sort of bent actuator to move things down more.

 


I came up with an idea for the joints where the arm stalks connect inside the body.  I obtained packs of M3 threaded servo rods and ball jointed tie rod ends ("Rose joints") for use with LD-F1, and I'll have extras.  If I mounted one of these standing up vertically, then put a short length of the M3 threaded rod through the hole in the ball with lock nuts on either side so that the rod is centered, I'd have the pivot point for a tilt/yaw lever.



In this setup, the rod and ball can also rotate in the housing. The arm can be attached on the outward end of the rod (with brass inserts or similar), and the servos can drive the end that's inside the body.


** Don't think for a second that I didn't immediately ponder how to accomplish the hover when I first saw the droid, though.  I debated whether you could do it with something like a thrust vectored EDF motor, similar to what James Bruton and Ivan Miranda built their rocket collab around.  But it would be super loud indoors and not leave much room for internal electronics.