Dyne's LD-F1 (D-O concept) build

[Dyne]

In honor of the first anniversary, here's a short clip from Dragon Con that I never got around to uploading until now.

[kresty]

I wonder how a power car antenna does it?

I believe they have a spool that works somewhat like a measuring tape (i.e. flexible enough to coil, rigid enough to extend the antenna without collapsing when unspooled).  

My friend suggested something like that, but I don't think I have anything that'd work as the tape (except maybe filament, and even then it'd probably require the spool to have a large diameter or it'd break unless it's nylon).  It probably also relies on the tube having a large enough ID for said tape to run all the way to the top, though, and this tubing has a pretty narrow hole, almost too small for the screw attaching it to the retainer.

I used to have one in an old car, and I think the spool might be more like a bowden tube. The nested bits of the antenna were solid.  So, I could imagine a cable being attached to the inside of the top.  Shoving the cable up makes it go up, pulling it down makes it go down.  When up it was "solid", like a normal antenna that you could manually push up & down.

[Dyne]

I've been tinkering with the top bay door, pondering how I want to set it up for servo control, and more specifically how to hinge the door.

Here's a simplified cross section of the bay, as if you are looking from the front of the droid toward the back (so port is to image right and and starboard is to image left).  Any dimensions specified here are accurate to the real model.  The black part is the door.


There are three types of hinges that I have been pondering:

1. A living hinge type of arrangement.  Think "piece of tape along the side of the bay and the underside of the door".  It is the most compact, cheapest, and easiest to implement, but it will look the worst.
   
2. A compact metal hinge, such as this antique type from amazon or something similar, arranged much the same way as the living hinge above.  I don't really think this would fit the aesthetics of the droid, even painted, and it would have to be glued on rather than attached with screws.
   
3. U-shaped hinges akin to those often used for R-series Astromech panels (astrogreeblies, Seth's hinges, etc)
   

If I wanted to place a hinge on the starboard side of the door, then the only thing that will work is the tape hinge.  The door sits directly on plastic with no gap between to allow for a compact metal hinge or for the mounts for the R-series hinges.  Also, the R-series hinges would extend into the interior cavity of the head, and on the starboard side, that'd likely be a problem because the existing head mechanics are right there.  And yeah, hinging the narrow side of the door would look a bit odd.

Putting the hinge on the port side of the door (which was my intent from the start) won't have quite as many limitations.  Obviously the tape hinge is still the simplest here.  But now a compact metal hinge could also work, but it would have to be shaved down on whichever side is attached to the bay wall (each side is (13/2) = 6.5 mm, and the wall is only 5.155 mm high.

As for the R-series hinge, this is problematic for a number of reasons, mostly because there are some tight constraints on where the pivot can be located.  If you drew an arc that represents the desired path of the lower corner of the door from its closed position to its maximum "open" position, then the center of that arc is the necessary pivot location.

The ideal pivot placement for LD-F1's situation would allow the door to rise faster than it travels horizontally (at least until it clears the top of the opening).  That means that the pivot would best be placed nearer the corner of the door vertically than it is horizontally.  If you imagine the pivot's rotation circle as a compass, it'd be best if the pivot were between East and East-by-Southeast from the door, or as a clock, between maybe 3:15 and 4:30 ... ideally closer to the 3:15 end.

Unfortunately, that's not very practical.  If you look at the dimensions given above, the pivot has to be at least (1 + 5.155 + 2) = 8.155 mm below the upper surface of the head, plus some offset from the inside surface (1 mm in this model).  That's dimension B in the next image.  Horizontally, on the other hand, anything more than 6 mm away from the port inside wall of the bay, will place the pivot axis inside the shorter section of the head.  That could be problematic.  It's 4.5 mm in this model, dimension A.

If the pivot fits the constraints given above (as A and B do), it will always be nearer the corner horizontally than it is vertically.  Have a look at this diagram:


In the above image, the dashed circle represents the path of the lower corner of the door.  You could only rotate the hinge about 6 degrees before it strikes the head, and by the time you've rotated more than about 24 degrees, the door would have reached the apex of its movement and would be moving further into the head (having never actually cleared it).

As for the dashed line coming off the bottom the pivot, then going left, and then going back up to the door, this is roughly the shape of the R-series style hinge arms that would be needed, and the lengths of those dashed lines are determined by where the pivot is.  For example, since the bottom dashed line rotates upward to become a vertical line as the door opens, its length must be >= dimension B plus some extra for clearance as well as the thickness of the arms perpendicular to it (represented as 0.5 mm in this example, but probably a fair bit larger).


There is another small problem that would make the pivots for the R2 style hinges harder to place because of the way I designed the head.  Below is the relevant section in CAD from a high angle to the side (the white shape shown in the head opening here is the ridged reinforcement pattern I'm likely adding to the bottom of the door itself. It is currently a separate model so I can check clearance.):


The main problem is the magnet mount (the protrusion beneath the top edge of the opening in the head, with the circular recesses for holding magnets) and how close it'll be to one of the hinge arms.  Here's the view from a lower angle:


And the bottom view:


I should have centered the magnet mount rather than putting it so far toward the rear of the head.  Notice how close the highlighted rectangle in the last image is to that horizontal rectangle. The pivot has to be closer to the top edge of the image than the highlighted rectangle is, which implies that the pivot on that side will end up directly adjacent to (and probably overlapping) that magnet mount).  It's also possible the new antenna mechanism could interfere, since it's in the area at its highest point, but i havent checked specifically.

So the upshot is that the R2-style hinges have several problems, and I would likely have to go with one of the other two options.

Edit: Upon further consideration, the other two also have a problem. If I leave the black door overhanging the port wall of the bay; then the door protrudes beyond the port side of the pivot point on that image, which means that the protrusion would rotate downward. Since it's already in contact with the lip the door sits on, it would be blocked from rotating at all. In order to avoid that, the pivot point ideally needs to be as close to the actual bottom right corner of the black door in that image as possible, and even that wouldn't strictly be enough. I'll have to think on it more. Might end up needing to embed a pivot directly into the wall of the bay.

[Dyne]

Before I get fully into the topic for today...

If anyone is seriously considering building this droid in the near term, let me know.  I'm not ready for the full file release yet, but I could gather up the STL files and instructions for the track links, track pads, and the little jig that lets you cut the link pins to the right length.  Those files are definitely not changing and though they aren't huge, they represent 213 of the ca. 321 printed parts needed for this build, so it would be a pretty decent head start.

Just be aware that for the rest of the droid, there will be several large parts -- like the rear body shell and rear chassis parts -- that you'll eventually need to figure out how to print. They won't be easy to split and currently won't fit on much smaller than a 320x320 build plate, even diagonally.  They are considerably longer than they are wide or tall, so a belt printer could also work.  (My printer is 400x400, and the only major parts that I've had to explicitly design to be printed in multiple pieces were those cosmetic track panels that I added in February.)  As for Z height, most parts are pretty short.  The starboard side of the head requires the most Z height, at around 224 mm, followed by the sprocket bodies and the body shells, which are printed on edge, as shown in Post #36 and Post #37.


Anyway, moving on.

I've been considering printing out a handful of small figures of LD-F1 to give away at Dragon Con (I know people sometimes give away small printed Mouse Droids or whatever to fans).  Last weekend I contemplated some of the details, and then started making variants of the main sub-assemblies (tracks, body, head, and neck) without all the screw holes and gaps in the walls and such.  

I specifically want his head to be pose-able (given that the head tilts are basically his signature move).  My friend suggested a small ring magnet-ball bearing pair, similar to what's used for some mecha miniatures, one in the head, one attached to the neck.  That'd allow tilt, nod, and rotation with no problem.

I'd prefer making these figures at 1/12th scale (similar to the scale of the Black Series figures and the Bandai droid models), but at that size I will need to beef up the neck substantially and probably print the parts on my resin printer.  To give you some idea of the range, if you notice the blue highlighted section of the neck in the image below, it has a diameter of 12 mm, so at scale that part would end up being only 1 mm in diameter.  The large disk is 50.8 mm (two inches) so it'd be slightly over 4mm.





I don't recall if I ever mentioned this, but since I was pondering the printable figure, I remembered that at some point after last Dragon Con I stumbled across what is (as far as I'm aware) the only other physical model of this Luke Fisher concept droid that exists.  It's a set of STL files by Colin/BigBangCollectibles on Etsy for printing a 1/6 scale static figure.

Colin interpreted the droid differently than I did -- instead of reading the taper as foreshortening like I have, he read it as a fairly exaggerated slant in a considerably shorter droid (lengthwise), resulting in the tracks being taller in the front than the rear .

As you may recall, I considered going that route myself very early on.  I commented on it way back in post #1.  Ultimately, the main reason I didn't was because the two panels on the side plates of the tracks seem like they are meant to be mirror images of one another.  If mirroring wasn't the intent, then I certainly wouldn't have designed them to look that way, and I doubt Luke Fisher would've either.

In a motorized droid, I'd worry that the short wheelbase would give it a tendency to faceplant unless you put a lot of counterweight in the rear, and I think the lift mechanism would be even less plausible due to the Etsy version's shortened wheelbase.  All that said, it's still a neat looking interpretation and as a static model it's fine.  It even has a more faithful version of the track pads.

[Dyne]

I know that I said I wanted to finish the release before making the major changes I have planned.  However, I have now entered crunch mode for Dragon Con, so if I'm going to do these changes, I need to do them soon.  I have already exported the various parts as they currently exist, I just need to finish going through and writing the instructions.  Unfortunately, I probably won't have time to do that before the con.  I don't think it's a big deal, because I don't think anyone is in a rush to build this, and releasing after these changes just means that a few parts will have two versions to choose from instead of one.

Here are the two major changes I presently want to make to the CAD:

1. Make the neck static
   
2. Improve the belt tensioning system
   

Since I'm actively working on the first one right now, I can talk about those modifications in more detail.  As a refresher, here are the parts we are concerned with:


In Black is the electronics cover.  There are two screw holes on each side that attach it to the rest of the droid.

In Green is the starboard side lower neck stalk (there is a mirror on the port side that isn't shown here).  It is split down the middle vertically, and on the end nearest the red part, the outer hub with the short diagonal slot is a separate cover held on by magnets.

In Red is the current upper neck stalk, which is split vertically as well, and has two embedded threaded rods for increased stiffness.  The rear rod extends into the head while the forward rod extends only just past the top of the neck.

I have three requirements for the changes to these parts:

1. The green part(s) will be truncated aft of the black part.  Instead of a functioning portion of the mechanism, they will be cosmetic parts held on with magnets.  (They need to be easily removable to provide access to the screws attaching the black part to the droid.)
   
2. The red part will be (to some extent) unified with the black part, rendering the neck static.
   
3. The threaded rod in the rear section must be preserved, because it serves as the pivot for the head's rotation.
   

Now, the issue with that second requirement is that it would normally make finding an acceptable print orientation challenging.  The black part is printed as shown, while the red part (being split down the middle) is printed with each half lying flat on the bed.  If these are merged, either the black part would be split down the middle as well (which would weaken it substantially) or else I would need a lot of support material to print the red part standing up.  Luckily one of my friends made a suggestion that will simplify printing: replace the central parts of the stalks with metal (or whatever).

You may recall that I called out the thinner portion of the rear stalk in my previous post as having a diameter of 12 mm.  That is near enough to the same outer diameter (13.05 mm) as the aluminum tubing that I have looked at for the head rotation of my Treadwell build and used as the mounting post for my 99-99 build.  The tubing's inner diameter (9.25 mm) is also large enough for the threaded rod (7.73 mm), which fulfills my third requirement (with some necessary attachment points at the top and bottom.  The thinnest sections of the front stalk (23.812 mm) are just a bit bigger than the outer diameter of 1/2" schedule 40 PVC pipe (21.4 mm), so that works too.  

I can adjust the rest of the neck greebles to match easily enough and the differences in size would be hard to see. With that done, then they can be printed in halves or upright and then epoxied around the inner pipe or tube.  This is similar to how James is doing his Treadwell neck.

Both the tubing and the pipe are easily stiff enough (at the short lengths necessary) to provide support for the head of this droid, assuming that the part that they slip into is strong enough to hold them.  (I have more concerns about whether the M3 screws holding the black part on will pull out the brass inserts that these screws thread into.)

Shown below is how the red and black parts have been merged with the black part.  I am pondering exactly where to make the splits to remove the greebles and leave space for the new cores.



The next image shows how I plan to truncate the green part(s).  As previously mentioned, the leftover parts will be held on with magnets, but I still need to add some supports underneath so they rest on the black part.


If you prefer to look at the physical droid's chassis, the truncation would happen near the small gap visible in the far wall of the droid's chassis, just to the right of the far motor.


As you may be able to tell here, it's not easy for someone to see far enough into this area to tell how far those arms go without getting pretty close, especially since the arms and the interior parts are all actually black.


As these changes get made, I will also be rewiring the droid to remove the existing fuse panel and make the battery easier to connect.


The picture of LD-F1's chassis above also has a clear view of the location where I am likely adding the new belt tensioner (if I work out the details in time).  

Specifically, just above the upper run of the secondary belt (the one not shown that runs between the silver intermediate pulley and the large drive sprocket on the left.   I'll go into this later if and when I get further along with it, but it basically involves a hidden rotating disc hidden inside the front chassis walls, attached to an idler on the outside pressing against the belt.  I'm also removing the slot in the intermediate pulley's mount while I'm at it, and the pulley's axle will be threaded into brass inserts instead of a just hidden (read: inaccessible) locknut that likes to loosen.  A better solution that I think I've mentioned before would be a hex head (not hex SOCKET head) M3x60 screw, but those are not common.


Edit: I have now adjusted for the new structure and split up the parts.  In the following images, the thicker white cylinder is the PVC pipe, and the thinner silver cylinder is the aluminum tubing.  The various colors represent which parts are printed as a unit (or as two halves, as convenient).  I may split up the red part further for easier printing.


The second image above is the cross section showing how the PVC and Aluminum tubes are situated internally, as well as the upper and lower holes for the threaded rod and the lower lock nut recess.  The upper end will be held on with the lock nut that sits above the rotation gear/neck.


As a side note: Probably to the great surprise of nobody that has read my build threads in the past, I have been unable to resist the urge to distract myself by starting another project for Dragon Con.  At the beginning of this month I began work on a mouse droid.  I really want to get the Mouse ready for Dragon Con because that's where most of my builds debut, but I have my doubts I can get much more than the chassis (from Michael Baddeley's printable chassis) up and running in time because I also have to do some coding to get my sound board and control type working for Penumbra.  The body shell is still TBD at this point.  The build thread for the Mouse project is on the Mouse club forum.

[Dyne]

I've completed printing of all of the new neck parts.   Raw black prints are annoying to photograph, but here they are laid out.


The Aluminum tubing and PVC pipe were cut to 250 mm length, which makes the neck the same height as it originally was. I have leftover threaded rod to use in the assembly, but I have to find my cutting wheel and figure out exactly how long it needs to be.

The two lower stalk arms (which will be held on with magnets, as previously mentioned) were printed in halves lying on the bed, as was the large disc that goes on the forward neck. These have already been assembled and epoxied together in the above photo.  The space between the marker lines on the PVC are roughly the distance that will be exposed and needs painting.

I'd printed the parts that go on the upper and lower ends of the neck first, and once they were done, I dry fitted them together with the tubes (lower part has a tight fit with the aluminum tube despite having same tolerance as upper part, so i increased that slightly for the remaining pieces.

To the right of the Aluminum tube in the above photo you can see three thin black cylinders.  The longest of those cylinders is the cover that brings the lower rear piston to its original length.  After my dry fit without the detail parts, I decided that I liked the look of the bare aluminum tube, so I made two shorter variants of the rear piston cover in order to expose more of the aluminum (it would be exposed only a centimeter or so at the original size). I decided to go with the medium length cover instead of the long or short one.

Here is the assembly being test fit together just before priming.  As you can see, I've used milliput to fill some of the gaps.  You may be able to see that I added some greebles to the top rear stalk that werent originally there.


And here is the assembly after priming with Rustoleum 2X Flat Black Paint + Primer.  The lower rear piston cover hasn't been glued on yet.  I've got a few rough bits to clean up, but most of it won't be easily visible unless you get really close.  After that, I'll give the parts a final coat of paint.


My rewiring plan requires involved disassembly of the front chassis section, but I dont think the new neck will add a huge amount to it. To remove the existing neck, there's the pivot bolt at the lower end of the lower stalk that I'll likely have to remove one or both tracks to get to.  There's also the pivot at the leading edge of the rear body shell which might be more of a pain, but hopefully not too bad.

Installing the new version is just a matter of bolting it on and adding the dummy lower arms. And that makes it removable for transport.

[Dyne]

Since the previous post, I've removed the old neck, and I didn't even need to take off the tracks thanks to fortuitous positioning of the bolts.  As predicted, the pivot on the upper shell was a bit more troublesome, but it wasn't as bad as I thought it might be.  Just a bit tedious.  Taking the old neck off let me test fit the new neck (to the extent possible, given the absolute state of LD-F1's electronics).  

With that done, with Dragon Con imminent, and with the replacement neck nearly ready for installation, I no longer had any excuse to continue to avoid it.  Yesterday, it was finally time for...

The Great Rewiring

I've been putting off the task of cleaning the electronics stuff up for the last 11 months.  This is the mess that I had to deal with:


No more.

The reason I'd been putting it off was, in order to get at the wiring, LD-F1 had to be about halfway disassembled, and I knew that would be annoying and would take me a few hours.  The tracks and track covers had to come off, of course, but also the outer carriers for the bogies (with the cover mounts attached), all four belts, the entire front axle assembly (sprockets, locknuts, bearings, etc.), and the front chassis.  The existing wiring also had be disconnected, ripped out, and redesigned.


As a side note, while I had the droid disassembled, I wanted to take the opportunity to replace that front outer wall that got chewed up last year.  It has somehow been working as-is, but I can foresee a situation where a belt damages it further and causes it to fail completely. So once the front chassis was off, I removed the old wall, transferred the pulley and drive motor to the replacement, and installed it. Here are the two parts.


I also made a modification to the outer carriers for the bogies, opening up a slot in the support for each intermediate pulley's M3 axle.  This is mostly just because I didn't like the axles being completely captive.  This will theoretically make the front chassis easier (though not exactly easy) to remove without disassembling the bogie carriers in the future, at least if you loosen the axles a bit first.



Taking the front main axle off requires chucking the threaded rod into a drill, as it is long and runs through two internal locknuts and a few small bits of plastic along the way.


This is the original main wiring in the front of the droid (with the fuse box removed), as seen from below.  Power comes in from the battery through the top XT60 connector (behind the horseshoe terminal with black insulation) and goes to the switch, then comes out through the white wire with the spade terminal to the fuse panel.  From the fuse panel, it goes to the two motor drivers via the other two XT60 connectors, and to the 5V step-down for the head via the red wire at the top of the image.  The ground wires from those components go back to the brass bus bar, which is the ground rail of the droid, then back to the battery.


With the fuse box removed, this area is much cleaner.  And here you can see the inserts that the new neck attaches to.


This is what the droid looks like with everything removed and the neck attached as a test.  It's a bit fiddly to get the starboard screws aligned to the inserts; they like to miss and push the part up instead.  I'll have to make sure there's no random bits of plastic pushing the neck out of alignment.  It's also awkward to put the forward screws in because the driver likes to hit the side of the neck.  Luckily I have ball-end drivers, so I can turn them at an angle.


All fuses have now been put in individual blade holders (and I added a new main fuse).  The main fuse on the 12V line lies in that electronics compartment beneath the neck, in between the switch and the nearer of the two Wago connectors in the rear of the droid.  Three more fused lines come off of that Wago connector.  The two next to the battery cover supply the Speed Controllers up near the front of the droid.  Routing back and forther adds a bit more length to the wiring but it's unavoidable.  The third fuse supplies the 12V to 5V buck converter mounted beneath the track (visible at right of image) by way of the rearmost Wago connector, mostly just as a convenience so I don't have to solder the fuses or a connector onto the wire.  The black cable here is the one that sends the 5V and ground lines from the buck converter up into the head.  The R/C receiver was also subsequently mounted on the inner wall on the left side of this image using double-sided tape.


The ground side of these three current paths run back to a third Wago connector in the front, which serves as the replacement for that brass grounding bus.  The ground Wago is then connected back to the battery negative terminal.  

The voltmeter was connected across one of the speed controller XT60 connectors and that grounding Wago connector.

The wiring was powered up via the benchtop supply to make sure all the components were working, and I got the reward of hearing the speed controllers use the motors to chime.  The voltmeter near the main switch and the one on the buck converter are both reading values that make sense.

Then came the laborious process of reinstalling everything.  I swapped out the spare intermediate pulley (the one I put on last year when I fixed the chainsaw incident) for one printed in PETG.  I also tightened up the pulleys on their axles

This is what the process for reinstalling the axle looks like.  You have to put it through one side using the drill (seen in the back), while looking through the hole on the opposite side to make sure it aligns with the hole correctly.  That's a bit annoying to do single-handedly.  

The other bits that are annoying to do by yourself are to take the belts off, put the belts back on, and to rejoin the ends of each track.

At any rate, the droid is now reassembled and all of the electronics are hidden within, which is a vast improvement.

[Dyne]

This is what LD-F1 looks like now:


I think changing to a static neck was an excellent decision.  The new neck feels pretty solid and stable so far, and I do like the addition of the aluminum tube.  The neck also no longer has a vertical seam down the center of the parts, and being able to just open the rear body to install the battery without having to hold the lower stalks up makes that process a lot easier.

Thanks to the neck being detachable from the body, reattaching the head is also a lot easier than it used to be ... aside from one screw that was difficult to correctly align with the brass insert.  Eventually I got it, though.


Post-reassembly testing went alright.  

He drives at least as well as he did before, possibly a little better.  I'm implementing a small amount of deadband on my radio sticks by setting up curves in the radio's model.  That will make it a little easier to (for example) drive straight rather than drifting to one side or another when the stick is slightly off axis, or look up/down without tilting the head at the same time (to the extent the servos behave themselves).

I mention behaving themselves because the head animatronics are working more or less like they did before. In other words, they still tend to jerk into place if you move the controls too fast, which I suspect is more an issue with the servos and the lack of proper smoothing than anything.  Also, when (for example) looking up/down, both servos are supposed to move together, but one servo will still sometimes respond noticeably later than the other. Unsure if that's a power draw issue or what.

I have, however, tightened up the slop in the rotation gearing by moving the servo's mounting point 1 mm closer to the neck on the rotation platform, and I also printed the new platform part solid, so the head shouldn't be able to skip teeth and lose centering anymore.  That said, I did have to add a bunch of subtrim to get it centered when the potentiometer is centered, so I think my gear is slightly offset in tooth position. I had it centered via servo tester but I needed to take the gear off tge servo horn to fix something during reassembly, so I may have put it back on rotated. I'll need to fix that before the end of the month.

The split loom on the servo wires up into the head causes the wire bundle to curve forward relative to the neck, which I'm not sure I like.  The loom also sometimes gets caught on the body shell without leaving enough slack to allow free head movement.  I might remove the split loom and go back to free wires, which is closer to the art anyway.

The ESP32 seems to no longer randomly reboot, but it still sometimes fails to power up in the first place (maybe the sketch is crashing at boot or something, not sure).

I ran into one further problem that I'll need to address.  Randomly, the droid started producing a buzzing noise.  The source turned out to be that servo that I left in the head for future use with the top door mechanism.  Somehow, the receiver triggered that channel into running continuously, likely running the servo into its end stops and/or running the servo horn into the inside of the head, and probably stripping the gearbox.  Either way, it spontaneously starts buzzing constantly and heating up so I've now disconnected it.

[Lichtbringer]

I'm not writing much these days....but i love watching. So cool, i love your work.

[Dyne]

I'm not writing much these days....but i love watching. So cool, i love your work.

Thank you, and no worries. I go through periods like that myself where I just lurk and admire.

Sent from my SM-A326U1 using Tapatalk