Monthly Archives: September 2014

ClearNav Joystick Part 6. Flap Grip Model Finished?

In Part 5 of this journey, I described our paradigm shift from a joystick mounted ClearNav remote caddy to a flap grip mounted version. When John received and tested the first try at this approach, he agreed we were on the right track, but there were some ‘issues’ (there are *always* issues!).

The cable channel in the flap grip piece is unusable, as routing the cable this way interferes with opening the canopy. Instead, John just used the single tie point on the caddy piece, and simply detached that piece when entering/exiting the glider
The rectangular slot in the flap grip piece was a bit too roomy in the forward/aft direction.
The finger scallops on the front of the grip didn’t really fit John’s fingers – oops!
The caddy piece would work better if it was offset slightly forward of the flap handle centerline, and rotated about 20 degrees ‘up’ relative to the top surface of the flap grip (i.e. toward the glider centerline)
The way that the caddy piece was attached to the flap grip prevented it from being rotated slightly to accommodate John’s actual thumb position; he suggested a round stud rather than a rectangular one.
The remote connector (a RJ-11 phone-style connector) sticks out the front of the caddy and is vulnerable to damage; John suggested I design in some protective walls for this.

Incorporating all the above elements into a new design involved some head-scratching, and quite a few 4-letter words, as I found that none of the 3D design tools I have so far employed (TinkerCad, 123D Design, and MeshMixer) would do the whole job. I wound up working with all of them at one point or another in order to get what I wanted. In addition, I was in the process of upgrading my PrintrBot 3D printer with a heated bed to accommodate future plans to print with ABS plastic in addition to PLA, and this turned out to be a lot more complicated than I had envisaged (I eventually reverted to my original hardware configuration, as I could not get consistent prints with the heated bed).

The first three items above were trivial to solve – just deleting the relevant ‘hole’ structures from the design, and modifying the rectangular slot dimension slightly.

The last three, however, required a complete re-thinking of the way the caddy piece coupled to the top of the flap grip piece. The flap grip has an elliptical cross-section, with a top surface that is parallel to the glider centerline, while the caddy piece has a rectangular cross-section that needed to be tilted up 20 degrees to the flap grip piece. In addition, the front portion of the caddy piece needed to have significant material added to support a side wall protection structure for the RJ-11 connector. I ran through several TinkerCad versions but ultimately realized that TinkerCad just kind ran out of horsepower to handle the complex surface morphing I was looking for. I just couldn’t get any kind of a smooth transition from the elliptical flap grip shape to the essentially rectangular caddy shape *and* provide for connector walls.

An early concept for the ‘new improved’ flap grip model

However, I had been playing with 123D Design as a possible alternative to TinkerCad, and realized that it’s ‘sketch lofting’ feature might be just what I needed (assuming I could get past all the 123D Design frustrations and actually make the *%$@!#%$#* thing work! The way the ‘lofting’ feature works is to take a ‘base’ 2D sketch in an X-Y plane (in my case, an ellipse representing the top of the flap grip), and morph it into a ‘top’ 2D sketch (an outline representing the bottom of the caddy section) over the Z-axis distance between the two sketches. By doing this I hoped to get a much smoother transition from one surface to the other. After playing around with this in 123D design for a while, I actually got this feature to work (an unusual occurrence with this program!). The following two images show the basics of the feature – here I have ‘lofted’ an ellipse in the Z=0 plane to a rectangle in the Z=50 plane, rotated about the Y-axis by 20 degrees.

2 different 2D surfaces

An ellipse ‘lofted’ into a tilted rectangle

For my purposes, I imported the flap grip and caddy pieces into 123D Design, and used them as patterns for the required 2D sketches. I still needed to add the connector-guard section onto the front of the caddy section sketch, but this was fairly straightforward using the 123D Design ‘2D polyline’ feature. Once I had the two surfaces mapped out and placed at the desired heights/angles, I could use the ‘loft’ feature to create a transitional 3D structure. The first run at this is shown in the following screenshot.

First try at blending the flap grip surface to the caddy surface.

This actually worked pretty well, but I realized that the connector-guard portion of the transition structure wasn’t going to be nearly deep enough (in the Z-axis direction) to actually do any connector guarding – bummer! So I tried again using a 3-surface model, with a middle surface using the non-rotated caddy outline plus the outline of the connector guard. With this setup, I got a much more radical initial transformation from the ellipse to the caddy outline (not so good), but much more available bulk in the Z-axis direction for the connector guard (very good). On balance, the need for the additional connector guard material outweighed the aesthetics of the smoother transition, so I wound up with a transition section that looked a bit like some prehistoric alligator (minus the legs) but…

More or less final transition section from the flap grip ellipse to the caddy

Caddy, transition section, and cylindrical mating stud

The next step was to import the transition section created in 123D Design back into TinkerCad for additional work and tuneup prior to 3D printing. It was right in here somewhere that TinkerCad (a ‘cloud-based’ application) mysteriously went away for about 24 hours, causing a major hiccup in my project and correspondingly major damage to my Wa (I had upwards of 80 design revisions in the TinkerCad cloud, and if they all went away…). Fortunately for me and my Wa, TinkerCad came back several hours later, and I was back in business. The following screenshots show the ‘final’ (as if this project will ever end!) version of the remote caddy section. The holes in the remote caddy bed were intended to accommodate the remote’s mounting screws (so they could be retained in the remote and not lost), but unfortunately the measurements were a bit off :-(. The side view shows the tie-wrap attachment ring on the front undersurface, and the mounting plug on the bottom. Both the mounting plug and the front portion of the caddy underside were shaved off a bit to create a larger attachment area for 3D printing

Final caddy version side view. Mating plug and front bottom portion shaved off for better 3D printing

Final remote caddy version. Note holes for remote mounting screws (didn’t work)

So now it was time to fire up my 3D printer and start making parts. Unfortunately, in the interim between this version and the last one I had decided to do the long-awaited upgrade to a heated print bed so I could eventually transition to stronger and more resilient ABS plastic instead of only PLA. The downside to a heated print bed is the much higher power requirements, and a host of other secondary problems associated with heated-bed printing. As it turned out, I couldn’t get a consistent print of the caddy piece with ABS or PLA using the heated bed, so I eventually had to downgrade my hardware setup and revert back to the original unheated bed, at least for these prints. The following screenshot shows the problem with PLA on the heated bed. The heat from the bed causes the PLA filament to stay soft too long and curl upward, even with forced air cooling.

An aborted attempt at printing the remote caddy piece. Note the curled up edges and numerous print failures

Anyway, after getting the hardware restored to its original unheated configuration, I started making trial prints of the remote caddy. The first several tries were failures for one reason or another. The first few trials failed because I had inadvertently disabled the ‘generate support structures’ option, and the next few failed because the support structures, when enabled, were too tightly bound to the main structure to remove when the print was finished! It took some tweaking and print re-orientation to get to a configuration that produced useful results, although I’m still not happy with the ‘final’ (see above) product. After the caddy piece, the flap grip piece was “a piece of cake”, as its geometry was much less complicated, although physically much bigger. The final caddy piece took about 1.5 hours to print, while the flap grip piece went for more than 4 hours!

: Note the tie-wrap mounting loop on the underside of the caddy piece

: View showing the rectangular slot for the flap control bar

: This view shows the holes meant to accommodate the original mounting screws for the remote (didn’t work)

: Rear view with the remote installed

: Front view with the remote installed

: Side view with the remote installed

After getting both parts printed, there was still quite a bit of work to be done, especially on the caddy piece. As it turned out, the cylindrical post on the bottom of the caddy piece was much too long for the matching hole in the top of the flap grip (measurement goof on my part), so the post had to be cut down by hand. Then I had to drill and tap the post for a 4-40 screw so John can tighten the caddy down on the flap grip when he gets the rotation angle correct for his hand placement. Then a 4-40 clearance hole had to be put into the top of the flap grip piece, and the 4-40 screw threaded up into the rectangular slot, through the clearance hole, and then screwed into the caddy mounting post (let’s hear it for double-sided tape and long, skinny screwdrivers!). Also, the bottom surface of the caddy was a lot rougher than I liked (printer misconfiguration?) so I had to spend some time with a file and a sanding block to get this surface even remotely acceptable. I think I’ll try some more experiments to see if I can do a better job at this, so I’ll be ready when John comes back with the next batch of ‘issues’! ;-).

Stay tuned!

Frank

ClearNav Joystick Part 5. Back to the Drawing Board Again Again

Leave a reply

The ‘minimalist’ joystick model from Part 4 was barely out the door before I had another brainstorm. I have always done some of my best thinking while in bed, waiting for sleep to come. My brain is still going over the events of the day, teasing at unsolved problems, and sometimes just as I’m about to drop off, a possible solution or different approach to a problem becomes clear.

In this particular case, it occurred to me that the joystick might not be the best place for a remote control caddy – maybe it would be better on the flap handle instead! This was such a powerful idea that I practically leaped out of bed, padded barefoot down the hall to my office/laboratory, wrote “Flap Handle!” on a post-it note and stuck it to my primary PC monitor. In the light of the next day, awake, alert, and with coffee in hand the idea seemed even better.

Pros:

In a flapped ship, the pilot’s left hand is almost always on the flap handle, and there aren’t any other controls or switches there to interfere with a remote control pad.
The flap handle is on the side of the glider, so wire routing should be easier.
Flap handle mechanical motion is also much more constrained than joystick movement, so interference with other cockpit objects/controls becomes a non-issue.
Button actuation with the left thumb should be much easier, assuming the CN remote caddy could be angled correctly with respect to the flap handle.

Cons:

The flap handle juts out over the pilot’s left leg, so anything mounted to the end of the flap handle is vulnerable to damage as the pilot gets into or out of the glider
The cable from the remote control to the ClearNav (CN) has to get from the end of the flap handle to the side of the glider somehow, and the length of the cable run to the CN changes significantly with the forward and aft motion of the flap control
Getting the right caddy/flap handle relative orientation might be tricky.

A completely independent, but quite significant ‘Pro’ for this idea was that it offered a perfect vehicle to demonstrate the power and flexibility of the 3D printing paradigm to my two grand-children (and their parents) who happened to be visiting over the Labor Day weekend. As it turned out, Danny (12 years old and younger of the two) dived in with great enthusiasm, and by the end of his stay was doing design modifications on his own!

When I broached this idea to my friend, he was also enthusiastic, and sent photos detailing the flap handle dimensions and a photo with the remote control held in what he thought might be the correct orientation.

Armed with this information, Danny and I started work. The overall strategy was to constuct a new handle, shaped more or less like the old one, but with the CN remote caddy section from the previous ‘minimalist’ joystick grip design (see Part IV) attached to the end at an orientation approximating the one shown above.

The first thing I did was to take a short trip to the local hardware store and purchase a 3′ section of 1″ x 1/4″ aluminum flat stock, slightly larger in both dimensions than the actual flap handle. I figured we could use this for initial hole sizing/testing, and adjust later if necessary. Next we designed and printed a test piece to confirm we had the rectangular hole size correct (we did). Next we started designing the replacement handle, incorporating the slot from the above test piece. Somewhere this step, we inadvertently scaled the handle and its internal slot down by a factor of about 2/3, so the first couple of handle prints came out way undersized. Fortunately, we were watching the print as it built up from the bottom, so we recognized the problem pretty quickly and aborted the print after 1/4″ or so. This actually turned out to be a pretty cool technique; we would make a full-length design, but only print a few mm or so, as the cross-section was more or less uniform throughout. As we got closer to what we thought would be a final design, we would let the print continue longer, finally letting the print go to completion. This way we could iterate much faster, and save filament too! ;-).

By about Rev 3 we had a full-sized constant-cross-section cylindrical flap handle printed up that we could slide onto the metal flap handle simulator, but we soon determined it was way undersized for an adult hand (it fit Danny, but…). So, back to TinkerCad for more revisions. Along the way, we also tried out some ideas for creating finger-grip indentations on the front side of the grip, and a corresponding palm indentation on the back side. As it turned out, we were able to use TinkerCad’s ‘Round Roof’ (a semi-cylinder) object for both – at a small scale for each finger indentation, and at a larger scale for the palm indentation.

Basic flap handle grip with finger and palm indentations

So now we had a basic flap handle grip, with finger and palm indentations and a rectangular hole for the flap handle. We still needed a way to attach the CN remote caddy to the end, and we needed some way of getting the control cable down the flap handle to the side of the glider. Our first run at this was a cylindrical cutout in the side of the grip. The idea was that the control cable could be pressed into the cutout through a small gap in the side of the grip, but wouldn’t come back out again without some effort, thus allowing for a smaller diameter hole because the end connectors (telephone style RJ-11’s) wouldn’t have to be accommodated. This was an OK idea in theory, but in practice we were stymied by a limitation in the ‘slicing’ software that converts the solid object model into thin wafer-slices that the 3D printer can handle. The slicer software kept opening up the gap to the point where the cable wasn’t retained – it would just fall right back out again. So, we took a closer look at the connectors and determined that although a 12 mm round hole would be required, we could get away with a 9 x 9 mm square hole! The smaller square hole dimensions would allow us to put the hole entirely inside the flap grip and still not interfere with the main rectangular hole for the flap handle itself – ah, the wonders of 3D printing! ;-).

Next we tackled the problem of how to connect the CN remote caddy section from the ‘minimalist’ joystick grip model (See Part IV) to the flap handle grip. We decided that if we carried the rectangular flap handle hole all the way through the grip, and made the grip a bit longer than the actual handle, we could use the hole as a socket, and make a corresponding ‘plug’ section that would mate with the CN caddy. We created the ‘plug’ section by simply slicing off about 1/4″ of the handle into its own part, changing the rectangular hole to a rectangular solid, and then mating that part with the CN remote caddy. However, we quickly discovered that while it would be possible to print the combined part, it would be ugly, as the combination would require an exorbitant amount of support material due to the weird mating angle. So, we sliced and diced yet again, and printed the plug and caddy sections as independent parts (although we did print them at the same time) and glue the two together post-print. As it turned out, the first revision of the plug didn’t have a big enough indentation for the caddy, so the resultant piece wasn’t strong enough. Back to TinkerCad, where we extended the plug a bit to provide more ‘meat’ for the mating interface. This also necessitated a slight modification to the design of the cord handling tunnel, as it had to be ‘bent’ slightly at the end to allow the cord to be threaded behind the remote caddy and into the tunnel.

TinkerCad drawing for the CN caddy plug with large indent

‘Small Indent’ and ‘Large Indent’ caddy plug versions

At this point we had designed and printed all the required pieces – the flap handle grip, the ‘plug’ to connect the grip to the remote caddy, and the caddy itself. All that remained was to glue the caddy to the plug (which turned out to be a mini-project in itself) and get the two pieces (caddy/plug and grip) down to John for him to try out on the actual glider. Stay tuned for the results! ;-).

: Completed grip system showing all three parts

: Completed system, with plug and caddy mated

: Completed grip

: Complete grip showing remote caddy orientation

: Flap grip with ball-nose spring plungers installed for better fit with flap control bar

: Flap handle simulator with ‘dimples’ installed for ball-nose spring plunger detent

As a not-insignificant side note to this project, my 12 year-old grandson got a great introduction to the world of 3D printing, and quickly became as much a participant as a spectator. The TinkerCad software GUI is very intuitive, and easily passed the 12 year-old usability test. By the end of his 2-day stay with us, he was designing parts in TinkerCad, downloading the STL files to my Linux box, and running my PrintrBot printer to actually print the parts, all with little or no supervision on my part. In fact, the only thing Danny had real trouble with was getting the printed parts off the print bed (protected by blue painter’s tape) because that required a bit more strength and precision with the removal tool (wood chisel with a 1/2″ blade) than he was able to muster.

Frank