totm feb 2020 Bendy Tech - stock mechanics in a new era

[Servo]

 

An Introduction to Bendy Tech

Bendy Tech is a new way to make pure stock mechanisms in KSP, combining reliability and simplicity in a simple-to-build way                            ...

The elevator of @Kronus_Aerospace's upcoming 747SP replica utilizing Bendy Tech

It's no secret that some builders such as myself, @Phantomic, and @HB Stratos continue building stock mechanics despite the presence of robotic parts from the DLC. Whether it's a desire to make your craft accessible to the largest number of people, desirable properties of stock mechanics, we prefer the challenge of keeping things stock, or just that we're old farts who refuse to believe that our once cutting-edge tech is now outdated, we keep building craft that swing their wings, have oversized control surfaces, or have other functions beyond what is conventionally possible with the pure stock parts.

A Stock Innovation born from the DLC

The Rigid Attachment and Same Vessel Interaction options in the SPH. 

The release of the Breaking Ground DLC added the option for easily building gears and other complex mechanisms which were all part of the same craft. With the mechanics of the time, parts could not collide with another part of the same vessel, so the option to toggle Same Vessel Interaction (SVI) was added in the editor. Notably, parts with SVI can only interact with other parts with SVI. If you combine this with "actuators" which have moving colliders (Airbrakes, elevons, tail fins, landing gear, and service bays) to force parts of the same craft to bend - without the use of a servo or hinge from the DLC. To facilitate the bending, chains of weak joints are added between the colliding parts, creating a "spring". This is further facilitated by the use of the Rigid Attachment feature on the actuators, as well as on any "guide" or "hinge" (static parts which either guide or transform the motion).

Introduction to Bendy Tech

A simple example of BendyTech - a 20 degree swing, useful for an all-moving tailplane

This is a very basic example of the principle at work here (fittingly so - it was my first test of the tech's potential). Hidden from view is a stack of six separators connecting the I-beam to the structural panel. The elevators are attached to the structural panels, which sandwich the I-beam. The elevators are set up to respond to yaw controls as pictured (the forwards pair have control authority set to -30, and the backwards ones +30), and have SVI and rigid attach enabled. The I-beam is the only other part with SVI enabled. If you are interested in using this tech yourself, I highly recommend attempting to replicate this setup, as it contains the essential elements of Bendy Tech designs.

Two Axis-Control, high-stress applications

A more complex mechanism allows for two-axis control, and stronger construction to defeat the inherent bendyness of Bendy Tech

The above image shows my first attempt at a practical Bendy Tech mechanism - adding a gimbal to the fixed Reliant engine. While functional, it is less than practical, due to the added size and weight. However, it very clearly shows how the tech can be evolved into more complex applications, while showing some of the limitations. Most notably, the fact that all parts with SVI enabled cannot overlap. The elevons must be arranged in such a way that they can be useful without crossing into each other's control area. 

A craft download for this craft can be found here: https://kerbalx.com/servo/Engine-with-custom-gimbal-no-DLC

For the extremely clever among you, there is a way to minimize the mechanism by rearranging the control inputs and I-beams down to a 1.25m profile. The doing of this is trivial and is left as an exercise to the reader. Or you can see how here: https://kerbalx.com/servo/Gimbal-Assortment

 

Advanced Bendy Tech: Hinges

Classic Stock Mechanics make a return in the Bendy Tech hinges. Here, they redirect the lateral displacement of the I-beam into rotation

Here is the area where I stop claiming innovations as my own. @HB Stratos was the first to see a way to improve the rotation ability of the tech by adding a simple RCS/Thermometer hinge on either side to stabilize the motion. Importantly, small radial-attach parts do not have the ability to enable SVI by default. However, there are two ways around that: Craft-file Editing (which I will leave to HB to explain in his own post), and action groups. You can set the relevant parts to enable SVI  on the stage action group (or whichever AG you use to actuate the parts), and it will function as normal. Another important advanced note here: if you are using hinges, you must use docking ports instead of separators as your spring. Otherwise the hinge will break out of its cage. For an exhausting look at the uses and construction of stock hinges, I recommend a stroll through this thread: 

Advanced Bendy Tech: Guides

Structural panels form the guides for the wings on HB's and my forthcoming F-111 Aardvark replica

The use of guides to keep parts in line is old hat to anyone who's built complex mechanisms (DLC or otherwise). Here, the advance is in the size of the guide - with stacked structural panels, you can fit the guide *inside* a single wing panel. This really takes advantage of the fact that Bendy Tech can fit inside of any part, so long as those parts don't have SVI enabled themselves. This eliminates one of the biggest weaknesses of previous stock tech, and I look forwards to seeing how it is abused in the future.

 

Limitations and Advanced Potential of Bendy Tech

While Bendy Tech is a new and very interesting development, it's not a perfect solution, so using it effectively is a matter of knowing its strengths and weaknesses.

Strengths and Weaknesses of Bendy Tech

Strength	Weakness
Single craft allows control of all elements throughout flight, using standard flight controls	bending parts behave poorly under physics warp
Using Deploy Adjust action groups, can create fine motion, including motion of less than 1m	Limited range of motion, cannot rotate indefinitely
Reliable mechanism not relying on docking ports or KLAW adapters	Long springs can increase part counts of craft
Mechanisms can fit inside of other parts	Without support structures such as guides, large parts are floppy under external loads

 

 

 

Applications and Inspiration board

@Kronus_Aerospace's large elevator. You can see the RCS hinge near the launch clamp. Actuation is provided by a Big-S elevon

Flaps and slats - the flexible tech is well-suited to creating complex mechanisms which expand, especially since control is maintained on any elevons along the length of the flap. 

Our WIP Aardvark, with wing actuation exposed. The wing swings through a 60 degree arc, and the craft also features Bendy Tech all-moving stabilators actuated by tail fins.

 

A Community of Builders

Getting Bendy Tech to this stage is not a singular accomplishment, but the combined work of a community of stock replica builders. If you would like to join us, drop me a PM and I'll get you added to our server

On behalf of HB_stratos, Kronus_aerospace, and others, I hope you'll join us in pushing the limits of what can be done with the base game

Edited February 9, 2020 by Servo

[Kronus_Aerospace]

Well said. One thing that I feel is very important to stress with bendy tech is it's reliability and ease of use in flight. Bendy tech mechanisms require no decoupling, don't need a large internal space to occupy, and one more massive advantage.

You see, when you time warp, the craft effectively reverts to whatever its original appearance was, as any flex is removed from the joints. Since Bendy tech mechanisms keep the craft as a single craft, even if they "break" you can simply time warp to fix them. This makes using them practically and bug fixing significantly easier for obvious reasons.

This ability to utilize time warp in this way is a prime example of the advantages that come with keeping the craft as a single vessel. And by for the largest for me, as I work with massive vessels where the revert and load times are quite significant. From my POV bendy tech is the future of stock mechanics. 

Edited January 8, 2020 by Kronus_Aerospace

[Bubbadevlin]

Damn. I never really considered the importance and usage of inter vessel collisions with single-craft joints... let alone people coming up with such a similar idea 4+ years later.

I really like the designs and its great to see more evolution in KSP mechanics. Plus the tech is DEFINITELY well suited for things like flaps and gimbals

Back in 2016 I had created a REALLY similar system, except that rather than using the new inter-part collisions as the guide bearing I used a decoupling shell. Because of the decoupling shell the joint basically had both the downsides of a traditional bearing as well as bendy tech, so it was really only ever used as a method of rotating engines on my Binary SSTO in order to keep them throttle able and as one single craft. (And yea, that was still in the era of landing gear bearings... man KSP builders have made such an improvement to mechanical things). Later I revisited the idea with my Binary 2, and redockable shell, but the design ended up being overly complex and convoluted. Other than that I only ended up using the idea as a the bearing for a tank turret (once again so that I could control both parts at the same time).  Due to the need for a decouple-able shell (and overall larger size and complexity of bearings) the only real advantage the design had over a standard joint was the ability to control across it. 

 

Also have you guys tried strings of Ant engines as the spring? I found them to be one of if not the most flexable parts. If i put 10-20 in a spiral (like a clock spring) I was able to get the joint to turn 180 degrees + in either direction pretty easily

Oh and i would definitely  like to join your community of builders

Edited January 9, 2020 by Bubbadevlin

[Servo]

A fine example of BendyTech at work here - providing an extremely simple, yet strong, mechanism to control my F8U Crusader's Variable-incidence wing

If you've read through the examples in the original post, the mechanism will be extremely familiar - it's basically copy-pasted from the third example, with the only change being the addition of a guide (a single structural panel sliding between two fixed ones) to prevent roll input from messing up the hinges. It's entirely probably that even that isn't necessary. Another interesting quirk of how Variable-Incidence works is that you can use the grandparent autostrut without it breaking the structure, since the parts matched via symmetry are moving together.

[sturmhauke]

I don't do these kinds of builds myself, but this is seriously impressive stuff.

[έķ νίĻĻάίή]

I've had a go at bendy tech

https://kerbalx.com/jeb_/Reliant-booster-with-stock-gimbal

It actually works ok under time warp and it doesn't spam around!

[Majorjim!]

On 1/8/2020 at 3:54 AM, Servo said:

SNIP

 

I'm way too drunk to soak all this in but it looks awesome! I kinda wish there where no such thing as stock hinges.

[James Kerman]

Congratulations @Servo, you have been awarded thread of the month for this impressive work.

[Servo]

12 hours ago, James Kerman said:

Congratulations @Servo, you have been awarded thread of the month for this impressive work.

Thanks! I'm glad that other people are joining in and pushing the boundaries of the base game

On 1/8/2020 at 4:54 AM, Servo said:

A Community of Builders

Getting Bendy Tech to this stage is not a singular accomplishment, but the combined work of a community of stock replica builders. I

On behalf of HB_stratos, Kronus_aerospace, and others, I hope you'll join us in pushing the limits of what can be done with the base game

I want to thank everyone who's been experimenting with Bendy Tech, and for inspiring me and everyone else. Again, this wasn't the effort of just me by any stretch. It was these guys who saw the potential in my little experiments and shaped it into something useful.

 

On 1/8/2020 at 4:54 AM, Servo said:

Strengths and Weaknesses of Bendy Tech

Strength	Weakness
Single craft allows control of all elements throughout flight, using standard flight controls	bending parts behave poorly under physics warp
Using Deploy Adjust action groups, can create fine motion, including motion of less than 1m	Limited range of motion, cannot rotate indefinitely
Reliable mechanism not relying on docking ports or KLAW adapters	Long springs can increase part counts of craft
Mechanisms can fit inside of other parts	Without support structures such as guides, large parts are floppy under external loads

 

With this news, I would like to share another major breakthrough in Bendy Tech that will allow even more complexity in mechanisms than I ever thought possible.

Fittingly, I have invented Bendy Tech servo motors

The best analogy for how this mechanism works is a marble machine stair lifter like the one below:

To step the motor clockwise, you alternate between pressing action groups 1 and 2. This bends the RCS ball ring 45 degrees, and clips it through the second ring (which has SVI disabled). When you hit the other action group, the sequences reverses and moves it another 45 degrees, by enabling the SVI on the second set of elevons and disabling it on the first. Then, the elevons return to their neutral position, which brings the ring of balls forward another step. 

To reverse the motion, a second pair of action groups is used, which works the same way.

 

Craft Download: https://kerbalx.com/servo/Servos-Servo-Servo

[boomchacle]

I've made a big ass Engine gimbal. A trick for making such large and heavy parts is to use a dumpling fuel tank as a joint held in place using inter craft collisions, otherwise the whole thing will wobble around way too much. It's already a bit wobbly as is, but it is actually useable. You can make parts without the tweakable to collide within craft do it anyways by using action groups, so these solar panels have it set to stage. This engine cluster has a vector engine because this ship is meant to be used as a ferry between a planet and moon, and it needs the extra thrust when the ship is low in the gravity well.

the joint

*edit. I can't make the gif any bigger for some reason, so you'll have to full screen it... weird.

 

 

Edited February 2, 2020 by boomchacle
Gif too small.

[Servo]

 

Improved the servo and made a cleaner video - functions the same, except that by rearranging how the elevons deploy, I have reduced the number of action groups required to just 2 - one to tick it one step, and the other to reverse the direction.

Now it leaves only to use the servo somehow. The mechanism is perfectly suited for tank turrets, but that was a little too expected for me...

So I'm building an excavator. 

I'm currently in the mechanical testing stage on the elbow joint, which has been giving me a bit of trouble. The shoulder joint is a very simple RCS/Thermo hinge (you don't need anything fancier 90% of the time), actuated by a single elevon. That elevon is AG'd to forwards/back deploy limit adjust, giving the entire joint a 60 degree range of motion. More would be possible with an airbrake, but that runs into problems with the airbrakes over-deploying when used with the deploy adjust (devs pls fix)

Because of the airbrake woes I'm having to try out a series of different joint designs for the elbow here, which needs 90 degree actuation, on a small profile. Unfortunately, those don't mesh too well, so I'm still looking for ideas (up to an including recreating the servo joint to drive this joint. Possible, just not implemented). This is an attempt at doubling the effective deploy of the elevon. I think this has the most promise, but the idea needs a bit more refinement to be proven.

Here's another idea that I tried shortly - double stacked elevons. Idea is really simple: one elevon pushes the second elevon, which pushes the target. Unfortunately, due to the fact that joints prefer to translate rather than rotate, this isn't the most useful design without needing a lot of hinges (or one hinge with a bunch of rotators inside it... I've messed with stock hinges that have multiple rotating elements inside a larger hinge, so I could imagine a large Thermo hinge frame packed with thin communotrons all sharing the same hinge frame)

 

If anyone wants to develop any of these ideas further, or has good ideas for how to multiply the rotation of elevons, I'd love to hear about it! I'm genuinely so happy that other people are joining in in my weird pursuit of springs and things, so I'd really like to see what you guys can come up with

[Servo]

 

Honestly, I'm still in awe that I managed to get this thing working. 

For all you curious engineers out there, here's the craft download:

https://kerbalx.com/servo/Excavator

[HB Stratos]

https://imgur.com/a/wKAXJiJ

I have managed to make another breakthrough, usually the locked heaviest part autostrut would prevent bendy tech gears but with the help of a KLAW on free pivot that can be tricked to work anyways. Click the link above for an animated verison.

[boomchacle]

Here's a cute little thing. It's an outboard motor for boats, and it has a few degrees of angling which it can handle. 

 

[Servo]

Made some progress today towards another holy grail of mine - full flaps/slats/spoilers on a pure bendy wing. This profile gives me the best chance of doing that, I think. I've integrated slats (the airbrake in the front is the main actuation) already, with plenty of room in the back for flaps, and at in the middle for spoilers. 

The slats are simple enough - airbrake driven pusher actuation, but with a linear guide instead of a hinge (I used a Mk0 LF tank inside a cage of 4 I-beams). Dead simple, worked on the first try.

The flaps are going to be another monster entirely. This weird looking mechanism is a linear servo which I built today. It works on the same principle as my other servo (stepped motion, each step moving the RCS balls one step out or back), and works reasonably well (given that the actuation is limited to 1.5m or less). More importantly, it lets you "set flaps to position 3" and do other fun things  of that sort.

[Bej Kerman]

What's the difference between this and Animated Attachments?

 

[Servo]

22 minutes ago, Bej Kerman said:

What's the difference between this and Animated Attachments?

 

Bendy Tech takes advantage of the way that the stock game handles non-rigid joints in order to create more or less the same functionality as AnimatedAttachment without requiring the use of any mods or DLCs. Because of this last fact, it's primarily popular with the stock replica community. It even has a few advantages over DLC parts in specific areas, particularly in high strength (stock parts produce an effectively unlimited amount of torque when moving) and dynamic applications such as aircraft horizontal stabilizers (where DLC hinges often have trouble damping vibration or strength, and also can't be controlled using SAS)

 

Here are a few examples of its uses recently:

My Mitsubishi F-2 uses Bendy Tech to create stock all-moving elevators which can hold up to high-G maneuvers 

https://kerbalx.com/servo/Mitsubishi-F-2

 

JCEverett's MiG-31 Foxhound, which despite using the DLC in other parts of the craft, also has Bendy Tech Elevators thanks to their strength, reliability, and the ability to trim them using normal controls.

https://kerbalx.com/jceverett/MiG-31-Foxhound

 

My and @HB Stratos's F-111 Aardvark. This was designed to see how far we could push BendyTech before it broke. We didn't quite find the limit.

Includes a BT Missile Bay, swing wings, swinging pylons, and elevator, and traditional stock mechanics on the landing gear for good measure

https://kerbalx.com/servo/F-111A-Aardvark

 

And to anyone worried about the part count of these planes, @AVeryNiceSpacePenguin managed to fit a bendy elevator into a 127 part F-16

https://kerbalx.com/KAS_Rejects/General-Dynamics-F-16-Fighting-Falcon

 

[Pds314]

WRT the bendy servos, are they a single craft? If so, do they have rotation limits? Like does overtwisting break them?

[Pds314]

Hmm... I have an idea... So awhile ago I made this swashplate. It works in the sense that every that should move does and everything that shouldn't move doesn't, but it's very fragile. Too fragile for liftoff RPM, because it's 13 separate crafts that need to be coaxed into moving coherently.

If the swashplate and pushrod system can be solidified into a single craft and the rotor another single craft though.....

Then maybe this wouldn't be so fragile and could withstand flight loads! I must learn this sorcery.

Edited July 23, 2020 by Pds314

[Servo]

1 hour ago, Pds314 said:

WRT the bendy servos, are they a single craft? If so, do they have rotation limits? Like does overtwisting break them?

Every bendy mechanism that I've built is a single craft. I've seen some hybrid techs, so it is possible (this frees up rotation and multiple SAS pointing solutions, used in a very interesting project to recreate flight instruments the hard way). The servo that I've built managed to take advantage of structural superpositions (I think) in unity's engine to allow for theoretically unlimited rotation. I've tried to recreate the behavior in a different design without success, so it's possible that it was a fluke. Either way, the servos are limited to the rotation speed of the small elevons, so they're less practical for any high-speed application. It did make a killer base for my excavator, though 

47 minutes ago, Pds314 said:

Hmm... I have an idea... So awhile ago I made this swashplate. It works in the sense that every that should move does and everything that shouldn't move doesn't, but it's very fragile. Too fragile for liftoff RPM, because it's 13 separate crafts that need to be coaxed into moving coherently.

If the swashplate and pushrod system can be solidified into a single craft and the rotor another single craft though.....

Then maybe this wouldn't be so fragile and could withstand flight loads! I must learn this sorcery.

That's a lovely sphagetti that you've got there. I've toyed with swashplates before but mainly in the BG sphere and with limited success. The failure there was the hinge strength, so perhaps Bendy Tech is the solution. If you experiment more, please keep us posted in this thread - I'm very curious about how it would go (and will be more than willing to contribute help/advice as I am able).

[Pds314]

On related news, I managed to make bill faint from G forces while he was bolted down...

14 minutes ago, Servo said:

Every bendy mechanism that I've built is a single craft. I've seen some hybrid techs, so it is possible (this frees up rotation and multiple SAS pointing solutions, used in a very interesting project to recreate flight instruments the hard way). The servo that I've built managed to take advantage of structural superpositions (I think) in unity's engine to allow for theoretically unlimited rotation. I've tried to recreate the behavior in a different design without success, so it's possible that it was a fluke. Either way, the servos are limited to the rotation speed of the small elevons, so they're less practical for any high-speed application. It did make a killer base for my excavator, though 

That's a lovely sphagetti that you've got there. I've toyed with swashplates before but mainly in the BG sphere and with limited success. The failure there was the hinge strength, so perhaps Bendy Tech is the solution. If you experiment more, please keep us posted in this thread - I'm very curious about how it would go (and will be more than willing to contribute help/advice as I am able).

Oh with BG I got a swashplate that worked pretty much first real mechanism I made.

Early vehicle with it. Hard to control. Reaction wheels are just for the SAS, they do not react to my control inputs. This also demonstrates how crazy maneuverable a proper rotor is.

 

Edited July 23, 2020 by Pds314

[Pds314]

@Servo
Later BG helicopter tech. I think this broke when they rebalanced the weight of motors and added conservation-of-energy-violating BG rotor blades though.

 

 

Edited July 23, 2020 by Pds314

[Servo]

I've been inspired to start experimenting again with these mechanisms, with the twin goals of minimizing the geometry restrictions and making the hinge strong enough to handle off-axis rotation. Both were challenges to our F-111 build, and I am happy to say that this setup mostly solves those problems.

The first major change was the spring structure itself. Rather than simply placing the ant engines in a single line connected head-to-tail, the ants are zig-zagged with distance between connected ants. This increases the radial offset that is achieved per ant engine, theoretically decreasing the number of parts required for a given range of motion. I haven't looked too hard into the geometry to see if this actually works or not, so it's possible this doesn't actually do anything.

The next change was shifting the actuation point (here, an airbrake) from the base to the actuating bit. This has a few benefits mainly having to do with limiting the space restrictions that we faced (needing the airbrake to swing fully across the fuselage, limiting replicas to ones with wide airframes, such as the -111, but not something like the Su-22). Additionally, this keeps the airbrakes at a constant angle to the airstream, preventing a major drop in performance from the aibrake angle (though I guess this simulates the lower-speed of extended wings...)

A final change is in the hinge structure itself - small SAS wheels surrounded by thermometers turned out to be plenty strong enough. Unfortunately, separators/decouplers were too thin and often clipped through the fins providing the top/bottom of the cage under load - had they worked the thickness of the mechanism would have been cut in half. However, as it stands now, the limiting thickness is the SAS wheel + 2 fins, which works out to be roughly the height of a cubic octagonal strut. All told, that's pretty darn good.

[Starboost88]

I just found this X-wing poking around on KerbalX that uses Bendy Tech for the S-foils.

[Capital_Asterisk]

Not sure if anybody else has done this yet, but enabling/disabling same vessel interaction can make for some very responsive mechanisms. Here's a video of a piston that enables SVI on girders to extend.

It's also possible to make rotating joints, but it seems to be a little harder. I've tried using this mechanism to build an ornithopter; I haven't had much success so far for a bunch of unrelated reasons.

Edited October 24, 2020 by Capital_Asterisk