speculating about motorised dolls

[curious_person]

There was an interesting conversation in the chat going over the issues around motorized skeletons(e.g. cost of motors, complexity of software etc).
r.e fully moitorised , I still think they're far off but it was interesting to investigate the possibilities and limitations. (there'd be a sliding scale of capability - I dont think they'd need to be able to walk to be interesting)


One aspect that bugs me is the abiltiy to access internals.
For something so mechanically complex, you'd want to be able to open it up for maintainance.
That would inevitably mean seams in the surface, right?

What directions could the design go in , what tradeoffs would there be.

Are there places where seems would work ok without being stretched when it bends

what would be the most popular positioning r.e. being able to cover it with clothing
would that be offputting ("this doll really must wear at least a suspender belt otherwise you're going to see a visible seam")

how many such seams would there need to be

Or would people rather put up with the need to cut and glue the TPE surface back together for fitting

[Technician]

IMO motors would be a terrible idea. Motors are noisy and in order to be able to do any level of useful work would need to be run through a gearbox resulting in painfully slow movement or be large, heavy, and inefficient.

The model that makes the most sense is the human body, and use hydraulic muscles or cylinders and a single pump with a valve/distribution manifold.

[rubherkitty]

I'm working on a doll that will have a flexible rubber hose back bone and rocking hips. Body will be a composite of different foam covered in a silicone coated cloth skin.
I plan to install 1 DC gear motor of low RPM and high torque w/a variable speed controller. Hopefully the motor will make the dolls torso thrust and hips rock slightly. If there's still plenty of power, I may be able to add in some other mild motions via pull cable and opposing spring.
Motor and battery pack access will be through the upper back. Will have to figure that all out when I get there.
viewtopic.php?f=6&t=125878
---------------------------------------------------------------------------------------------
Actually here is the write up on that phase of the project. I haven't added it to the doll build thread as I'm not really there yet.

Overall doll is reinforced plastic skeleton, foam body and silicone coated cloth skin.

Overall doll motion plan is:
12v single battery or 10 C-cells. A single 12v could be left in doll or removed for recharge or replacement. The 10 C-cells will be removed from doll for recharge or replacement. I've read it is best to charge and monitor individual batteries solo as each has their own condition and life span.
This is why if I choose the C-cell route I want to use a battery holder and not buy a wrapped C-cell pack.

Battery connects to simple ON/OFF switch or a variable speed controller.

Motor will be a DC gear motor, possibly right angle.
IG32 Right Angle 12V 1 to 71 83 RPM
IG32 Right Angle 12V 1 to 100 60 RPM

Mechanics:
Motor will be attached inside rib cage to back bone.
A clamp on short arm will be attached to motor output shaft.
The short arm will have a post to operate actuation tube.
Top end of tube will have a Heim joint to allow for rotation on shaft.
Bottom end of tube will have a ball joint as this is a non rotating connection.

Actuation:
Motor output shaft rotates output arm.
Output arm drives actuation tube up and down.
Hip assembly rocks either up or down depending upon stroke cycle off central pivot point.
Hip assembly has rotation stops that allows a limited motion. Consider 25% of the actuation tube stroke in either direction.
When the hip rocking limit is reached the actuation rod connecting ends then force the rib cage assembly and hip assembly to either separate or contract depending upon stroke cycle. This is allowed by flexing of a soft rubber flex tube connection.
Desired effect is a hip rocking motion along with a chest thrusting and arching motion. Motion should be mild and fluid. Not herky, jerky robotic.

Mechanics illustration.

IMAG0002.JPG (52.38 KiB) Viewed 3937 times

Motor output shaft arm "which will control the length of stroke", actuation tube OAL and connecting point on hip will all be adjustable. While the hip rocking motion will have direction and distance limits the chest to hip connection motion will be unknown until the system is installed. The foam body and position of doll will also effect the motion. Cowgirl, missionary or doggy could all have a separate motion due to gravity or resistance to the item the doll is placed upon.

Electrical basics.
Using a ON/OFF switch a fully charged 12V battery will rotate the shaft of a 60 RPM motor 60 times in a minute.
As the battery runs down a loss of speed and/or torque is to be expected.
A variable speed controller will only be able to reduce the speed of a set RPM motor, never increase it. So if 60 strokes of the doll is optimal, but you want to vary it from 50 - 75 strokes you will want a higher base RPM motor, but use a speed controller.
RPM of a motor can be reduced via a voltage controller, but that simply cuts voltage to the motor reducing speed and possibly torque. Optimally you want to control speed, but still allow full torque of the motor.
In this doll project having a speed controlled that allows for reverse direction is not necessary, but may be desired if the reversed actuated of the doll offers a different and desirable motion.

I'm not looking to install an Arduino, etc at this time to code specific motions or timing sequence. Although I have studied into finding a simple controller that would allow the motor/ doll actuation to match the beat of music, but found any ideas or instructions lacking.

Additionally, if the motor has the power surplus above this basic motion plan. additional smaller actuation rods or cables could possibly be attached to the base of the skull, hinged jaw, shoulder blades, etc to offer additional motions for making the doll more animated.
-------------------------------------------------------------------------------------------
But this is a simple 1 motor design.
As Technician said, if you want a doll to move w/a number of various independent motions it is probably best to use some type of hydraulic system. Pump, manifold, hoses and cylinders. You will still need a bank of micro servo's to operated the hydraulic control valves and a controller to give them the signal. Either a pre written script or reactionary to stimuli. You will need battery power for the hyd pump, servo's and controller.

[curious_person]

ok so.. Hydraulic dolls, not motorised dolls. (just a motor pump and valves). interesting. I wonder how much "hydraulic routing" could be 3d printed? or would you definitely need pipes and so on (some parts more rigid, others flexible) dual material printing? sounds like such a thing would be a nightmare to assemble

[rubherkitty]

Why would you even try to 3D print hydraulic tubing or hose when you can buy it off the shelf already?
Never have researched how small you can get it though and what pressures they can withstand. I don't think you would want to use over 25-30 lbs of pressure to move a doll limb so anything rated for 50lbs should be good.
In use on hydraulic equipment, lines are generally steel tubing except where they cross hinged joints then they use hose.

I'm not sold on hydraulics myself for a doll. One major issue is the motor/ pump has to be running all the time to be at the ready when a valve is actuated. With servos, no motor is "running" until the electrical signal is sent to the servo. Servos come in so many sizes and power options. They can also be individually controlled for speed and length of movement. And the only thing running to a servo is some small wires.
I guess a hydraulic cylinder can be controlled for speed and length of movement, but you will still need a computer controlled electric servo or linear actuator to control the valve! Basically you are creating 2 systems instead of just 1. Now if when a servo pushes a valve and the hydraulic system actuates 6 cylinders of desired movement you have a gain of 5 actions.

Most of the animatronic makers use servo's, but their animatronics are for display and so light duty such as light metal frames, light foam and thin latex skin.
Now if you're talking a robotic doll that can fuck you it's going to have to strong, tough skinned and will probably be quite heavy. You might need hydraulics for the heavy movements and maybe get by using servo's for the light stuff.

[curious_person]

Why would you even try to 3D print hydraulic tubing or hose when you can buy it off the shelf already?

the question is how much of the full assembly (hydraulic tubes -> "muscles", surrounding flesh) could be 3d printed (e.g. consider the extreme depicted in the westworld intro).
I see them 3d printing rocket engines, although admitedly that's pretty high end equipment required

[rubherkitty]

Why would you even try to 3D print hydraulic tubing or hose when you can buy it off the shelf already?

the question is how much of the full assembly (hydraulic tubes -> "muscles", surrounding flesh) could be 3d printed (e.g. consider the extreme depicted in the westworld intro).
I see them 3d printing rocket engines, although admitedly that's pretty high end equipment required

I was just considering small sealed hydraulic cylinders that would push/pull the bones of the doll. Hydraulic expanding bags/muscles is something I have not studied.

Small brake tubing and hose used on a motorcycle would probably work fine for a doll.

[Technician]

There's plenty of plastic tubing and fittings that could handle the pressure, they get commonly used in pneumatics.. I kind of pictured a pressure reservoir of some sort and a reed valve piston style pump moving constantly and slowly to simulate a heartbeat.

[doll-o-grapher]

Hydraulic and Pneumatic are going to be MUCH louder then decent brushless motors or stepper motors, both of the later can be made even quieter using helical Cut gears in the reduction gearbox.

And all of the above has too many moving parts.

The term you want to put in to you favorite search engine is "Linear Motor" or "Linear Actuator"

Attach a cable from one side of a hinge/joint (lets say elbow) to side A of the linear motor.
Attach a cable from side B of the linear motor to an anchor point a few cm below the shoulder on the same limb.

Now when activated the two halves of the linear motor pull the joint towards the anchor.

If you want to get real fancy, sheath the motor in fiberglass weave, and cover that in super soft TPE, so when the motor gets shorter the TPE bulges under the dolls skin.

Use a distributed network of Arduino's or Raspbery Pi's one for controlling each leg and foot, one for each arm, one for each hand and one for the spine and neck, and a final one for the face, all back to a central one to send commands and adjust according to gyro's from a Quadcopter flight controller.

You did want this to walk, right?

[curious_person]

> You did want this to walk, right?

I'm guessing a doll is not going to walk for less than $100,000... autoposition change and adjustment is the first goal

[Mr Franz]

Valentina Girl just announced a motorized version of their dolls. Still in the development stage and won't walk or anything like that, but apparently it'll do most of the work in bed.

http://www.valentinagirls.com/sexrobot.html

[rubherkitty]

You did want this to walk, right?

First she learned to talk.
Then she learned to walk.
Yesterday she ran.
Today I went to the store for anal lube.
Tonight I sit all alone, she learned how to turn a door knob.

[Mr Franz]

You did want this to walk, right?

First she learned to talk.
Then she learned to walk.
Yesterday she ran.
Today I went to the store for anal lube.
Tonight I sit all alone, she learned how to turn a door knob.

Hey, as long as you have the anal lube and access to the doorknob...

[doll-o-grapher]

I give you an answer on simple mechanisims, and you blokes fixate on the joke at the end?

[rubherkitty]

From what I've studied about linear actuators is they're kind of a slow movement due to their rack and pinion gear design. Great torque though. Fine in some doll applications, but doesn't fit the bill for everything. They also get hot quickly due to their high RPM forward and reversing of the drive motor.

Thanks for the input and ideas.