Up in arms – Tandy Armatron Dissection

I’ve not been very well these past few weeks, but things are brightening up. To get myself back in the saddle, I decided to do some tinkering and sate a long-standing curiosity.
In the late 70s to mid 80s, Tandy/Tomy made the Armatron. A robotic arm with six degrees of freedom. I never got to own one as a child, as they were off the market by then. But earlier this year I found one going cheaply on ebay, and pounced.
While the magic of a robot arm has perhaps faded a little, it was the conversion of one to steam by CrabFu Steamworks that pointed out the unique property of this toy. The whole thing is driven from one electric motor. The full 6 degrees of motion are driven entirely mechanicly, via two joysticks. The CrabFu article had one tantalising view of the mechanism within, but no details on how it actually managed to channel the power via joysticks.
I had to know how this gearbox worked!

The untouched Armatron

More after the cut!

The Armatron I acquired had seen better days certainly, but largely worked fine. The ornamental piping on the top of the arm was missing, and the lower bushing for the “C” joint. The only functional issue turned out that the wrist of the claw wouldn’t rotate clockwise. A puzzle for later on.

The first thing to do was to label up the degrees of freedom so I could better make notes as I went, and properly ID each drive train.

Each joint on the arm is basicly a set crown gears in the centre of the axis, allowing the drive train to turn corners. Not especially interesting, but worth noting. This end of the drive train wouldn’t be taken apart, as it’s visible already.

The shoulder of the arm is a different story though, and will come later.

Out came two old D-cells, 6 screws that held the base on (yellow arrows), and one wide-flanged screw that held the base of the arm centred (red, obscured by batter hatch).

The two halves simply pull apart.

Next, matching each gear to it’s axis. Each gear was set at it’s own height, which made sense.

In the second image you can see the contacts for the timer. The Armatron was made so you had to perform challenges against the clock. It’s just a cylinder on a worm-gear off the motor though. When it’s turned so all-black shows, the contacts break the circuit and the motor stops. Easy to remove if you want your Armatron to run indefinitely.

To begin exposing the gearbox proper though we have to remove another 5 mounting screws. 4 are obvious, and one is hidden under the radial gears shown above. They just lift out fortunately, and you can see the gears which drive them. (below)

You’ll also need to undo the contacts for the start switch, or you won’t be able to lift the cover out of the way.

And finally, the gearbox itself!

From above you can see two lines of sprockets. The lower one connects to the various gear arrangements that take the power to the base of the arm. The upper sprockets connect to the lower ones.

For those of you that have noticed the studs on the green drums, you may already know how this gearbox works.

Each joystick has three degrees of freedom. Left-right, up-down, and being turned clockwise-anti. At this stage the joysticks lift out of their carriers. Each has a stud underneath and one out from the side. Undoing another two screws lets the cover for the joystick mechanism lift off.

These guides turn the motion of the joysticks into simple back and forth motions of the pins that touch the studs on the green drums.

These green drums are gear carriers.

Running down the centre of these carriers is the main drive shaft, and within each carrier is a drive gear on that shaft, and an orbiting gear attached to the carrier itself.

The joystick pins at rest sit against the wider centre stud of each carrier’s outer surface. There are two type of carrier in this gearbox. Two shorter ones that run from the turning motion of the joysticks to the claw (E) and wrist (F). The four longer carriers have 5 studs on their circumference, while the short ones have three. the longer carriers also have two gears, while the shorter ones just have one.

With some coaxing of the motor, the way this works became obvious quickly.

The “rest position” stud holds the carrier-gear away from any other gear, spinning uselessly. However when the joystick moves, the drive shaft turns the carrier until one of the other studs makes contact with the pin. These pins are positioned so that the carrier rotates around for the carrier-gear to mesh with one of the output gears. Moving the pin to the right stops the carrier at the right stud and the carrier-gear engaged with the bottom output gear. Moved to the left the carrier turns further and stops with the carrier-gear engaged with the upper output gear. That upper gear outputs into the lower gear, so with the extra gear in the train, the output is reversed!

It’s all so simple in retrospect. But it didn’t explain the extra studs on the longer gears. The extra gear on the longer carriers did though.

The longer carriers operate like the shorter ones, except that when the outermost studs are used the second two-tier gear makes contact instead. This has a higher gearing ratio, so the output is much faster. So pushing the joystick a little way runs the movement slowly, whereas pushing it further runs it at higher speed. This wasn’t something I’d previously noticed, so came as a very pleasant surprise!

While investigating this, I noticed the wrist control stuck at the gearbox level, and by turning it manually I eventually spotted a tiny speck of grit, and the marks it’d left in the carrier-gear.

With it removed, the gearbox now works fully again. Bonus! Always amazing that such a small thing can cause such big problems.

With the exploration at a satisfactory end, I reassembled the mechanism. It all still worked, so I decided to press just a little further.

While the axis of the arm are straightforward, the “shoulder” seemed it might be more complex due to the number of drive trains that had to pass through it. So apart it came.

For future reference, if you ever decide to open up the shoulder, you should do it while the arm is detached from the base unit, and while held upside down.

Or all the gears fall out.

The latter shows the gears as they should appear.

Overall a very entertaining and insightful look into how you can control a complex process using purely mechanical systems. Certainly inspiring in a children’s toy that’s older than I am, though with the sheer parts count I can understand why the following versions shifted to individual motors for each axis.

But never forget what can be done with just gears.

With my curiousity now satisfied, I think the Armatron will go back onto ebay this week, with the hope it’ll find a nice new home.

Tags: , ,