AotPR part 4 - Adding a Camera

Having built the successful prototype and started using it to test the programming needed, I dug out the aluminium profile and started building something more functional...

Click to see video.

fig I: gearing and pulley with aluminium structure

Click to see video.

fig II: from the other side, showing the gearing in more detail

Lucky Coincidences...

After building the first test model, I wanted to double-check that the motors would be strong enough to lift the camera - without being geared down so far that it would take forever... This required a much stronger structure than would be easily built with LEGO, so I decided to build the outer U-frame out of aluminium profile, and somehow fix bits of LEGO to it. By some quirk of Fate the M5 bolts used for attaching things to the t-slot nuts just happen to be a really snug fit in the holes in Technic LEGO:

fig III: bolt through LEGO

fig IV: t-slot nut in the t-slot

fig V: attaching LEGO at 90? - to the right you can see one of the angle brackets available to go with the alu profile

Because of this rather lucky accident, I will be using rather more LEGO pieces in the final model than originally planned. All the main structural pieces will still be alu profile and angle brackets (including the central square section that holds the camera), but more of the fixings etc. will be LEGO.

Gearing

LEGO gears are not particularly strong, although the new wider teeth ones are better. For that reason, most of the gears used were doubled-up to reduce the pressure on each individual gear:

fig VI: overhead view of gearing

The only 'single' gear pair is the one from the motor itself as there is little torque involved at this point. Overall, the gear train has a 45:1 ratio from the motor to the 'winch drum'.

fig VII: main pulley, on the side of the camera mounting frame

If you are being observant, you will spot that the cord is actually tied to each pulley rather than it actually acting as a pulley. The weight of a Canon 5D and 15mm fisheye with enough offset for the no-parallax-point results in about 100Ncm turning force. That's roughly the same as hanging 1 kilo weight off a pulley with 10cm radius.With a 3cm radius pulley such as the one used you would have to hang a weight of over 3kg from the string just to counterbalance the weight of the camera - far more than any regular pulley could cope with without slipping.
The current plan is to use toothed belts and pulleys for this section which only slip under extreme circumstances. The other options would be a continuous geartrain to the camera-holding frame or a 'closed-loop' pulley system where the cord/belt goes all the way around but is still fixed in place on one or more of the pulleys rather than being loose like a regular pulley/belt system.
For reference, the small-to-large pulleys add roughly another 4:1 ratio for a grand total of about 180:1 from the motor to the camera mounting frame.

Rotating the Head

Rotation of the head (yaw movement) will use a similar mechanism, which a large toothed pulley fixed to the part of the head which attaches to the tripod. A smaller toothed pulley will effectively 'pull' the head around. This should require far less gearing than above as there will be much smaller forces involved.

Shutter Release

The next step was a shutter release. Rather than trying to accurately place something over the shutter button on the camera I had already decided to go for a remote release so that the motor for this purpose would be stationary. This way it would not add to the weight of the camera/lens and would not need to be moved every time a different lens was mounted with a different no-parallax-point.
A full press on the remote release (a cheap far eastern one) takes a lot of pressure, even with it's spring removed. One option would have been to remove the main button and press directly on the micro-switch, but it makes a very small target. Another issue is that the motor has to move relatively freely - if it comes up against too much resistance it will stall and drain the batteries very quickly, if not causing permananent damage.

After fiddling around with multiple variations of levers, springs and rubber bands, I came up with the following quite elegant solution:

fig VIII: shutter release

The remote release is held on to two parallel beams using a rubber band - easy to remove again for other purposes. Then a long lever is held against the button by a rubber band leading to an arm on the motor. The motor turns through about 90? (without the motor stalling!), increasing the pressure on the release button to fire the camera shutter. By controlling how long the motor stays in that position before returning, we can control how long the shutter is pressed - very useful for bracketed shots.

For a closer view, here it is again with a few pieces removed:

fig IX: shutter release at rest

fig X: shutter release during action

So, with the addition of a shutter release, we can see two of the main controls in action. I've added a red bar to the shutter release motor so that it's movement is visible in the video, but if you turn up the sound you should be able to hear sets of three distinct shutter clicks as bracketed shots are fired:

Click to see video.

fig XI: gearing, pulley and shutter release

That's all for now. I'm going to be busy with a software project for the next couple of weeks so you aren't likely to see much progress for awhile, but it's coming along nicely!

Ian