The View Up Here

Random scribblings about kites, photography, machining, and anything else

Re-Thinking the T2i Rig

Posted by Tom Benedict on 08/07/2011

I’m still finishing up a number of projects, so I haven’t been able to begin construction of the T2i panorama rig just yet. In some ways I wish I had the rig in-hand this very moment. In others I’m glad it’s taking a while: I re-designed it. Again. And as soon as I thought I was done, I re-designed it again.

The initial design called for a Geneva mechanism that would index the camera through a number of pan positions at a fixed rate. I planned in a Hall-effect sensor and magnet so I could compare the rotation rate to a watch crystal and continuously tune it through a PID loop so it would remain consistent even as the rig’s batteries drained. This had all the benefits of Vertigo’s burst-KAP technique without the drawback of the high shutter speeds required to freeze the action. Since the goal of this rig is to do nighttime or late-evening panoramas, stopping the camera at each position is a requirement, and the Geneva fit the requirements perfectly.

Since then I’ve revised the design several times. The first revision was to add the ability to install a two-axis flywheel on the rig’s pan axis. This is to damp the high frequency rig motion in roll, pitch, and yaw, and allow for the slower shutter speeds necessary for late evening photography. I still think there’s a lot to be said for this approach, and plan to pursue it in the future. But for now I’m just adding the bolt pattern so the flywheels can be added later.

Meanwhile two other discussions have happened that changed my mind on a couple of things. The first was a thread on the KAP Forum from Mike LeDuc, who built a 3-axis stabilized pendulum rig that he’s used with great success. In addition to using an IMU to control three servos to take out extraneous motion in roll, pitch, and yaw, it also has the provision to trip the shutter on the camera only when it is at the apex of a swing, when the motion is at a minimum. I loved the idea as soon as I saw it, and still like it. I think it compliments the flywheel in terms of adding rig stability, and would result in sharper pictures at longer shutter speeds.

While trying to design around this idea, another thread on the forums came up, in which a fellow KAPer who goes by the name of yurik_ryba posted a link to a Pololu Robotics product, the Micro Maestro USB servo controller. I have a long-standing love of the design work done by the folks at Pololu, so I had to give it a look-see.

The Maestro line of products are all servo controllers with attached microcontrollers. They’re intended for robotics or animatronics, so there’s a provision for programming a sequence of motions into them. A Windows-based application lets you write the scripts that control what they do. Even better, each I/O slot can either drive a servo, act as a digital input or output, or act as an analog input. The Micro Maestro has six channels, but by the time you get to the largest version there are 24 channels of I/O and the added capability of having your I/O channels act as PWM outputs for DC speed control. They’re wild!

Of course I ordered one.

So for the moment the plan is to go with a mechanically simple rig. It’s basically a Brooxes BEAK with a geared pan axis. No sense re-inventing the wheel, so I plan to use his Deluxe Gear Set, which interfaces directly to the Brooxes Utility Frame I plan to use as the framework of the rig, similar to how it is done in the BEAK. The pan gear will be modified to include a ring of 1/8″ diameter, 1/16″ thick rare earth magnets. A ratiometric Hall-effect sensor will then be wired in on one of the unused I/O channels as an analog input device. Careful placement of the magnets should let me use this as a fairly accurate encoder.

One of the cooler features of the Maestro controllers is that each axis can have its own acceleration and maximum velocity settings. This should let me replicate the behavior of a Geneva without the mechanical complexity, and by using an encoder that gives a continuous near-sinusoidal output, I should be able to have it stop at an arbitrary number of positions as it completes its circle. The scripting language includes a command to detect if all of the axes have stopped, so with a sufficiently slow pan velocity and acceleration, I should be able to time things so the rig has settled down after each move.

Since this rig only has one axis, one shutter, and one encoder input, that still leaves me three I/O lines to play with. Sparkfun offers several three-axis gyro devices that would let me sample the rate of roll, pitch, and yaw. Once the rig is built and tested, if further stabilization is required I should be able to add Mike LeDuc’s apex sensing to my rig as a plug-and-play option. And adding the flywheels is a matter of bolting them on. This wouldn’t offer me the three-axis stabilization that Mike’s rig has, but I can cross that bridge when I get to it. Adding three-axis stabilization of a DSLR camera is not a lightweight affair, and up to this point the rig would weigh very little compared to the weight of the camera and lens. Light, simple, and effective is what I’m after.

In case anyone is reading this and is shaking their head, thinking I’m rushing headlong toward disappointment, please understand I know the risks. Earlier this week I talked about some of this with Michael Layefsky, whose nighttime picture of Jack London Square in 2009 got me started on this whole idea. He convinced me that night time panoramas from a kite probably are a fool’s errand. Even so, it’s a risk I’m willing to take. Look at it this way: At the very worst it won’t work for panoramas, but will occasionally (say 0.5% of the time) produce a good nighttime photograph. So be it. Roll the clock back into the sunset hour and it will still provide better panoramas than I can currently make. Roll the clock back a little further to the golden hour and it will beat my current setup hands-down. No matter how I cut it, this is a good direction for me to go.

And even if the whole thing is an utter and complete flop, the rig is entirely constructed of Brooxes components. That’s like building a KAP rig out of Legos or Meccano parts. If all else fails, take it apart and build something else that does work. Nothing is wasted except for my time, and even that’s not a waste since I’ll come out of it knowing more than when I went in.

All in all, life’s pretty good.

– Tom


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: