For KAP a gimbal offers great potential, viable video should be possible, slow exposures in low light and smooth camera movements should be possible with 3 axis active stabilisation.
Gimbal technology has come on leaps and bounds as the drone market goes mainstream. A decent gimbal for anything other than an action cam is still a pricey item but self-build is a reality thanks to dedicated enthusiast work on firmware and software as well as a boom in cut price components. In the 1990s an IMU required an end user defence certificate, now you can pick them up for small change.
Given the huge choice of controller boards, brush-less motors, IMUs, firmware and software it’s a bit daunting to make a leap and get started. It’s all to easy to buy stuff that might be cheap but ends up as scrap componentry when programming gets tough or the bits turn out to be incompatible with each other. It’s bad enough with radios never mind chipsets and the like.
Al ‘Skysnaps 99’ has done the homework and shared details of a beautifully executed gimbal rig on the KAPforum. It is based on the ‘Storm32BCG’ controller board and shipped, complete with matched motors, as a reasonably priced gopro gimbal which he has adapted for both Canon S110 and Ricoh GR cameras. With it he has shot superb still and video footage of kite surfers in action at Woodman Point, Freemantle, Australia.
A really impressive aspect of this gimbal is the software avaible to set up a custom build. OlliW has provided the means to assign the RC channels, control the motor directions and access the PID algorithm parameters. A good instructional video on using the software was another encouragement from Al.
Getting the parts is easy. Adapting the gimbal for the Ricoh GR is tricky, not being too skilled with the soldering iron, (I have lived with the curse of the connector on far too many projects) I decided to do as much as possible retaining the wiring loom intact. This means progress is continually hampered by wires. The reach of the pitch, roll and yaw axes needs to be extended and stretching the loom to fit turned out to be a fiddly business. The longest wires are the ones to the pitch motor and, despite patient fiddling, they are just not long enough to thread their entire length outside the swept area so they cut a corner at the yaw motor support.
Following Als lead, I began by swapping over the motor supports from roll to pitch.
The fixings are miniature Allen headed bolts. The motor supports are a neat clamp, unfortunately I had one freeze on me and had to cut it, with some difficulty I replaced the bolt with a plier tight zip-tie. The output plates were cut through to free the loom. They are stamped aluminium plate which is easily cut with wire cutters. The ferrite ring was snapped off which increases the length of the camera IMU leads. It will be intresting to see if this causes any RF interference.
The new parts are simple enough, cut from 1.6mm x 25mm strip they are achieved with saw, drill and file.
No problem then? Just find the camera x,y and z centres and off we go… but there is a problem and it turns out to be a big one.
The rig must be balanced to work. The balance includes the mass of the motors, and their attendant supports ( including the newly cut parts) as well as the camera, in 3 axes. This is 2 steps forward one step back stuff, each piece needs to be measured, cut, fitted, checked for balance and then re-cut …each fit getting closer than the last ..until I discover the Ricoh has a completely different CoG with its lens retracted ..the pitch balance turns out to be the most difficult part of the job.
So many re-tries:
The IMU housing is cut down from the original, the alloy casting is easy to work.
Once the pitch is done the whole thing (camera, cables, lens in ‘out’ poition etc) can be balanced on the roll axis.
The yaw arm extension is a 3 piece assembly comprising 2 ‘L’ section parts cut down from 25mm x 25mm x 1.5mm angle section with the original output plate to allow for the height and reach required. The path for the loom is routed through a half-and-half cut-out and a notch:
A JR extension cable at the board for the pitch motor feed wires (mot 0) released just enough length to fit the yaw extension arm, 5cm seems to be the shortest available.
On power up the new moments are detected and an aproximate, twitchy, function achieved.
The gimbal assembly under test with a simple string suspension. Not a wire cut, all functions reading ‘normal’.
Next, more balance tuning and PID tweaking…then suspension, RC and down-link to do…busy busy!