Weekly Report #4: Trial and error

Mischa Daams
2016-08-29 12:03
This week's report is a bit technical, but since this residency for large part is about making an instrument and controlling it, the making process that is involved reflects a large part of my activities: building and learning about mechanics, motor control and precision from the bottom up.

My concrete idea is the following: I want to build a computer-controlled robotic camera rig that allows me to have precise control over the X-Y movement of the camera and over the light intensity, the focus wheel and the zoom wheel.
I will use this system to explore the character and dynamics of visually moving and emerging growth patterns that result from a 'conversational' feedback loop of information between two basic media: a video camera and lcd television screen. This feedback loop principle will serve as an interactive video installation that generates a videostream of visually complex organic growth patterns that are manipulated by the pre-programmed machine Γ‘nd by intervention of the audience. I will get into further detail on what drives me in this project and my approach in next weeks post.

Because I'm a noob in technical engineering it took me some time to get acquainted with what I needed and hence I spent weeks before going to Korea, browsing the internet looking for apparatuses and DIY projects that sort of do the things I want my rig to do. Now, there are several commercial systems on the market that do similar thing to what I want, but not as specifically.

Most electronically controlled pan-tilt heads have a dc motor of which you can simply control the strength of rotation, which usually just defines the speed of the rotation. I want to be able to tell the motors that control all the parameters of the camera, to make slow movements and tell them precisely where they need to go, like a choreography. This is important because this would allow me to repeat certain positions and movements in the same way again and again. This exact repeatability in control is important because the system that I want to control is in itself already very unpredictable and complex; it derives its 'random-ishness' from the translations taking place between the used media themselves. So in a system in which random factors are already very present and visible, it is good to exert tight control over other factors if you want to start to understand the underlying mechanics and behaviors: which is an essential part of my research.

Regarding the pan-tilt mechanism that should move the cameras; after much research and comparison between different model approaches I found a solid -open source- example for the moving base of the camera. A description of how it is built, the drawings of the parts and even the parts that it uses are publicized online. Since I never worked with these mechanics I was happy to find such an in-depth, reliable source as a basis for my project.

However, this basis did not work with my motors and after much trial and error in trying to apply my motors retroactively I decided to completely draw all the hardware involved in my project from scratch using precise computer illustrations, so that laser cut parts would actually fit the rest of the parts and mechanics exactly. This took a couple of days, but was really worth it. At the end it really felt I had achieved something πŸ™‚

For my project I need precise motors. The E.I. Lab where I work has recommended to use Smart Servo's for my project. These motors are ideal because they give feedback on their location or speed, the load that they carry, their temperature and they are very precisely and repeatably controllable. So I ordered a bunch of these Korean servo motors: Robotis Dynamixels; basic ones and very strong ones.

Along the course of my project I have needed to make custom parts that I could not find elsewhere in Korea, using a laser cutter and 3d printers. With open-source code I was able to customize gear rings to exactly fit my camera lenses. Overally speaking I learned a lot about the parts that that make up a mechanical system. At the end of this week I was able to assemble the pan-tilt rig.
Total 1

  • 2016-09-12 15:54
    이번 μ£Ό λ¦¬ν¬νŠΈλŠ” 기술적인 뢀뢄에 κ΄€ν•œ 이야기 μž…λ‹ˆλ‹€. 사싀 이번 λ ˆμ§€λ˜μ‹œμ—μ„œ μ œκ°€ μ§‘μ€‘μ μœΌλ‘œ λ‹€λ£¨λŠ” 뢀뢄듀이, 도ꡬλ₯Ό λ§Œλ“€κ³  그것을 μ‘°μ’…ν•˜λŠ” λΆ€λΆ„μž…λ‹ˆλ‹€. λ”°λΌμ„œ μ €λŠ” 기계에 λŒ€ν•΄ κ³΅λΆ€ν•˜κ³ , λͺ¨ν„°λ₯Ό 더 μ •ν™•ν•˜κ²Œ μ‘°μ’…ν•˜λŠ” 뢀뢄에 λŒ€ν•΄ μ—°κ΅¬ν•˜κ³  μžˆμŠ΅λ‹ˆλ‹€.
    μ €λŠ” μ»΄ν“¨ν„°λ‘œ 맀우 μ„Έμ„Έν•œ λΆ€λΆ„κΉŒμ§€ μ‘°μ’…λ˜λŠ” 카메라 λ‘œλ΄‡μ„ λ§Œλ“€ κ³„νšμž…λ‹ˆλ‹€. κ·Έλž˜μ„œ λΉ› 세기와 쀌 ν¬μ»€μŠ€μ— μ„Έμ‹¬ν•œ 주의λ₯Ό 기울이고 μžˆμŠ΅λ‹ˆλ‹€. 이 μ‹œμŠ€ν…œμ€ λΉ„λ””μ˜€ 카메라도 LCD TV 슀크린 μ‚¬μ΄μ˜ κ΄€κ³„μ—μ„œ λ°œν–‰ν•˜λŠ” μΌμ’…μ˜ λŒ€ν™” ν˜•νƒœμ˜ 루프 μ˜μƒμœΌλ‘œ λ‹€μ–‘ν•œ ν˜•νƒœλ‘œ μ›€μ§μ΄λŠ” λͺ¨μŠ΅κ³Ό μ¦ν­λ˜λŠ” ν˜•νƒœμ˜ νŒ¨ν„΄μ„ 결과뢈둜 λ³΄μ—¬μ€λ‹ˆλ‹€. λ‹€μŒ λ¦¬ν¬νŠΈμ—μ„œ 제 μž‘ν’ˆμ— λŒ€ν•œ 더 μžμ„Έν•œ 이야기λ₯Ό λ“€λ € λ“œλ¦΄κ»˜μš”.
    사싀 μ €λŠ” 기계 곡학 μ „λ¬Έκ°€κ°€ μ•„λ‹ˆλΌμ„œ ν•œκ΅­μ— 였기 일주일 전에 μ œκ°€ 직접 λ§Œλ“€κ³  μ‹Άμ—ˆλ˜ ν”„λ‘œμ νŠΈμ— λŒ€ν•΄ κ΅¬μƒν•˜λ©΄μ„œ 도ꡬ듀을 κ²€μƒ‰ν•΄λ΄€μŠ΅λ‹ˆλ‹€. λ¬Όλ‘ , 기계 μƒμ μ—μ„œ λΉ„μŠ·ν•œ 것을 ꡬ할 μˆ˜λŠ” μžˆμ—ˆμ§€λ§Œ μ œκ°€ μ›ν•˜λŠ” μ™„λ²½ν•œ λͺ¨μŠ΅μ€ λ‚˜μ˜€μ§€ μ•Šμ•˜λ˜ 뢀뢄도 μžˆμ—ˆμŠ΅λ‹ˆλ‹€.
    λŒ€λΆ€λΆ„ μ „μžμ œν’ˆμ—μ„œλŠ” DC λͺ¨ν„°λ₯Ό μ΄μš©ν•΄μ„œ νšŒμ „μ˜ 속도λ₯Ό μ‘°μ’…ν•˜λŠ”λ°μš”. μ €λŠ” 이 속도λ₯Ό 천천히 μ›€μ§μ΄κ²Œ ν•¨μœΌλ‘œμ¨ 마치 λͺ¨ν„°λ₯Ό μ΄μš©ν•˜λŠ” μ•ˆλ¬΄ 같은 λͺ¨μŠ΅μ€ μ—°μΆœν•˜κ³  μ‹Άμ—ˆμŠ΅λ‹ˆλ‹€. 같은 μœ„μΉ˜μ™€ ν–‰μœ„λ₯Ό λ°˜λ³΅ν•  수 μžˆλ„λ‘ ν•˜λŠ” 것은 제 μž‘ν’ˆμ—μ„œ ꡉμž₯히 μ€‘μš”ν•œ λΆ€λΆ„μž…λ‹ˆλ‹€. λ¬Όλ‘  이미 μž‘ν’ˆ μ†μ—μ„œ μš°μ—°ν•˜κ²Œ λ“±μž₯ν•˜λŠ” ν˜•νƒœλ“€μ€ μ œκ°€ μ‘°μ’…ν•  수 μ—†λŠ” λΆ€λΆ„μ΄μ§€λ§Œ, μ €λŠ” 이 뢀뢄듀을 쑰금 더 μ •ν™•ν•˜κ²Œ μ‘°μ’…ν•  수 μžˆλ„λ‘ μ‹œλ„ν•˜λŠ” 것이 κΈ°κ³„μ˜ 행동을 μ΄ν•΄ν•˜λŠ”λ° 도움이 λœλ‹€κ³  μƒκ°ν•˜κ³  이것은 제 연ꡬ에 μžˆμ–΄μ„œ ꡉμž₯히 μ€‘μš”ν•œ λ‚΄μš©μ΄μ£ .
    팬 ν‹ΈνŠΈ ν—€λ“œ (μ΄¬μ˜κΈ°κ°€ μˆ˜ν‰μ΄λ‚˜ 수직으둜 움직일 수 μžˆλ„λ‘ μ‚Όκ°λŒ€ μœ„μ— μžˆλŠ” 마운트) 에 λŒ€ν•˜μ—¬ μ—°κ΅¬ν•œ 끝에 카메라λ₯Ό 움직일 수 μžˆλŠ” μ˜€ν”ˆ μ†ŒμŠ€λ₯Ό 찾게 λ˜μ—ˆμŠ΅λ‹ˆλ‹€. μ–΄λ–»κ²Œ μ„€κ³„ν•˜λŠ”μ§€ 그리고 λ§Œλ“€μ–΄μ•Ό ν•˜λŠ”μ§€ 각각의 λΆ€ν’ˆμ˜ λ„λ©΄κΉŒμ§€ λͺ¨λ‘λ‹€ μ˜¨λΌμΈμ— κ³΅μœ λ˜μ–΄ μžˆμ–΄μ„œ μ €λŠ” λ‚œμƒ 처음으둜 제 μž‘ν’ˆ 세계λ₯Ό λ”μš± 깊게 연ꡬ할 수 μžˆλŠ” 이 기계λ₯Ό λ§Œλ“œλŠ” 것을 μ‹œλ„ν•˜κ²Œ λ˜μ—ˆμŠ΅λ‹ˆλ‹€.
    ν•˜μ§€λ§Œ 이런 기초적인 뢀뢄듀이 제 λͺ¨ν„°μ™€ 잘 μž‘λ™ν•˜μ§€ μ•Šμ•„μ„œ μ—¬λŸ¬ 번의 μ‹œλ„ 끝에 λ ˆμ΄μ €μ»€ν„°λ₯Ό μ΄μš©ν•΄μ„œ μ •ν™•ν•œ ν˜•νƒœλ₯Ό λ§Œλ“€κΈ°λ‘œ κ²°μ •ν–ˆμ–΄μš”. λ¬Όλ‘ , λ§Žμ€ μ‹œκ°„μ΄ μ†Œμš”λ˜μ—ˆμ§€λ§Œ 개인적으둜 ꡉμž₯히 λΏŒλ“―ν•œ μ„±κ³Όμ˜€λ‹€κ³  μƒκ°ν•΄μš”. 이 ν”„λ‘œμ νŠΈλ₯Ό μœ„ν•΄μ„œλŠ” μ •ν™•ν•œ 규격의 λͺ¨ν„°κ°€ ꡉμž₯히 μ€‘μš”ν–ˆμŠ΅λ‹ˆλ‹€. λ‚˜λΉ„μ˜ E.I.Labμ—μ„œλŠ” Smart Servo λΌλŠ” 것을 제 ν”„λ‘œμ νŠΈλ₯Ό μœ„ν•΄μ„œ κΆŒν•΄μ£Όμ—ˆκ³  μ €λŠ” Robotis Dynamixels μ΄λΌλŠ” ν•œκ΅­μ˜ λͺ¨ν„°λ₯Ό λ‹€λŸ‰μœΌλ‘œ κ΅¬λ§€ν•˜κ²Œ λ˜μ—ˆμ–΄μš”. λ¬Όλ‘ , μ œκ°€ μ›ν•˜λŠ” 거리와 μŠ€ν”Όλ“œλ₯Ό λ‹΄μ•„λ‚Ό 수 μžˆλ„λ‘ 잘 μž‘λ™ν•΄ μ£Όμ–΄μ„œ 정말 기쁘게 μƒκ°ν•˜κ³  μžˆμ–΄μš”.
    λ¬Όλ‘ , μ œκ°€ ꡬ체적으둜 μ°ΎλŠ” λΆ€ν’ˆλ“€μ€ ν•œκ΅­μ—λ„ μ‘΄μž¬ν•˜μ§€ μ•Šμ•„μ„œ 3D ν”„λ¦°νŒ…μœΌλ‘œ μ œμž‘ν•΄μ•Ό ν•˜λŠ” 뢀뢄듀이 λ§Žμ•˜μŠ΅λ‹ˆλ‹€. μ˜€ν”ˆ μ†ŒμŠ€λ“€μ„ ν†΅ν•΄μ„œ μ œκ°€ μ‚¬μš©ν•˜λŠ” 카메라에 μ ν•©ν•œ κΈ°μ–΄λ₯Ό μ£Όλ¬Έ μ œμž‘ ν•΄μ•Ό ν•˜λŠ” 상황듀을 자주 λ§ˆμ£Όν•΄μ•Ό ν–ˆμ§€λ§Œ, 그만큼 기계λ₯Ό μ œμž‘ν•˜λŠ” 과정에 λŒ€ν•΄μ„œ 더 λ§Žμ€ 것듀 것 배울 수 μžˆμ—ˆλ˜ μ†Œμ€‘ν•œ κ²½ν—˜μ΄μ—ˆμŠ΅λ‹ˆλ‹€.