ALGORITHMIC
LINE ART
Explores algorithmic drawing system through custom-built plotters, robots and traditional artistic tools
Drawing Bot V1 / DBV1 — The First Working System
Back in 2020, I became deeply interested in plotters, drawing machines, and robots capable of constructing an image not as a conventional printer would, but as a mechanical mediator between the artist, the algorithm, and the physical material. What attracted me was not automation for its own sake, but the possibility of creating a device that could translate a digital idea into real physical movement — through a pen, a marker, a brush, a pencil, the resistance of paper, friction, delay, error, and chance.
I began to see the plotter not as a technical tool, but as an extension of the artistic gesture. The machine was meant to become a mediator: not to replace the artist, but to expand his physical capabilities.
In 2022, I built my first plotter — Drawing Bot V1, or DBV1 for short. It was an experimental machine with a working area of 300 × 400 mm, created to study the basic principles of coordinate motion, pen control, and the execution of small-scale artistic experiments.
I began to see the plotter not as a technical tool, but as an extension of the artistic gesture. The machine was meant to become a mediator: not to replace the artist, but to expand his physical capabilities.
In 2022, I built my first plotter — Drawing Bot V1, or DBV1 for short. It was an experimental machine with a working area of 300 × 400 mm, created to study the basic principles of coordinate motion, pen control, and the execution of small-scale artistic experiments.
The first version was based on an Arduino board, a CNC Shield, and GRBL firmware. It was connected to the computer via USB, and GRBL-Plotter served as the control software. This program made it possible to control movement, work with G-code, manage multiple tools, insert auxiliary commands, and conduct my first experiments with automation.
The main advantage of DBV1 was its low cost and relative simplicity. It was a useful learning platform: open, understandable, flexible, and suitable for rapid experimentation. But its limitations became obvious very quickly.
The main advantage of DBV1 was its low cost and relative simplicity. It was a useful learning platform: open, understandable, flexible, and suitable for rapid experimentation. But its limitations became obvious very quickly.
The USB connection turned out to be the weak link. With long files, complex trajectories, and extended execution times, the risk of stopping, losing connection, freezing the computer, or encountering data transmission errors increased significantly. For small tests this was acceptable, but for serious work it was not. Already at this stage, I understood that any plotter dependent on a constant connection to a computer via USB or a parallel port could not be reliable enough for complex projects where G-code might run not for hours, but for days.
The automatic pen-changing system was especially problematic. The mechanics were unstable, repeatability was insufficient, and even a small error could ruin the entire work. DBV1 gave me the most important insight: an artistic machine must not only be precise — it must be durable. It must be able to work for long periods of time, without pauses, without constant supervision, and without the possibility of a random failure.
The automatic pen-changing system was especially problematic. The mechanics were unstable, repeatability was insufficient, and even a small error could ruin the entire work. DBV1 gave me the most important insight: an artistic machine must not only be precise — it must be durable. It must be able to work for long periods of time, without pauses, without constant supervision, and without the possibility of a random failure.
Drawing Bot V1.1 / DBV1.1 — The Shift to an Offline Controller
After the first experiments, I abandoned the idea of a plotter that was fully dependent on a computer. The next step was to move to an offline industrial CNC controller. In my case, this was the DDCS V3.1 — a controller that had proven itself in affordable CNC systems and allowed G-code to be executed directly, without continuously streaming commands from a computer.
This is how the transitional version, Drawing Bot V1.1, appeared — a small-format plotter based on an offline controller. In essence, it was not so much a new artistic machine as a technical bridge between a laboratory prototype and a future large-format system.
The main purpose of DBV1.1 was to test the stability of offline control. Now the G-code was loaded directly into the controller, and the machine could operate independently from the computer. This changed the philosophy of the process itself. The plotter stopped being a peripheral device and became an autonomous execution machine.
For me, this was a decisive turning point. I realized that if I wanted to create large, complex, long-duration works, I needed an architecture closer to an industrial machine than to hobby electronics.
The main purpose of DBV1.1 was to test the stability of offline control. Now the G-code was loaded directly into the controller, and the machine could operate independently from the computer. This changed the philosophy of the process itself. The plotter stopped being a peripheral device and became an autonomous execution machine.
For me, this was a decisive turning point. I realized that if I wanted to create large, complex, long-duration works, I needed an architecture closer to an industrial machine than to hobby electronics.
Drawing Bot V2 / DBV2 — The First Large-Format Plotter
In 2023, already having experience with an offline controller, I built a large-format model — Drawing Bot V2, or DBV2. The working area of this machine was 1700 × 1000 mm.
The structure was assembled from accessible and widely available components: aluminum construction profiles, NEMA 17 stepper motors, belt drives, carriages, and an offline CNC controller. This was no longer a tabletop experimental machine, but a fully functional large-format system capable of working on a large sheet.
On DBV2, I was able to carry out one of the key technical tests: an isometric projection of a view of New York measuring 1260 × 726 mm. The work was executed with a high density of lines and a pen movement resolution of 0.1 mm. The final G-code contained more than 32.6 million lines.
The structure was assembled from accessible and widely available components: aluminum construction profiles, NEMA 17 stepper motors, belt drives, carriages, and an offline CNC controller. This was no longer a tabletop experimental machine, but a fully functional large-format system capable of working on a large sheet.
On DBV2, I was able to carry out one of the key technical tests: an isometric projection of a view of New York measuring 1260 × 726 mm. The work was executed with a high density of lines and a pen movement resolution of 0.1 mm. The final G-code contained more than 32.6 million lines.
The execution took 21 days of continuous operation at a speed of F8000. The controller worked for three weeks without stopping and did not produce a single failure.
For me, this was an important moment. The machine passed the test of time. Not one hour, not one day, not a short demonstration, but twenty-one days of uninterrupted movement. Millions of commands. Constant acceleration, braking, directional changes, and a pen trajectory stretching for kilometers. It was then that I became convinced that the offline architecture had been the right decision.
But the same test also revealed the weak points of the mechanics.
For me, this was an important moment. The machine passed the test of time. Not one hour, not one day, not a short demonstration, but twenty-one days of uninterrupted movement. Millions of commands. Constant acceleration, braking, directional changes, and a pen trajectory stretching for kilometers. It was then that I became convinced that the offline architecture had been the right decision.
But the same test also revealed the weak points of the mechanics.
Due to the bulkiness of the structure and the type of kinematics often found in inexpensive 3D printers, parasitic vibration appeared when the drawing head decelerated sharply. Its amplitude could reach 1–2 mm, creating visible defects in certain areas of the image. The machine was capable of completing the work, but it did not yet possess the rigidity and dynamic stability required for a high-level artistic result.
Another problem was the rubber wheels that moved along the aluminum profile. They were not designed for continuous multi-week motion cycles. Over time, the rollers wore out, lost their geometry, developed play, and reduced the repeatability of movement.
DBV2 proved that a large-format autonomous drawing machine was possible. But it also showed that for true precision, electronics and correct G-code are not enough. The mechanics must be as serious as the idea itself.
Another problem was the rubber wheels that moved along the aluminum profile. They were not designed for continuous multi-week motion cycles. Over time, the rollers wore out, lost their geometry, developed play, and reduced the repeatability of movement.
DBV2 proved that a large-format autonomous drawing machine was possible. But it also showed that for true precision, electronics and correct G-code are not enough. The mechanics must be as serious as the idea itself.
Drawing Bot V3 / DBV3 - The Fight Against Vibration and Inertia
In 2024, I completely redesigned the drawing head and the motion assemblies. The main goal was to reduce the mass of the moving parts, lower inertia, and eliminate the parasitic oscillations that had appeared on DBV2.
I abandoned the short-lived rollers that moved along the aluminum profile and switched to MGN12 linear rails with sliding carriages. This is how the next version appeared — Drawing Bot V3, or DBV3.
I abandoned the short-lived rollers that moved along the aluminum profile and switched to MGN12 linear rails with sliding carriages. This is how the next version appeared — Drawing Bot V3, or DBV3.
The transition to linear rails radically changed the behavior of the machine. Movement became more composed, rigid, and predictable. The defects caused by wheel wear and micro-play disappeared. Reducing the weight of the drawing head lowered inertia during acceleration and braking. The plotter was no longer simply moving a tool through coordinates — it was doing so with far greater mechanical discipline.
DBV3 was an important stage in the maturation of the system. If DBV2 was proof of scale, then DBV3 was proof of control.
DBV3 was an important stage in the maturation of the system. If DBV2 was proof of scale, then DBV3 was proof of control.
However, DBV2 and DBV3 still had one significant limitation: they worked with a fixed tool. For multicolor works, I had to manually change the pen, marker, brush, or other instrument. This interrupted the continuity of the process, increased the risk of error, and turned complex multicolor work into a sequence of manual interventions.
The next task was obvious: the machine had to learn how to change tools by itself.
The next task was obvious: the machine had to learn how to change tools by itself.
Drawing Bot V4 / DBV4 — Automatic Tool Changing
In 2025, I built Drawing Bot V4, or DBV4 — a plotter with an automatic tool-changing system. This version became the most complex stage in the entire evolution of the project.
The working area of DBV4 is 1050 × 1050 mm. The machine uses an offline four-axis controller and a tool-changing system for 31 instruments. Now the plotter can work not only with a single pen, but with an entire set of pens, markers, brushes, and other tools, switching between them according to a predefined logic within the G-code.
This changed the very principle of image construction. Color, line, tool, and sequence of application became part of the algorithmic composition. The machine gained the ability not merely to draw a trajectory, but to perform a complex score: one instrument enters, another falls silent, a third returns after several hours, and a fourth is used only for a brief accent.
The working area of DBV4 is 1050 × 1050 mm. The machine uses an offline four-axis controller and a tool-changing system for 31 instruments. Now the plotter can work not only with a single pen, but with an entire set of pens, markers, brushes, and other tools, switching between them according to a predefined logic within the G-code.
This changed the very principle of image construction. Color, line, tool, and sequence of application became part of the algorithmic composition. The machine gained the ability not merely to draw a trajectory, but to perform a complex score: one instrument enters, another falls silent, a third returns after several hours, and a fourth is used only for a brief accent.
To prepare G-code for DBV4, I built a custom workflow. It includes image preparation, separation into color and tool layers, trajectory generation, tool assignment, optimization of the drawing order, insertion of tool-change commands, parking commands, tool lift and lowering commands, as well as additional service commands for reliable long-duration operation.
Throughout the first half of 2026, I refined the moving assemblies, tested the repeatability of tool changes, and explored different approaches to preventing working instruments from drying out during long waiting periods. This turned out to be a separate engineering challenge: in a multicolor work, some markers or pens may wait for their turn for hours, and sometimes for days. Therefore, the system for storing, parking, and protecting the tips became no less important than the precision of coordinate movement.
Throughout the first half of 2026, I refined the moving assemblies, tested the repeatability of tool changes, and explored different approaches to preventing working instruments from drying out during long waiting periods. This turned out to be a separate engineering challenge: in a multicolor work, some markers or pens may wait for their turn for hours, and sometimes for days. Therefore, the system for storing, parking, and protecting the tips became no less important than the precision of coordinate movement.
Works using more than 60 colors were created on DBV4. This became possible through the combination of automatic tool changing, a carefully planned sequence of passes, and the preparation of color layers in advance. The machine ceased to be just a plotter. It became a complex drawing system in which mechanics, code, color, and physical material operate as a single organism.
Conclusion
The path from DBV1 to DBV4 was not merely a technical evolution for me. It was a journey from a simple homemade device to an autonomous artistic machine capable of executing complex multicolor works over extended periods of time.
Each version solved a specific problem left by the previous one.
DBV1 taught me the fundamentals and revealed the limits of a USB-dependent system.
DBV1.1 proved the necessity of an offline controller.
DBV2 opened the possibility of large-scale format and multi-week continuous operation.
DBV3 provided mechanical stability, rigidity, and precision.
DBV4 added automatic tool changing and transformed the plotter into a full execution system for complex algorithmic works.
For me, Drawing Bot is not a machine that draws instead of the artist. It is a tool for extending the artistic body. It is a mechanical mediator between idea and matter, between digital code and physical line, between control and unpredictability.
It is precisely within this tension — between algorithm, tool, error, time, and movement — that my artistic method is born.
Each version solved a specific problem left by the previous one.
DBV1 taught me the fundamentals and revealed the limits of a USB-dependent system.
DBV1.1 proved the necessity of an offline controller.
DBV2 opened the possibility of large-scale format and multi-week continuous operation.
DBV3 provided mechanical stability, rigidity, and precision.
DBV4 added automatic tool changing and transformed the plotter into a full execution system for complex algorithmic works.
For me, Drawing Bot is not a machine that draws instead of the artist. It is a tool for extending the artistic body. It is a mechanical mediator between idea and matter, between digital code and physical line, between control and unpredictability.
It is precisely within this tension — between algorithm, tool, error, time, and movement — that my artistic method is born.
