There is a point where a breadboard prototype stops feeling temporary and starts feeling unnecessarily cramped. For me, that point usually arrived as soon as I added an ESP32.

Most ESP32 development boards are wide enough to consume nearly all the useful space across a standard breadboard. Once the board is installed, only a single row of connection points may remain accessible on either side. The ESP32 technically fits, but every jumper wire becomes a small routing problem and the breadboard has very little room left for the circuit it was supposed to hold.

The usual workaround is to bridge the ESP32 across two breadboards. That restores access to the pins, but it also doubles the footprint and allows the development board to dictate the shape of the entire prototype. What began as a simple test circuit quickly becomes a wide collection of breadboards, jumper wires, and loosely positioned modules.

After fighting with that arrangement one too many times, I started with a different question: **what if the ESP32 did not need to occupy the breadboard at all?**

::card[gallery] {"title":"ESP32 DIN-rail prototyping system","images":[{"src":"/assets/blog/esp32-din-prototyping-breadboard.webp","alt":"Close-up of an ESP32 mounted beside a breadboard on a blue DIN-rail carrier, with USB and Dupont wiring connected.","aspectRatio":"1312/1024"},{"src":"/assets/blog/esp32-din-prototyping-application.webp","alt":"DIN-mounted ESP32 and prototyping board connected to an illuminated geared mechanism beside a foot pedal.","aspectRatio":"1600/898"},{"src":"/assets/blog/esp32-din-prototyping-system.webp","alt":"Vertical DIN-rail prototype combining a breadboard, blue breakout mounts, and two Ethernet ports.","aspectRatio":"800/1200"},{"src":"/assets/blog/esp32-din-prototyping-wiring.webp","alt":"Close-up of the blue ESP32 breakout carriers with Dupont wires routed beneath the pin headers.","aspectRatio":"1600/1067"},{"src":"/assets/blog/esp32-din-prototyping-hero.webp","alt":"Blue ESP32 pin-breakout carriers on a DIN rail, showing the exposed sockets and bundled Dupont wiring.","aspectRatio":"1536/1024"}]}{placement=inline}

## The breadboard is carrying the wrong thing

A breadboard is most useful when its rows are available for the circuit being explored. Resistors, sensors, level shifters, driver components, and temporary connections all need somewhere to go. A wide development board consumes that space before the experiment has really started.

Using two breadboards solves the electrical-access problem, but not the physical-layout problem. The controller remains in the middle, USB access competes with jumper wires, and every additional module extends the arrangement in another direction. It works, but it becomes difficult to move, revisit, or keep assembled.

The problem was not the ESP32 or the breadboard. It was asking one surface to hold both the controller and the circuit.

## Move the controller, not the circuit

The [ESP32 DIN-mount prototyping system](https://www.printables.com/model/1825235-esp32-breadboard-prototyping-din-mount-more) moves the development board off the breadboard and mounts it alongside the prototyping area. Dupont connectors can be fitted only to the ESP32 pins the circuit actually uses, then routed to a separately mounted breadboard or prototyping board.

![CAD preview of the two-piece ESP32 DIN-rail mount and pin-breakout carriers.](/assets/blog/esp32-din-prototyping-mount-render.webp)

This separates two jobs that are normally forced into the same space. The printed mount holds the controller and keeps its headers accessible; the breadboard is free to hold the circuit. The ESP32 can stay connected over USB while the experimental wiring remains visible and replaceable.

The mount also turns the controller into one module within the prototype rather than the object everything else must be arranged around.

## A rail instead of a pile

DIN rail is useful here because it gives otherwise temporary parts a shared physical reference. The pieces can still be moved, replaced, or expanded, but they no longer need to balance on the desk as a loose chain of breadboards and cables.

That makes the system useful in the space between a quick experiment and a finished enclosure. A circuit can remain easy to change while becoming structured enough to move, store, demonstrate, or continue developing later.

## What this changes

Moving the ESP32 away from the breadboard provides a few practical advantages:

- The ESP32 no longer consumes most of the breadboard's usable rows.
- USB, buttons, headers, and connected pins remain accessible.
- Breadboards and modules can be arranged around the circuit instead of around the controller.
- Larger prototypes can grow by adding DIN-mounted modules rather than rebuilding the entire layout.
- A successful experiment can remain assembled without becoming a loose pile of boards and jumper wires.

The intention is not to make breadboarding permanent or turn every experiment into an electrical cabinet. It is to preserve the flexibility of breadboard prototyping while giving the physical arrangement a little more structure.

## The initial system

The first release includes mounts for a 26 mm-wide, 30-pin ESP32 development board, breadboards, 20 × 40 and 20 × 80 prototyping boards, and Ethernet ports. Support for other development-board formats can follow the same modular approach without changing the rest of the rail.

Together, these parts create a foundation that can expand as a project moves from a quick experiment toward something more substantial. The breadboard remains temporary. The layout no longer has to be accidental.

## Model files

The printable parts, complete gallery, and current model revisions are available on [Printables](https://www.printables.com/model/1825235-esp32-breadboard-prototyping-din-mount-more).

A modular 3D-printed DIN-rail system that moves an ESP32 beside the breadboard, preserves prototyping space, and organizes larger electronics experiments.

ESP32 DIN-Rail Breadboard Prototyping System

Taking the ESP32 Off the Breadboard

micropython-ota-updater

Update MicroPython application files on an ESP32 from a GitHub branch using commit SHAs, staged downloads, startup confirmation, and rollback.

MicroPython OTA Updates from GitHub on ESP32

# How I Designed a 3D-Printed Laminar Flow Nozzle

::card[feature] {"eyebrow":"Featured on Hackaday","description":"An earlier iteration of this project was featured in “Garden Hose Gets Laminar Flow” in June 2025.","url":"https://hackaday.com/2025/06/08/garden-hose-gets-laminar-flow/","linkLabel":"Read Garden Hose Gets Laminar Flow on Hackaday","secondaryLabel":"Also featured","secondaryTitle":"Matt’s “PrusaPrinters Picks” · Prusa Live #26 · 13:05","secondaryUrl":"https://www.youtube.com/live/HgwusaEWNuk?t=785","secondaryLinkLabel":"Watch Matt’s PrusaPrinters Picks feature on Prusa Live #26 at 13:05"}{placement=inline}

There is something mildly unreasonable about connecting an ordinary garden hose to a printed part and expecting the water to leave as a clear, smooth rod.

::card[gallery] {"title":"Laminar-flow nozzle stream tests","aspectRatio":"16/9","images":[{"src":"/assets/projects/laminar-flow-intro-stream.webp","fullSrc":"/assets/projects/laminar-flow-intro-stream-full.webp","thumbSrc":"/assets/projects/laminar-flow-intro-stream-thumb.webp","alt":"Mk2.2 3D-printed laminar-flow nozzle producing a smooth water stream outdoors.","aspectRatio":"1600 / 884"},{"src":"/assets/projects/laminar-flow-mk2.jpg","fullSrc":"/assets/projects/laminar-flow-mk2.jpg","alt":"Mk2 3D-printed laminar-flow garden hose head.","aspectRatio":"1200 / 630"}]}{placement=inline}

That was the attraction. A hose normally gives you a lively, turbulent stream. I wanted to see how much of that turbulence could be negotiated away with geometry that came directly off an FDM printer.

The result became three generations of garden-hose laminar-flow nozzles. Mk1 proved that the idea could work, but relied on packed diffuser material and leaked more than I wanted. Mk2 moved much more of the flow conditioning into the print. Mk3 separated the mechanism into swappable stages so I could test the diffuser, straightener, exit chamber, and outlet lens independently.

## Two key lightbulb moments

Two ideas eventually changed how I approached every later revision: work with the texture produced by the printer, and pay close attention to pressure stability. A key to maintaining laminar flow is stable, evenly distributed pressure; using more straws creates additional surface area that helps dampen disturbances and keep the flow consistent.

### Print with the flow

FDM passage walls are not perfectly smooth. When water crosses the rounded edges of successive printed layers, each ridge is another opportunity to disturb the flow near the wall. Orienting the straightener so its dominant layer texture runs downstream reduces those cross-stream interruptions. It does not remove surface roughness; it turns that roughness in the less disruptive direction.

![Good and bad FDM layer orientations for water flow through a printed straightener](/assets/projects/laminar-flow-layer-lines.svg)

A squeezed hexagonal lattice made that orientation practical. It creates many long, narrow, self-supporting passages as repeatable printed geometry, replacing the bundle of drinking straws traditionally used as a flow straightener.

### Normalize pressure instead of adding length

The early Mk2 body treated extra length as extra settling space. In practice, an elongated body could preserve an uneven axial pressure path all the way to the outlet. From Mk2.2 onward, the design became deliberately “fat”: a shorter, wider chamber gives the flow a broader region in which to equalize pressure before it reaches the lens.

![Conceptual drawing of a 3D-printed laminar flow nozzle showing the entry chamber, diffuser, straightener, exit chamber, and outlet lens](/assets/projects/laminar-flow-nozzle-stages.svg)

1. Entry chamber — The hose enters at 90 degrees instead of pointing directly at the outlet. The chamber gives the incoming water room to settle before it reaches the conditioning stages.
2. Diffuser — A printed cross-hatch interrupts and dampens the larger turbulent structures created by the hose, fitting, and sharp inlet transition.
3. Straightener — Many narrow, parallel passages encourage the water to move in the outlet direction. This part is printed on its side so the layer lines run parallel to the flow.
4. Exit chamber — A short, wide volume lets the conditioned streams meet again while encouraging more uniform pressure across the outlet.
5. Lens — The interchangeable nozzle insert controls the final diameter, edge quality, and shape of the water column.

In theory the strategy is to calm the incoming water, divide it into smaller channels, give those channels enough length to straighten, bring them back together without reintroducing a large pressure difference, and then pass the combined flow through a clean [[stage:lens]] outlet.

That last part matters more than it first appears. A polished, accurately printed lens can preserve the conditioned flow. A rough edge, damaged opening, or badly sealed insert can undo much of the work performed upstream.

This article is the design and printing guide behind those models. If you only want the files, the [3D Printable - Laminar Flow project page](/project/3d-printable-laminar-flow) collects all three versions in one place.

::project[3d-printable-laminar-flow]{placement=gutter}

## Behold: the Mk3

![Completed Mk3 3D-printed laminar-flow nozzle](/assets/projects/laminar-flow-mk3-final-result.webp)

The best versions produce the familiar glass-rod effect: a stream that looks almost motionless until it breaks up farther from the outlet.

::embed[3D-printable laminar flow garden hose nozzle Mk3 demonstration](https://www.youtube-nocookie.com/embed/zeuPyQmIuW4)

## What is laminar flow?

::card[concept] {"eyebrow":"FLUID DYNAMICS","title":"Reynolds number","intro":"One common way to reason about the transition is the Reynolds number:","formula":"Re = ρvD / μ","body":"Density, velocity, characteristic diameter, and viscosity all matter.","threshold":"Re < 2300","caption":"Working marker for laminar flow inside a pipe","footer":["That number is useful here, but it is not a magic certification for the complete nozzle or the free jet after the water leaves it.","I use it as a design guide for the straightener channels."]}{placement=gutter}

Laminar flow describes fluid moving in orderly layers rather than mixing through chaotic eddies.

Splitting one large, fast stream into many smaller passages reduces the characteristic diameter and makes orderly flow easier to support. The visual test is then whether the streams can recombine and leave the lens with a stable profile.

So when I call this a laminar-flow nozzle, I am describing both the internal design goal and the stable glass-rod stream it produces.

## Mk1: proving a printable hose nozzle could work

::card[printables](https://www.printables.com/model/64165-laminar-flow-garden-hose-head-mk1){placement=gutter}

Mk1 was the useful kind of first prototype: good enough to prove the idea and awkward enough to make the next version obvious. It used a small piece of steel wool or Brillo-style material between the hose attachment and the middle section. That packed material acted as a diffuser and helped equalize pressure before the water entered the straightening channels.

![Mk1 3D-printed laminar flow garden hose head](/assets/projects/laminar-flow-mk1.jpg)

The output could become impressively smooth at gentle flow rates, but the nozzle had several limitations:

- The body and joints could leak.
- The stream was not always uniform across the outlet.
- Loose diffuser material made assembly and repetition less predictable.
- More of the final result depended on finding and packing non-printed material correctly.

The model includes STEP files, so it still has value as a starting point for remixes. For my own progression, its main contribution was proving that a garden-hose attachment could produce a stable enough stream to justify another round of design.

## Mk2: moving the work into the print

::card[printables](https://www.printables.com/model/1315244-laminar-flow-garden-hose-head-mk2){placement=gutter}

Five years later, Mk2 revisited the same problem with more modelling and printing experience. The central question changed from “Can this work?” to “How much of the mechanism can the printer create for me?”

Mk2 uses a print-in-place body with a settling region, printed diffuser geometry, a straightener, and interchangeable outlet lenses. The straightener applies the first epiphany with a squeezed hexagonal lattice in place of a bundle of loose tubes. The hexagons self-support at printable angles, and the channels provide a length-to-diameter ratio greater than 40:1.

Printing the body sideways makes that geometry practical. It avoids a tall, narrow print and aligns the straightener's dominant layer texture with the water instead of presenting each bead edge across the channel. The included 6 mm, 7 mm, and 8 mm lenses make the outlet easy to change without redesigning the complete body.

Mk2 also exposed the pressure problem. Its early body was long and narrow, the water entered along the same axis as the outlet, and the narrowing exit could preserve an uneven pressure profile through the middle of the nozzle. Those observations led directly to the shorter, wider Mk2.2 and became part of the design brief for Mk3.

### Why Mk2.2 may be the best first print

The later Mk2.2 “fat” revision applies the pressure-normalization epiphany in a smaller, simpler body:

- A shorter, wider form encourages more uniform pressure.
- The hexagonal body reduces difficult overhangs.
- The diffuser is included in the print.
- No screws are required.
- The footprint remains smaller than Mk3.
- Initial testing produced a much cleaner stream than the earlier Mk2 revisions.

If your goal is to print a working garden-hose laminar-flow nozzle rather than to experiment with every internal stage, Mk2.2 is the version I would start with.

## Mk3: making every stage testable

::card[printables](https://www.printables.com/model/1319709-laminar-flow-garden-hose-head-mk3){placement=gutter}

Mk3 trades some simplicity for control. The body uses a clamshell arrangement so the diffuser and straightener can be printed in deliberate orientations, removed, replaced, and compared. Water enters from the side, passes through the diffuser and straightener, expands into a wider exit chamber, and leaves through a push-fit lens.

![Mk3 stator-style flow core and internal channels](/assets/projects/laminar-flow-mk3-side.webp)

The discrete stages make Mk3 a better platform for experiments:

- Diffuser patterns can be changed without reprinting the complete body.
- Straightener channels can be printed with continuous layer lines parallel to the water.
- Lenses can be swapped while the hose is still running.
- Different lens diameters, materials, surface finishes, and shapes can be compared quickly.
- The larger exit chamber makes it easier to pursue uniform pressure before the outlet.

The tradeoff is a larger print and more assembly. The current design uses a printed base and cap, separate diffuser and straightener parts, a lens holder, a TPU lens gasket, and M3 hardware. Mk3 is the version I prefer for understanding and modifying the mechanism; Mk2.2 is the more accessible recommendation when you mainly want the finished stream.

## Which version should you print?

### Mk1: the original experiment

Print Mk1 if you want a simpler model to remix or the history of how the series began. Expect loose diffuser material and more attention to sealing.

### Mk2.2: the approachable first print

Print Mk2.2 if you want a smaller footprint, no screws, printed flow-conditioning geometry, and a strong visual result.

### Mk3: the modular test platform

Print Mk3 if you want independently printable stages, swappable internal geometry, and hot-swappable outlet lenses.

All three models are available from the [laminar-flow project hub](/project/3d-printable-laminar-flow), along with the image galleries and direct Printables downloads.

## Printing notes

The supplied Mk2 and Mk3 print files were developed around a 0.4 mm nozzle and 0.25 mm layers. PETG is my default recommendation for an outdoor hose attachment, and it is the material used for the prototypes described on Printables.

Several details matter more than chasing an unusually small layer height:

- Verify that the thin walls in the hexagonal or straightener pattern actually appear in the slicer. With a different nozzle width, some channels may disappear completely.
- Print the straightener in the supplied orientation. On Mk3, the part is placed so its layer lines run parallel to the direction of the water.
- Increase bridging speed if the slicer predicts an extremely long Mk2 print. The early settling region uses many bridges, and minor cosmetic artifacts there are less important than preserving the internal structure.
- Use enough perimeters and infill to make the hose connection mechanically strong and water resistant.
- Inspect and clean the outlet lens carefully. A small string, burr, or damaged edge can disturb the stream.
- Test for leaks at low pressure before opening the hose farther.

Mk2 needs a 3/8-inch O-ring at the Gardena-style connector. Mk3 uses a TPU gasket around the lens system and M3 hardware to join the clamshell and lens holder. A thin application of silicone sealant around the lens interface can help if water escapes around the insert instead of through it.

## Troubleshooting an uneven stream

If the stream breaks up immediately, start at the outlet. Check the lens for a rough inner edge, debris, print strings, distortion, or an uneven seal.

If the center of the stream behaves differently from the edges, reduce the flow rate and look for a pressure imbalance. A direct axial inlet, small exit chamber, blocked straightener channels, or uneven diffuser can preserve a higher-pressure path through the middle.

If the nozzle only behaves at very low flow, the straightener passages may be too large or too short for the supplied flow rate. Reducing the lens diameter may make the visual result worse by increasing velocity, even though the smaller opening looks cleaner on the workbench.

## Up Next: Can we FDM print a laminar-flow hood? 🤔

A laminar-flow hood applies similar flow-conditioning ideas to air instead of water: diffuse an uneven incoming flow, equalize pressure, and produce a smooth, uniform stream. Could an FDM-printed plenum, filter housing, and airflow straightener work with a suitably sized blower and HEPA filter? The real challenge would be controlling leaks and turbulence, maintaining even pressure across the filter, and verifying airflow and particle performance.

## Related printed water projects

The [Modular Hydroponic Tower Garden](/project/modular-hydroponic-tower-garden) and [Infinity Bowl](/project/the-infinity-bowl) approach water from different directions, but they share the same practical lesson: printed geometry is only part of the system. Fittings, seals, pressure, maintenance, print orientation, and revision history matter just as much as the shape on screen.

The complete laminar-flow series is available here:

- [Mk1 on Printables](https://www.printables.com/model/64165-laminar-flow-garden-hose-head-mk1)
- [Mk2 and Mk2.2 on Printables](https://www.printables.com/model/1315244-laminar-flow-garden-hose-head-mk2)
- [Mk3 on Printables](https://www.printables.com/model/1319709-laminar-flow-garden-hose-head-mk3)

For the consolidated project history, galleries, and links, visit [3D Printable - Laminar Flow](/project/3d-printable-laminar-flow).

3D Printable – Laminar Flow

A practical look at three 3D-printed garden-hose laminar-flow nozzles: design stages, printing lessons, failures, and free model files.

3D-Printed Laminar Flow Nozzle: Design Lessons

How I Designed a 3D-Printed Laminar Flow Nozzle

# Models section is now live

The models section is now live. It pulls together the 3D-printable things I have put on Printables over the years: hydroponic parts, hose experiments, pegboard bits, brackets, adapters, mounts, small household fixes, and the occasional odd object that exists because I wanted to see if the printer could make the problem go away.

That is the nice thing about small printed parts. They are usually not grand inventions. They are little arguments with the physical world: a hose thread that needs one more fitting, a controller that should have had a wall mount, a bucket system that almost works, a pet station that turns into plumbing, sensors, drawers, and too many printed revisions.

A few of the more popular ones are below. The full list is now on the [models page](/models).

::card[printables](https://www.printables.com/@smysnk)

## 50mm / 2in Hydroponic Net Cup

::card[printables](https://www.printables.com/model/64138-50mm-2in-hydroponic-net-cup)

A simple hydroponic net cup, which is probably why it has done well. It solves one job, fits into a larger grow setup, and does not ask the printer to be clever for no reason.

---

## Modular Hydroponic Tower Garden

::card[printables](https://www.printables.com/model/64154-modular-hydroponic-tower-garden)

This is the bigger hydroponics project: modular tower sections, plant-cup offshoots, a center-routed hose path, and normal garden hardware. It is one of those builds where the printed part is only half the story; the other half is water taking every opportunity to remind you it has opinions.

---

## Laminar Flow Garden Hose Head

::card[printables](https://www.printables.com/model/1315244-laminar-flow-garden-hose-head-mk2)

The laminar-flow hose heads exist for a very serious engineering reason: smooth water looks cool. Mk1 used packed diffuser material, Mk2 moved more of the flow conditioning into the print, and Mk3 pushes toward a cleaner stator-style core.

---

## The Infinity Bowl

::card[printables](https://www.printables.com/model/1277888-the-infinity-bowl)

The Infinity Bowl is the overbuilt one: an in-wall pet feeding and hydration station with printed parts, plumbing, sensors, a drawer mechanism, and Arduino-oriented control. It started as a practical pet project and turned into a small mechanical ecosystem.

---

## Borg Cube

::card[printables](https://www.printables.com/model/98618-borg-cube)

Borg Cube is on the other end of the spectrum: less practical fixture, more object. It is there because not every model needs to justify itself as a tool. Sometimes the printer is allowed to make a cube with personality.

---

## Ratcheting Toothpaste Squeezer

::card[printables](https://www.printables.com/model/64133-ratcheting-toothpaste-squeezer)

The toothpaste squeezer is exactly what it sounds like: a small household mechanism for a tiny repeated annoyance. These are the kinds of models I like keeping around because they show the practical side of printing without making the whole thing feel precious.

The full collection is on the [models page](/models), with sorting by likes, makes, downloads, views, and date.

Models section is now live!

The projects section now has a real home for the tools, CAD builds, 3D-printable models, browser experiments, and old-code revivals I keep working on. The cards are the quick tour; the notes below add the context and links.

## m68k-interpreter

::project[m68k-interpreter]{placement=gutter}

The [browser-based M68K Interpreter](/m68k-interpreter) is my fork of Gianluca Rea's Motorola 68000 assembly IDE. It runs entirely in the browser and combines an editor, assembler, CPU execution, breakpoints, single-stepping, register inspection, memory views, and hardware panels in one workspace. The current build targets the MC68000 and MC68010 instruction sets and tracks 116 instruction forms, while preserving the Easy68K-compatible workflow that made the original project useful.

I extended the runtime to revive a Motorola 68000 assembly version of Nibbles that I wrote for a college course in 2007. The terminal path supports keyboard-driven programs, so you can [play Nibbles online](/nibbles), inspect the source as it runs, and switch back to the debugger without installing an emulator. The [interpreter help and compatibility guide](/m68k-interpreter/help) documents the supported CPU models, instructions, hardware devices, examples, and current limitations.

The [smysnk fork source](https://github.com/smysnk/m68k-interpreter) includes the IDE, assembler, interpreter, examples, and tests. It remains explicitly credited to [Gianluca Rea's original m68k-interpreter](https://github.com/gianlucarea/m68k-interpreter), which provides the foundation for this browser-based 68K development environment.
## Nibbles 68000

::project[nibbles-68000]{placement=gutter}

Nibbles 68000 is the old game itself: the original `nibbles.asm`, now wired into the browser interpreter with a report path so the revived version can stay playable.

---

## varcad.io

::project[varcad-io]{placement=gutter}

varcad.io is my browser workspace for parametric CAD: editing, OpenSCAD/JSCAD compilation, GitHub sync, STL output, and report review in one place.

---

## ascii.gallery

::project[ascii-gallery]{placement=gutter}

ascii.gallery is my ANSI/ASCII art viewer and import pipeline. It crawls old textmode artwork, validates and converts files, then renders them in a browser with scrolling playback.

---

## sikuli-go

::project[sikuli-go]{placement=gutter}

sikuli-go is my Go layer for Sikuli-style desktop automation: image matching, OCR, clicks, typing, and reporting for tests where the screen is the API.

---

## autoresearch-jung

::project[autoresearch-jung]{placement=gutter}

autoresearch-jung is my Jungian fork of autoresearch. It keeps thesis/antithesis/synthesis notes, Codex prepare/reflect phases, Runpod runs, logs, telemetry, and Experiment Atlas close to the code that produced them.

---

## micropython-ota-updater

::project[micropython-ota-updater]{placement=gutter}

micropython-ota-updater adapts rdehuyss's updater for ESP32 projects that should pull from a GitHub branch at boot, update changed files, and leave local `env.py` secrets alone.

---

## test-station

::project[test-station]{placement=gutter}

test-station gives mixed-stack projects a consistent `report.json` and HTML report: coverage, ownership, failures, fixtures, and drilldown views in one place.

---

## gulp-rev-all

::project[gulp-rev-all]{placement=gutter}

gulp-rev-all is the Gulp plugin I maintain for cache-safe asset revisioning. It hashes files, rewrites references, emits manifests, and keeps parent filenames changing when child assets change.

---

## React Retro Screen

::project[react-retro-screen]{placement=gutter}

React Retro Screen is the reusable terminal-style renderer that came out of ascii.gallery: ANSI playback, editable screens, color modes, grid behavior, prompts, and input handling without bringing the whole gallery along.

---

## sikuli-framework

::project[sikuli-framework]{placement=gutter}

sikuli-framework is the Python/Robot Framework layer over sikuli-go, so tests can describe windows, controls, menus, and app behavior instead of repeating raw screenshot steps.

---

## jungstein.com

::project[jungstein-com]{placement=gutter}

jungstein.com is the writing and media side: essays, notes, and videos around science, psychology, cognition, consciousness, and the parts of those subjects I keep circling back to.

---

## project-commander

::project[project-commander]{placement=gutter}

project-commander is my local control plane for repos, terminals, dev servers, logs, and remote Go agents. It exists because my development setup kept getting messy in the same repeatable ways.

---

## 3D Printable - Laminar Flow

::project[3d-printable-laminar-flow]{placement=gutter}

::card[printables](https://www.printables.com/model/64165-laminar-flow-garden-hose-head-mk1){placement=gutter}

::card[printables](https://www.printables.com/model/1315244-laminar-flow-garden-hose-head-mk2){placement=gutter}

::card[printables](https://www.printables.com/model/1319709-laminar-flow-garden-hose-head-mk3){placement=gutter}

A three-version garden-hose laminar-flow nozzle project. Mk1 used packed diffuser material, Mk2 moved more flow conditioning into the print, and Mk3 uses a stator-style core with cleaner internal channels.

---

## The Infinity Bowl

::project[the-infinity-bowl]{placement=gutter}

::card[printables](https://www.printables.com/model/1277888-the-infinity-bowl){placement=gutter}

The Infinity Bowl is my in-wall pet feeding and hydration build: printed parts, water and drain plumbing, ultrasonic sensing, solenoid flow control, LED feedback, and a pull-out service drawer.

---

## Modular Hydroponic Tower Garden

::project[modular-hydroponic-tower-garden]{placement=gutter}

::card[printables](https://www.printables.com/model/64154-modular-hydroponic-tower-garden){placement=gutter}

A modular hydroponic tower model reworked around smaller print beds, capped plant-cup offshoots, a center-routed hose path, and normal garden hardware.

Explore a browser-based Motorola 68000 assembly interpreter, a playable 68K Nibbles game, developer tools, and 3D-printing projects by smysnk.

68K Interpreter and Browser Projects | smysnk

Projects: Browser Tools, Retro Computing, and 3D Printing

The fork keeps stepping, registers, memory inspection, and worker-backed execution while adding the terminal runtime path needed for keyboard-driven Nibbles play.

My fork of Gianluca Rea's browser-based Motorola 68000 interpreter, adapted so my old college Nibbles game can run in the browser.

GitHub

Test Report

Upstream

Hosted browser interpreter running Nibbles

Live demo

m68k-interpreter

It collects essays, notes, and videos around empirical science, depth psychology, cognitive models, consciousness, and the overlap between them.

The writing and media side of my work around science, psychology, cognition, consciousness, and where those subjects overlap.

Site

YouTube

jungstein.com

The build combines 3D-printed mechanics, water and drain plumbing, ultrasonic sensing, solenoid-controlled flow, LED faucet feedback, and Arduino-oriented control.

My in-wall pet feeding and hydration build with printed parts, plumbing, sensing, LED feedback, and a serviceable pull-out drawer.

Printables

Overview Video

Contest

The Infinity Bowl

Mk1 used packed diffuser material, Mk2 moved more flow conditioning into the print, and Mk3 uses cleaner internal channels with a stator-style core.

My three-version 3D-printable garden hose laminar-flow nozzle project, moving from packed diffuser material to a stator-style core.

Latest laminar flow garden hose head model

Printables Mk3

Second laminar flow garden hose head model

Printables Mk2

Original laminar flow garden hose head model

Printables Mk1

It wraps GUI automation in a Go runtime with Python and Node.js entry points, deterministic matching, OCR, and test-station reporting.

My Go take on Sikuli-style desktop automation with image matching, OCR, clicking, typing, local APIs, and reporting.

PyPI

Docs

sikuli-go

It lets tests describe windows, controls, menus, and application behavior instead of repeating raw screenshot steps everywhere.

The Python and Robot Framework layer I use on top of sikuli-go for higher-level GUI automation flows.

Engine

sikuli-framework

It ties together editing, sandboxed compilation, repository sync, and generated STL/report review so CAD iteration can stay in one browser workspace.

My browser workspace for parametric CAD projects, OpenSCAD/JSCAD compilation, GitHub sync, and generated STL/report review.

varcad.io

It packages color modes, grid behavior, touch input, prompts, editable retro screens, and xterm-checked playback into a reusable component.

A reusable React terminal-style screen component split out from ascii.gallery for ANSI playback, editable screens, prompts, and input handling.

Interactive component demos hosted on smysnk.com

Storybook

React Retro Screen

It normalizes results into `report.json` and renders reports with coverage, ownership, and drilldown views for mixed test suites.

My test runner and reporting tool for projects that mix validation stacks but still need one readable HTML report.

Demo

test-station

It checks the remote version on boot, downloads changed application files, and preserves device-local secrets in `env.py`.

My adaptation of rdehuyss's MicroPython OTA updater for ESP32 projects that should pull updates from a GitHub branch.

It combines Next.js, GraphQL, websocket log streaming, terminal sessions, and Go master/slave agents so local and remote dev machines are easier to manage.

A browser workspace for my too-many-repos setup, with project discovery, process control, logs, terminals, and remote Go agents.

MCP Plan

project-commander

I reworked the tower around smaller printer beds, capped plant-cup offshoots, foam-insert sizing, garden-hose hardware, and a center-fed water path.

My compact 3D-printed hydroponic tower model with smaller stackable segments, cup-retaining offshoots, and a center-routed hose path.

Modular Hydroponic Tower Garden model on Printables

Thingiverse

Modular Hydroponic Tower Garden

It crawls and imports old textmode files, converts and validates them, and renders them through a GraphQL-backed browser viewer with scrolling playback.

My ANSI and ASCII art gallery with crawling, import, conversion, validation, and browser playback for old textmode artwork.

ascii.gallery

The original `nibbles.asm` stays in the repo while the adapted m68k runtime and test-station report keep the browser-playable version easier to check.

My 2007 Motorola 68000 Nibbles game, wired up so the original assembly source can run again through the browser interpreter.

Interpreter

Play

Nibbles 68000

It hashes build output, rewrites asset references, emits manifests, and makes parent files change when the child assets they point at change.

An older Gulp plugin I maintain for cache-safe asset revisioning, reference rewriting, manifests, and parent hash propagation.

gulp-rev-all

It adds Codex prepare/reflect phases, Runpod execution, canonical logs, Experiment Atlas, telemetry, and a knowledge base around the training runs.

My Jungian fork of autoresearch, keeping thesis, antithesis, synthesis notes, and run telemetry beside the LLM training experiments.

Software projects by smysnk