{"id":4329,"date":"2023-02-05T17:37:00","date_gmt":"2023-02-05T22:37:00","guid":{"rendered":"https:\/\/jimsghost.place\/?p=4329"},"modified":"2024-12-28T18:02:02","modified_gmt":"2024-12-28T23:02:02","slug":"doomsun-general-assembly-en","status":"publish","type":"post","link":"https:\/\/jimsghost.place\/?p=4329","title":{"rendered":"DOOMSUN E32-FNC: General Assembly"},"content":{"rendered":"\n
\n

“The fragmentation of objects in the world is merely a trial,
designed for us to unite them as one.”<\/em><\/p>\n\n\n\n

— Satya-Sangama: The Essence of Unification<\/em><\/p>\n<\/div>\n\n\n\n\n\n\n\n

This will be a lengthy article, as this stage of the project involved assembling the entire mechanical structure of the machine \u2014 along with the accompanying basic electrical components. At first glance, it might seem like a straightforward task, given that by this point there was already a complete kit with all the necessary parts and fasteners. However, in reality, it turned out to be an endeavor where one could easily get bogged down for quite some time. Even painting the parts alone devoured almost three weeks.<\/p>\n\n\n\n

And then there was the challenge of not just screwing everything together \u2014 but managing to do so accurately and aligned \u2014 overcoming all the flaws inherently present in the design. For instance, to ensure that a circle wouldn\u2019t turn into an oval or that a rectangle wouldn\u2019t end up as a trapezoid or some kind of parallelogram \u2014 heaven forbid, let\u2019s not even speak of such horrors!<\/p>\n\n\n\n

\n\n\n\n
\n
\"\"<\/a><\/figure>\n<\/div>\n\n\n\n

Painting parts before assembly is my least favorite process due to its incredible tediousness and duration \u2014 regardless of its form. Be it painting walls in a room or detailing a model tank on a desk. Quite often, I manage to offload this part of the work onto my Wife. But not this time\u2026<\/p>\n\n\n\n

For this project, I\u2019ll lay out this personal nightmare as a sort of instruction manual, so that others, if the need arises, can fully experience the depths of despair, monotony, and hopelessness of this dismal routine.<\/p>\n\n\n\n

Considering that, in this case, the “original” paint could only symbolically be called that, the first step requires nothing more than a simple metal scraper. With it, the “paint” can almost entirely be scraped off the parts, resembling dried-up mucus.<\/p>\n\n\n\n

Next, the part is treated with sandpaper and a Scotch Brite “polishing sponge.” Among DIY enthusiasts, “Scotch-Brite” has become akin to the term “Xerox” \u2014 when a brand name turns into the name of a tool. It\u2019s a fantastic thing to always have stocked in the workshop. It works not only on metal but also on wood\u2026 and practically anything you\u2019d want to clean or polish:<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

All sorts of holes, nooks, and crannies are cleaned with acetone. The main thing is to avoid inhaling this stuff yourself. It\u2019s preferable to wait for good weather and perform all these operations outdoors.<\/p>\n\n\n\n

Afterward, the part is thoroughly washed in plain water with dish soap. A toothbrush and a regular green kitchen sponge will come in handy here. I have no brand preferences for these items. Whatever “shampoo” I can stealthily swipe from my wife\u2019s kitchen works fine. The same goes for the green kitchen sponge.<\/p>\n\n\n\n

\n

The theft is most easily blamed on the Cat later. That\u2019s his purpose in life. You can blame him for absolutely anything \u2014 from missing sponges to wood shavings all over the house. My Wife won\u2019t scold him because “he\u2019s got paws,” and he himself couldn\u2019t care less. The perfect scapegoat! Tested and proven many times.<\/p>\n<\/blockquote>\n\n\n\n

After several hours of drying, the part is rinsed in acetone once more. From this point on, it\u2019s considered unacceptable to touch it with bare hands. Gloves only \u2014 no exceptions! Human hands are greasy. They leave that grease on everything they touch. It doesn\u2019t matter what you think about this or how clean you believe your hands are. They\u2019re still greasy. And this will inevitably show during painting. Don\u2019t skimp on acetone and wipes!<\/p>\n\n\n\n

\n

As for the question, “Where did all the wipes in the house go again?” refer to the note above. By the way, don\u2019t blame the Dog. Unlike the Cat, he cares! He\u2019ll look at you with such sad reproach that you\u2019ll want to die of shame\u2026<\/p>\n<\/blockquote>\n\n\n\n

Aluminum is particularly tricky to paint due to its almost instantaneous surface oxidation. No matter how much you wash it with soap and acetone or protect it from greasy hands, everything will be in vain without a special primer. That\u2019s why a self-etching primer was used to prepare the surface:<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

This primer had already proven itself in my 3D printer<\/a> project. It still looks like new \u2014 no chips or peeling. This time, I also factored in the experience gained back then, making the process more deliberate.<\/p>\n\n\n\n

The instructions claim that you can sand and paint over this primer after 20\u201330 minutes. Probably true. But \u2014 through my own experience and the recommendations of other DIY enthusiasts \u2014 it\u2019s better to let this primer “set” on the part for at least a day. It\u2019s like giving it time to calmly “eat” the oxide beneath and fully bond with the fresh, tasty aluminum.<\/p>\n\n\n\n

I\u2019ve tried both the method suggested online and the one recommended in the instructions. Subjectively, there does seem to be a difference. So now I stick to the “day-long” method from the internet. As for the actual reasons\u2026 Humidity and temperature definitely play a role. In our local climate, it might genuinely be necessary to follow a procedure different from the one recommended by the manufacturer.<\/p>\n\n\n\n

What happens next depends on the circumstances. If I know the surface beneath the first layer of primer isn\u2019t great, I carefully sand the primer using blocks with 600->800->1000 grit sandpaper. If the surface was initially smooth or doesn\u2019t need to meet any specific “decorative” requirements, everything is lightly polished with the same Scotch-Brite sponge.<\/p>\n\n\n\n

A second layer of self-etching primer is applied, thicker this time, but not so much that it starts to drip. Another day for drying (in my case, again).<\/p>\n\n\n\n

After the second layer of primer, the final polish with a Scotch-Brite sponge is done, and the part is ready for paint. It\u2019s applied in two, three, or four thin coats, with intervals of half an hour to two hours. Generally, for each subsequent coat of paint, the drying time of the previous coat increases by half an hour. That is, 30 minutes after the first coat, an hour after the second, an hour and a half after the third, and so on.<\/p>\n\n\n\n

Of course, this depends on the type of paint, its brand, and specific properties, but I\u2019ve developed this general rule through experience, specifically for Krylon paints, which I use most often.<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

Whether to polish each layer or not depends on your preference. If you want a “mirror-like” finish, polish it. If you\u2019re after a “textured” look, don\u2019t polish it. The thinner and more even each coat is, the better the results.<\/p>\n\n\n\n

Oh, and one more thing\u2026 At all stages, protect the part from direct sunlight. This is crucial! I forgot about it again, and on a couple of parts, the paint “boiled” before it even began to dry\u2026 Although, why put it in quotes? It really does boil in our sunlight\u2026 In short, I had to strip everything and repaint. Which didn\u2019t make the process any more fun, not in the slightest\u2026<\/p>\n\n\n\n

\n\n\n\n

Finally, after receiving the complete set of painted parts and bidding a final farewell to my nervous system, I reached the much-anticipated assembly phase, which was meant to compensate for all the utter monotony of the painting stage.<\/p>\n\n\n\n

As always, there isn\u2019t much to say about the most enjoyable stage. The parts were simply put together and bolted into place:<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

Most of the fasteners were paired with split lock washers. Where split lock washers were unsuitable, the threads were secured with 222 Loctite (for areas requiring higher stability, 242 Loctite was used). When a nut was used, it was always a nylon lock nut. In theory, no vibrations should bother this machine now.<\/p>\n\n\n\n

\n

By the way, the color scheme for the parts was copied exactly from my favorite “Dusya” (see the first part of this series<\/a>). I have no regrets. It turned out to be a cheerful and neat device, rather than yet another black monstrosity.<\/p>\n<\/blockquote>\n\n\n\n

The initial assembly went smoothly, and it was time to move on to the next steps.<\/p>\n\n\n\n

\n\n\n\n

Ideally, this stage should have included installing the new motors. However, they arrived much later. As a temporary solution, the machine was fitted with its old motors.<\/p>\n\n\n\n

Surprisingly, the old motors turned out to be quite functional. I even took the time to clean off the stubborn paint and restore them (on the right \u2014 slightly cleaned; on the left \u2014 fully restored):<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

I even managed to control them without the control unit.<\/p>\n\n\n\n

Back during the construction of the 3D printer, I had put together a kind of “universal stepper motor driver.” A little box with the grandiose name “Elite eXtreme” (don\u2019t ask why\u2026):<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

It can drive a motor forward or backward and adjust the speed. Settings include selecting the required current power and the number of steps per revolution. This allows it to be connected to a fairly wide range of stepper motors.<\/p>\n\n\n\n

There\u2019s no separate article about it, as there\u2019s not much to write: an Arduino + a cheap external motor driver + a couple of switches + a couple of indicator lights + a three-line firmware. Nothing extraordinary, but it has saved me on multiple occasions and has earned its place in the lineup of “useful tools that should always be on hand.” Now it lives on the shelf above my desk, alongside the “Resisting Mantis<\/a>” \u2014 within arm\u2019s reach.<\/p>\n\n\n\n

Using this magical little box, the focus during assembly was not so much on kinematics per se, but rather on verifying alignment, parallelism, and perpendicularity of movements along all axes:<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n

When, right after assembly \u2014 without any fine-tuning or bolt-tightening \u2014 the marker hit all corners of a template square, it became clear that this whole endeavor might actually amount to something, and all the prior struggles to correct misaligned parts were worth it. The discrepancy was about 0.1 mm, which was quickly corrected by tightening a couple of bolts in the frame.<\/p>\n\n\n\n

The circular test was postponed until the control unit arrived, as the magical box can only control one motor at a time, and you can\u2019t draw a circle with just one motor.<\/p>\n\n\n\n

\n\n\n\n

In addition to hardware, the machine gradually acquired various auxiliary components \u2014 including ones it never had before.<\/p>\n\n\n\n

For example, limit switches. Oddly enough, the original version of this machine wasn\u2019t equipped with limit switches at all. This meant it couldn\u2019t automatically home itself to a “home” position, nor did it have any safeguards against exceeding the working area due to errors.<\/p>\n\n\n\n

Standard micro switches were used as limit switches.<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

I have plenty of them, and they have their own story! But I\u2019m not going to share it\u2026 If I were to tell the story behind every single nut and bolt, this series of articles would never end. Almost every part I\u2019ve salvaged has its own history \u2014 like the bricks for Uncle Pumpkin\u2019s house.<\/p>\n\n\n\n

\n

“This is the very brick I got ten years ago for Christmas,” Uncle Pumpkin would recall, picking up one of the bricks and stroking it like a kitten. “I bought it with the money I\u2019d saved for a holiday chicken.”<\/em><\/p>\n<\/blockquote>\n\n\n\n

Why micro switches instead of optical, inductive, or any other modern alternatives? Well, I must admit I caved to the opinion of the “general public.” In their blogs and YouTube channels, people strongly advised against using progressive modern limit switches for “metallic” CNC machines. They claim optical switches can be triggered by stray chips flying into their sensor area, or they can go haywire due to the presence of a nearby metal workpiece, or they simply fail to withstand vibrations. The consensus seems to be that a simple micro switch is the best option \u2014 no need to overcomplicate things.<\/p>\n\n\n\n

I personally have no experience comparing different types of limit switches in this context. However, the arguments presented seemed reasonable to me, and I couldn\u2019t come up with any counterarguments off the bat. So, I didn\u2019t overthink it and used the stockpile of micro switches I already had.<\/p>\n\n\n\n

Later on, special covers were made to protect the sensitive parts of the limit switches:<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

These covers were designed not only to shield the switches but also to prevent accidental triggers.<\/p>\n\n\n\n

\n\n\n\n

Since some auxiliary components were missing from the original design, as mentioned earlier, they had to be developed from scratch. No matter how thoroughly you model them in advance or test-fit them on a computer screen, there\u2019s always a chance something won\u2019t align properly. Having a 3D printer significantly simplifies this task, as it allows you to first print a prototype, test-fit it in place, make any necessary adjustments, and then print the final, accurate part.<\/p>\n\n\n\n

\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n\n\n\n
\"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

Prototypes for various mounts, holders, supports, and so on were printed using red filament. This served as a visual reminder to later reprint them in proper quality using black filament, should they fit correctly on the first try.<\/p>\n\n\n\n

\n

For prototypes, I use a print mode with “one top\/bottom layer, two perimeters, 10% infill, 0.24\u20130.28 mm layer height.” This, naturally, results in a noticeable boost in printing speed. In just 10\u201315 minutes, you can have something to check for hole alignment and dimensions. However, such a part will be mechanically very weak compared to a final print with “three top\/bottom layers, three perimeters, 35% infill, 0.16\u20130.20 mm layer height.” Thus, all prototypes must be replaced with proper parts; otherwise, they won\u2019t last long.<\/p>\n<\/blockquote>\n\n\n\n

Interestingly, in some cases, the seemingly out-of-place red color unexpectedly turned out to be quite useful, despite not matching the machine\u2019s color scheme. As a result, certain components were deliberately printed in red filament for their final version. This included, for example, the mounts for the very same limit switches. The red color immediately signals that this is a sensitive part of the system, encouraging one to instinctively handle that area with extra care \u2014 keeping it clean, avoiding damage, and generally paying heightened attention to the component.<\/p>\n\n\n\n

\n\n\n\n

The cherry on top of the auxiliary components was the lighting. After all, what machine is complete without it?<\/p>\n\n\n\n

To implement this, I had to modify the overall design of the gantry in the following section:<\/p>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n
    \n
  1. Top Cover<\/li>\n\n\n\n
  2. Aluminum Base of the Gantry\u2019s Horizontal Beam<\/li>\n\n\n\n
  3. Bottom Cover<\/li>\n\n\n\n
  4. What we can conditionally call a “lampshade,” which also functions as a diffuser for the light sources.<\/li>\n\n\n\n
  5. LED Strip<\/li>\n<\/ol>\n\n\n\n

    The primary purpose of covers 1 and 3 is to act as channels for routing wires from the limit switches and other auxiliary components. These wires are hidden inside when connections need to be run across the machine along the horizontal beam:<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    Additionally, the bottom cover includes an attached diffuser with an LED strip:<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    This setup ensures that there\u2019s no stray “spaghetti” of wires within the machine\u2019s workspace that could get caught in the cutting tool or interfere with the workpiece. At the same time, it provides some background lighting:<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    However, this is only ambient lighting! It\u2019s more for aesthetics and ambiance, with little practical necessity.<\/p>\n\n\n\n

    On the other hand, lighting on the spindle is entirely justified. No matter how well-lit the workshop is, the machine\u2019s own components cast shadows within the working area. To see more, better, and further, additional lighting is essential in the workspace:<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    This was kept as simple as possible, using six standard bright white LEDs:<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    Naturally, all these additional components (lighting, limit switches, brackets, covers, etc.) were simultaneously incorporated into the overall model and drawings, becoming an integral part of the project at the level of full documentation.<\/p>\n\n\n\n

    \n\n\n\n

    Many assembly operations were carried out simultaneously, so in the photos, the machine appears sometimes with the old motors, sometimes with the new ones, sometimes with the old spindle, and sometimes with the new one. In this article, I find it more convenient to group events by topic rather than the sequence of operations.<\/p>\n\n\n\n

    To add to the confusion, various adventures disrupted the process. For example, installing the new motors triggered an unforeseen incident (photo below: the “long” one is new, and the “short” one is old, for comparison):<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    The new motors are Nema 23, 2.8A, 269 oz.in. The old ones were also Nema 23 but were so ancient that I couldn\u2019t find their specifications. However, they were significantly weaker \u2014 that much is certain.<\/p>\n\n\n\n

    I discovered this during a test run of the Y-axis when the new motor, without any smooth acceleration (since the controller wasn\u2019t in place yet), went full throttle with such torque that it outright broke the coupling connecting it to the worm gear:<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    To prevent similar incidents in the future, I had to replace all the couplings with more robust ones \u2014 membrane couplings:<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    These do everything exactly the same as standard “spiral” couplings but can withstand significantly higher loads. Overall, they are also much more durable.<\/p>\n\n\n\n

    \n\n\n\n

    The spindle was also replaced at some point during the process.<\/p>\n\n\n\n

    The updated machine produced its first bzzzz sounds using its original spindle. This spindle had been restored, cleaned, lubricated, and even fitted with new brushes:<\/p>\n\n\n\n

    \n
    \"\"<\/figure>\n\n\n\n
    \"\"<\/figure>\n<\/figure>\n\n\n\n

    It\u2019s a decent motor, in principle, if you overlook its power supply requirements. I even managed to dig up something resembling a datasheet for it:<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    It\u2019s not very suitable for milling, but it works well enough for light engraving. I\u2019ve put it aside in storage \u2014 it might come in handy someday.<\/p>\n\n\n\n

    On active duty, it was replaced by a new spindle, about three times more powerful and equipped with a full suite of accessories and mounting hardware:<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    Daedalus is considered a decent brand in this weight class (under one kilowatt). We\u2019ll see how it performs\u2026<\/p>\n\n\n\n

    \n

    A note from the future:<\/strong>
    We\u2019ve tested it. For its price, power, and RPM \u2014 it\u2019s quite solid. No miracles, of course, but it delivers what was promised and what was expected. It cuts aluminum with reasonable confidence. It could use more RPM, but for now, we\u2019ll work with what we have. In any case, it\u2019s leagues ahead of what was originally installed.<\/p>\n<\/blockquote>\n\n\n\n

    \n\n\n\n

    When it comes to wiring, a CNC mill is no different from a 3D printer \u2014 both are more about “the science of wires” than mechanics. There are a lot of wires. Just an overwhelming number! At first glance, you\u2019d think: three motors and a spindle, no big deal. But then you add five limit switches (two for X and Y axes, one for Z), lighting, probes (which I\u2019ll explain later), camera connections, and so on. Some components need 12 volts, others 24 volts, and yet others 5 volts. And the list keeps growing\u2026<\/p>\n\n\n\n

    Fortunately, I knew what I was getting into. That\u2019s why I immediately designed spacious, flexible cable chains into the system \u2014 roomy enough to pack in all the spaghetti, no matter how much there ends up being.<\/p>\n\n\n\n

    \n
    \"\"<\/a><\/figure>\n\n\n\n
    \"\"<\/a><\/figure>\n<\/figure>\n\n\n\n

    I used the same flexible cable chains as for my 3D printer. The printer has even more wires, but everything fit perfectly in those chains. These chains didn\u2019t disappoint this time, either.<\/p>\n\n\n\n

    \n

    In the past, I tried many different chains from various manufacturers. Of course, I\u2019m talking about the lower price range. Most chains in this category suffer from poor flexibility, bending in unintended directions, breaking at joints, and so on. After plenty of trials and countless returns, I finally found some genuinely decent ones.<\/p>\n<\/blockquote>\n\n\n\n

    As usual, I followed the concept of “cable autonomy.” This concept has proven itself 100% in previous projects! The idea is that no wire runs continuously from the power source to the consumer. Instead, it\u2019s broken into segments linked to specific flexible chains and machine modules. Wires are connected between chains using connectors, forming necessary branches. This can be visualized like so:<\/p>\n\n\n\n

    \"\"<\/a><\/figure>\n\n\n\n

    In the schematic (loosely defined):<\/p>\n\n\n\n