site banner
ArjinFerman  Tinfoil Gigachad  20hr ago (text post)   253 thread views

Tinker Tuesday for May 13, 2025

This thread is for anyone working on personal projects to share their progress, and hold themselves somewhat accountable to a group of peers.

Post your project, your progress from last week, and what you hope to accomplish this week.

If you want to be pinged with a reminder asking about your project, let me know, and I'll harass you each week until you cancel the service

Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

Good news: my bass body arrived early.

Bad news: the neck pocket is cut to the wrong spec, and it's out of spec even if it were the correct one.

I have contacted the seller. We'll see what happens.

114k on my NaNoWriMo project. I would like to say that the end is in sight, but who knows. It actually feels like it's getting harder and harder the closer I get to the finish line. If I'd known that, when I'd finished my initial 50k words at the end of November, that would represent less than half of a first draft - I'm not sure I would have started in the first place.

So, the last three weeks having been a battle against the loss of morale realizing you did things kinda wrong and there is no going back. But I'll get to that.

I milled the rest of the lumber for my back chair legs, and got them rough cut out, working around knots and weirdness as best I could. Some of the layouts didn't leave a lot of room. But for the most part I got 4 chair legs out of plank, in pairs of two. This took for fucking ever with the jigsaw, and my urge to buy a bandsaw has intensified. I saw Harbor Freight recently released one with an 8" resaw capacity that's only $600 which is very tempting. Maybe if I get a Christmas bonus. Anyways, it got done eventually.

Next comes attaching the template to each leg with some double sided tape. You do a pass on the router table with a templating bit, remove the template, and then hit it again to remove the rest of the rough edge. It takes a few passes. I even had to hit this with a flush trim bit from the top after I got as far up as I could with the template bit from the bottom. Cleaning them up with a sanding drum on my drill press and they aren't looking too bad. At some point I realized, after enough really ugly tear out, that I needed to reverse my feed direction. Normally I hate this, because instead of the bit pulling the work piece in closer against a fense or other positive stop, it wants to climb the workpiece and spit it out. But I had to trade that safety for not consistently destroying my workpieces with massive tear out. The curve of the legs just offer too much chance for the bit to catch an odd grain and rip a huge chunk off. A few weren't so bad and sanded out, but one piece that tore away was so large I was able to glue it back into place and resume my work.

But I said I messed up, and here is what happened. I should have used MDF for my template instead of a piece of 1/2" sheathing. I figured sheet goods are sheet goods and it's what I had. But it turns out, sheet goods are not sheet goods. The sheathing had all sorts of warp too it, and flexed too much. Every leg I routed using it came out just a little bit different. This would not have happened with MDF which basically has no give at all because it's more glue than wood. Alas. Still, with 8 legs for 4 chairs, I was able to pair each leg off with it's closest match. So fingers crossed the final products don't drive me insane with their dissimilarity.

Next up came getting started on the jig for mortising the back legs. Fingers crossed it works out when I finally go to use it. It should handle the mortising for the back rest and the front to back apron pieces. I'll need to use a seperate jig for the side to side apron pieces and the brace down at the bottom which is going to be the most challenging mortise.

With the legs behind, I'm excited to plow on ahead towards dry fitting my first complete chair.

Still working on learning Blender, health and family allowing, and building a list of thing I want to make as time goes on. Current goal is still the texturing workflow. Made some good progress through the current tutorial this weekend, but finding time has been tough.

A few years ago I had a bit of a 3d modelling streak, but I mostly used Maya thanks to my school getting the adobe licenses. I really enjoyed the actual modelling over things like texturing or rigging, and never really learned them. What's motivating you to learn Blender? I was mostly into it to make things I thought were cool (guns, tanks). I've been halfheartedly pondering downloading blender over the summer to try and make some sort of short action scene. What's your current project?

I've been an artist in the video games world for more than a decade, but almost all of that time was spent on projects that used fairly lo-fi art, and early on I moved to a stripped-down modelling program. The project I'm current on is wrapping up, and I'm not sure I'll have a job once it does, so I want to brush up on my generalist skills in preparation for a job hunt. I've generally made guns and tanks and spaceships in my free time, but I've always wanted to get deeper into the tech, simply because it allows you to make cooler things; a gun is great, but a gun held by a character is better. A mech is great, but a mech with full texture and rigging stomping around an environment is better.

I've tried to learn Blender before, and it was going somewhat well, but I ended up going back to my old low-key modelling package for work, which killed muscle memory, and I never got back to it. Also, I don't think my approach was nearly organized enough to handle the complexity of Blender. This time, I'm taking a more serious approach:

  • First, I'm burning my ships. I'm making a personal commitment to never touch my old modelling package again. Blender or bust.

  • Second, I'm being much more structured about my approach to learning. There's a bunch of tutorials online for learning blender, but rather than just working through them, I've set up a google doc and am taking copious notes of hotkeys, techniques and so forth, essentially writing my own manual. This helps a ton with both retention, and with having a quick reference when I forget things.

  • Third, there's a ton of "quick tip" content online, showing random disconnected features of the program. I used to see this stuff and think "wow, that's neat", and then forget it. Currently, I'm collecting those links in the same doc, with the goal to work them into my notes and workflow when they become relevant.

Last time I tried learning Blender, the goal was to get to modelling as quickly as possible. This time I'm aiming more to learn the deeper toolset; it's obvious to me that I've wasted a huge amount of time doing things the hard way, when tools are available to make those things much easier, and that's bad both in a professional sense and an artistic one.

My current project is a gun; I've got a pretty good low-poly model, but I want a high-poly, normal bake, textures and simple function animations, the standard game-ready pipeline. I have a bunch of spaceships I'd like to finish the same way, and I've got some characters/environments/action scenes I'd like to try as well, plus a ton of other stuff; I've got a huge backlog of projects and Ideas that I've never had the chops to really execute on.

This tutorial is a really good introduction to the modelling end of things. It's a bit of a slog, but it handles setting up hotkeys and a number of addons for a more efficient workflow, goes through the basic modelling tools, and dips into materials, modifiers, a touch of simulation, and rendering. Be warned, some of the plugins he recommends (Jmesh in particular) currently don't work, and in some cases you'll need to figure out how to work around them.

This tutorial is a good follow-up, and is the one I'm currently working through. The modelling is more intense, and Jmesh being broken meant I needed to take a detour to learn some other methods for handling circular arrays, but it demonstrates a ton of useful techniques and approaches to a lot of standard modelling problems, and gives good working practice with the modifier system.

I've also been working through overviews of the shader and geometry node types, in preparation for texturing and environment generation. I've dipped a bit into some of the simple rigging, but want to tackle true high-poly > normal maps > texturing > rendering before diving into that.

Overall, the thing that makes me excited about Blender is seeing examples of how it seems to bust the general trend of overspecialization. videos like this one are pretty astonishing in terms of what is actually possible with a deeper understanding of the tools. The multiplication of effort is real, and I've got a wealth of experience of how limiting it is to be stuck with the old, slow way of doing things. An obvious example: I've spent probably hundreds of hours modelling rock over the years for various environments. I'm pretty sure I can spend maybe twenty hours learning the geometry nodes system, and then produce rock environments with ten times the fidelity in a small fraction of the time. I'm tired of doing it the stupid way.

Also, if you're comfortable posting any of your art, I'd love to see it!

This is going to be, quite-possibly, a below-illiterate tech question, so please bear with me and save all openly expressed disdain for the end.

I run a small business that makes about $20K a year, as a side hustle. I started soon after AI hit the mainstream and have found the $20 a month tier of ChatGPT to be invaluable for streamlining administrivia type tasks such as boilerplate emails, plus helpful for very early brainstorming and having a minimally effective sales pitch in 15 minutes. I still do some amount of cleaning up for these processes, less for the boilerplate things, more for the creative things. I have trained a couple of GPT’s to be focused on the specific tasks I need them for, and will continue to do so as/if I expand.

My question is, am I missing out on some capability by only using the basic bitch version of GPT? Could I be getting more bang for my buck, better sales emails, better crafted first pass sales pitches, more automation, etc, by changing products? Should I use a different LLM company, or pay for API access, or buy a good GPU and train my own sandslave, or what? Or am I fine where I’m at?

By paying for ChatGPT Plus, you have access to o3, which is the best model in OAI's stable, and a cutting edge LLM.

That being said, it likely isn't the best model out there, as Gemini 2.5 Pro is equivalent/slightly better. I don't think it would be a noticeable difference, in the use cases you mentioned. However, you can use it for free, and without limits too, on Google AI Studio. No catch, beyond the absolutely not worth worrying about fact they'll use your chats as training data. That would save you $20/m, but beyond that you're doing things fine.

https://aistudio.google.com/prompts/new_chat

All you need to do is link your Google account and you're set.

In brief, you're probably fine where you're at. Keep the 80/20 rule in mind, you're already getting a lot of value out of the "basic bitch" level. If upgrading your tool doesn't directly solve a specific problem you're having, I wouldn't. There are infinity arguments for why upgrading could potentially allow you to do more things: if you don't have a plan for exactly what those more things are and how an upgraded LLM fits into that process, don't do it.

So, as mentioned in previous weeks, I've been trying to TRONify my cargo bike and kids' bike helmets the last few weeks and I've never quite appreciated how much of small electronics work is shopping in general and specifically, shopping for plastic pieces of shit. This isn't what I thought I would struggle with. I'm a fairly skilled dev so I thought I would struggle with EE concepts or the dexterity required for soldering or even simply being able to work with a magnifying glass but no, shopping, which I find borderline triggering due to hatred of clutter, is the actual limiting factor for me.

And shopping for an enclosure for housing all of this junk is the challenge. For the bikes themselves I can use standard Hammond plastic enclosures and consumer battery packs, but for bike helmets all of the standard stuff adds way, way too much bulk.

The rough sketch of components I need housed are

  1. ESP32 board, mounted to a perfboard
  2. a boost converter to bring the 3.7v put out by the 18650 cell up to 12v for LED lights
  3. a buck converter to bring 12v down to 5.5 for the ESP32 board
  4. a charging board with USB-C[1]
  5. an SPST(?) switch for physically breaking the connection from the battery to everything else
  6. an N-channel MOSFET for controlling COB LEDs (if I supplement the WS2815s)
  7. a P-channel MOSFET for having the rest of the load shut off if it detects that a USB charger is connected
  8. nylon standoffs for the various boards
  9. a holder/receptacle for the 18650 battery itself

Finding an enclosure that isn't as bulky as shit for this seems impossible, so I'll probably have to 3d print something. Which means I need a 3d printer[2]. From my time being involved with a local Makerspace, my opinion of 3d printers is that they spend most of their time being broken, but I've heard from trusted advisors that that's because my experience has been with 3d printers made by the decadent and pathetic Western concerns and that the Bambu 3d printers from China have changed everything[3]

So... maybe that's my next purchase. Perhaps I can justify the 3d printer as some kind of educational value for the kids.

Does my experience here sound right so far? Small electronics success often hinges on shopping skill?

Notes

  1. I do see combination boost converter/USB chargers for 18650 cells, and that would cut down on complexity/bulk/work, but their amperage is much too low to run LED strips off of so I'm stuck buying individual components for this.

  2. Though I think if I want a really cool looking transparent enclosure so we can see circuit boards and blinken lights I'll need acrylic covers? Which requires getting a laser cutter too? I wonder what educational value for kids these have...

  3. So long as you have no opsec concerns from running proprietary firmware that requires a cloud connection to do anything from a nation state that we might go to war with in the near future. Though I'd be kind of amused to see the worst that can happen.

Switching gears (ha), but to avoid making another post, I'll consolidate into this one.

Aliexpress.com has incredible deals but takes forever, so that stalls my TRON helmet project out.

So, in the meantime, I've gotten to triangle dodge charger skydome in my 3d game (pic attached) that was inspired by binge watching the Fast and the Furious series while feverish. I made some generic cyberspace background while waiting for the right time to take a 360 panorama pic of a skyline in my town.

I fixed the texture banding issue. The camera now follows the car. The car can steer and accelerate. It runs okay on my hardware. So... I guess the next priority is to make a race track and add some collision detection to this bitch? A computer controlled car to drive against? Maybe incorporate some engine revving sounds?

I suppose a true Ride or Die Homies game needs something more inspired than a race track though. Like outrunning a nuclear explosion. Or battling a mechanized raptor.

I should be doing this smart and using Unreal Engine or something, instead of writing a 3d engine from scratch, but I got into computers in the first place because I wanted to write a 3d engine (before getting distracted by the world of Linux and networking), so coming back to this feels like addressing some unresolved spiritual concerns.

I've used https://www.taydaelectronics.com for a few projects and found them to have a limited but pretty useful inventory at very reasonable prices, aimed at hobbyists rather than manufacturers.

On the power side, I'd at least consider 3S LiPo battery pack, running the LEDs from that voltage, and just using one buck converter for ESP32 (which can be a tiny 5 watt one). That's only a nominal "12v", and really a 10ish-12.6 range, so you'll want to double-check the datasheet for your specific LEDs, but it's well within spec for the WS2815s. You can get 3S 18650 chargers cheap (eg here, not endorsed), though premade packs are so widely available (and tend to have much better low-current protection) that it's hard to justify building your own packs unless you need 5+ amp.

While series is generally pretty safe, never use LiPo in parallel until you have a very good familiarity with the technology and necessary safety precautions. Further instruction on that matter not available here.

On the circuit design work, yeah, part control is a Problem. As you get more experience you'll start to collect a bunch of standard parts and design with them around whatever your target is, and only need to order a handful of specialized things, but on the path there you'll end up with literally thousands of bags of stupid little components that are worth fractions of a penny and you'll never use again. And even once you have your 'jellybean' parts together, you'll always need something specific to a given project, or find that your old parts aren't available anymore, yada yada. Breaking projects into modules can help, but then you're juggling them, too.

There are some "MakerSpace Starter Kits" (and a much broader number of order lists) for component-level parts, but as you go from components to prebuilt boards they tend to get a more 'cheap plastic toy rather than tool' problems. There used to be some good lists of common sensor boards for educational uses that I'd cribbed from, but SnapCircuits and Cubelets et all have kinda cut the bottom out from that market; nowadays, you're probably better off just spending a birthday at a Microcenter or on Adafruit.

For 3d printers, yes, for FDM the entry-level ones (eg Ender3s) tend to be glorified junk that will need maintenance within days, if not hours, and owning them is an exercise in modification and repair. This can be a great educational experience, but it's also a lot of ship of thesus for a tool. You can pay a bit more and get a professional- or educator- grade one that can go hundreds of hours without as much work or a lot more and get an industrial-grade one that gets thousands of hours before serious repairs (eg, GenFabCo), but unless you want to Be The 3d Printing Guy, it's hard to justify the cost of either.

For small parts, resin printers are another option. They're still fiddly, because support material is even less of a solved problem, but they don't so much have maintenance schedules so much as they just have parts you replace, and even the absolute cheapo ones (eg Elego) tend to work just as a 'replace screen and bulb every few hundred hours' sorta thing. They don't scale well beyond about six-eight inch size parts, though, and they're very much not things to be used without proper ventilation.

If you're only needing a handful and had little or no interest in 3d printing, seriously consider various fabrication services, or local marker spaces (or some libraries), rather than buying a full 3d printer.

Laser cutters are ... iffy as a value proposition; they're extremely limited in both what they can cut or etch (some woods, acrylic, anodized coatings), and also how (both depth and angle). The big value proposition to a comparable CNC is ease of use and operating costs, but this is a very rough tradeoff. I'm very hesitant to recommend them for any structural component, just because anything capable of cutting deep enough into transparent plastic to make a useful part is going to be ludicrously expensive.

For flexible, light-weight, shaped transparent pieces with little 3d complexity, my go-to has been vacuum-forming. Big ones get surprisingly complicated fast, but for anything under a couple feet square you can get away with stuff you probably have in your house, and the big trouble becomes sourcing the right types of plastic.

I'd also recommend considering Traditional Manufacturing -- just as there's a lot of people 3d printing what could be made with a handsaw in ten seconds, you may well be engineering something that could be sewn together in a good half-hour. If you want something flexible and comfortable for long-term wear, sewing is kinda the way to go. You'll still want some boxes for the batteries and protoboards to provide impact resistance, but it'll give a lot more space to consider multiple small project boxes or such, and those are a lot easier to source.

Does my experience here sound right so far? Small electronics success often hinges on shopping skill?

You know, once upon a time I wanted to do more electronics stuff. But this is more or less the brick wall I ran up again. Sourcing parts is bonkers complicated, and simply was not a part of the hobby I was going to enjoy. I repaired a few old motherboards, and developed enough soldering skill to repair the odd toaster, mouse or audio speaker, and more or less decided to leave it there. Although I'm always have my eye out for an inspiring woodworking/electronics project.

It's really the most hellish part. I'm trying to just take it easy and not worry about how soon I get this done to avoid stressing.

Starter kits help a bit, but they're only like $10 whereas I want something like $100-200 that contains the top 1000 things people use for, e.g. Arduino or LEDs.

I fantasize about taking a two week vacation to Shenzhen and hiring some Chinese EEs that speak English to show me around and going home with a suitcase full of stuff I bought off the street.

You know, that really strikes at the duality of it for me. I don't want to be fucking around with arduino components like electronic legos, but then the ecosystem of all electronic components is so vast and wide and deep that you really are just adrift.

My dream for a time was to build my own 8-bit computer with some cheap 6502, VGA output and synthesizer audio as a fun learning exercise. But it turns out nobody makes VGA chips anymore? Or audio? Basically everything is SOC, and every project you see along these lines has a legit 8-bit Z80 or 6502, and then uses an SOC as a co-processor for it's VGA and sound, or has you using 20-30 year old salvaged chips.

Oh, yeah. Seems like people want small low-power computers to drive widgets rather than make an IBM PC from 1986 on a cocktail napkin.

Ideally all of this should could be end-to-end done with CAD software, or by writing out a .spec file and uploading it to a magic box or web site and then a microwave dings and it's ready. Or it arrives in the mail. Maybe one day...

Maybe it's possible to 3d print ghetto PCBs and also 3d print conductive filament and snap it into place?

pic attached in last message wasn't showing up. lets try this one.

/images/17471501354596198.webp

What is this place?

This website is a place for people who want to move past shady thinking and test their ideas in a court of people who don't all share the same biases. Our goal is to optimize for light, not heat; this is a group effort, and all commentators are asked to do their part.

The weekly Culture War threads host the most controversial topics and are the most visible aspect of The Motte. However, many other topics are appropriate here. We encourage people to post anything related to science, politics, or philosophy; if in doubt, post!

Check out The Vault for an archive of old quality posts. You are encouraged to crosspost these elsewhere.


Why are you called The Motte?

A motte is a stone keep on a raised earthwork common in early medieval fortifications. More pertinently, it's an element in a rhetorical move called a "Motte-and-Bailey", originally identified by philosopher Nicholas Shackel. It describes the tendency in discourse for people to move from a controversial but high value claim to a defensible but less exciting one upon any resistance to the former. He likens this to the medieval fortification, where a desirable land (the bailey) is abandoned when in danger for the more easily defended motte. In Shackel's words, "The Motte represents the defensible but undesired propositions to which one retreats when hard pressed."

On The Motte, always attempt to remain inside your defensible territory, even if you are not being pressed.


New post guidelines

If you're posting something that isn't related to the culture war, we encourage you to post a thread for it. A submission statement is highly appreciated, but isn't necessary for text posts or links to largely-text posts such as blogs or news articles; if we're unsure of the value of your post, we might remove it until you add a submission statement. A submission statement is required for non-text sources (videos, podcasts, images).

Culture war posts go in the culture war thread; all links must either include a submission statement or significant commentary. Bare links without those will be removed.

If in doubt, please post it!


Rules


Recommended Posts And Communities

Recommended Realtime Chats