leon_sbt

I saw Span a few times. It is a nice idea, but I'm personally not a fan. Doing switching at the panel level seems very clunky. In reality, switching should be done at the device level. Ie per fan , light etc. If they had a self contained server similar to IOTAWATT that just instruments the power draw of each circuit then I would be a fan.

Another vote for IotAWatt . I installed one a few months ago. No cloud app. visit an IP to view the local webserver. Has a neat graphing functions, and can pipe data to InfluxDB. + Home Assistant intergration. Love it. I wish they had a 30 circuit variant, so I can monitor every single circuit in my panel.

I was close to this exact situation. I regret it.

I'm assuming your in your 20's. Find a well paying job in an area you want to live in, and moonlight your business. When your business pays more than your day job, decisions and risk calculations are easier to make.

It also forces you to delegate from the get-go since your only have 4 hours "side project" time to work after work. If you're building something big, you will have to work with a team. If I was more time constrained, I would have probably made more tactical decisions.

If your esports thing works, great! If not, you hedged your risk and you didn't suffer from opportunity cost.

Throw everything I said out the window if you achieved product market fit and are growing at a viral rate.

Ultimately do what trpv said. "Go with your gut"

Also employed self-taught developer here. Learned to code through launching a software-oriented business. It took about 1.5 years for me to figure out the basics of (Frontend,Backend,AI).

Hands down have a project in mind, and start building it. It really helps to have a friend that codes the same stack, so they can provide guidance as needed.

If that is hard to come by, then pay a contractor on Up work to give you advice. I never solicited for outside help, and I deeply regret the time I wasted on figuring out things the hard way. Best of luck!

You have two options. Call up a local machine shop (check Google), ask if they do one-off parts. (They probably won't); then ask them if they have any solid recommendation for a local small-time machinist that enjoys one-off work. They probably have a couple on speed dial for jobs that aren't worth their time.

You can also checkout cnczone.com , practical machinist and a number of hobby/home machinist Facebook groups.

I've dealt pretty heavily in this area (Bought and operated several CNCs over the years) Your price point is going to limit your machines to almost exclusively Chinese imports

Lasers: Boss laser imports machines from China and provides great American-based support. If you can figure things your self, check out Cloudray laser. They ship straight from China; keep in mind the customs/shipping risks. I've purchased from both before. If you're not running production jobs, 75W C02 should be fine. I prefer 150W RECI C02 laser tubes. The go-to Chinese controller is made by Ruida. At the end of the day; most of these Chinese laser brands all have the same part suppliers, just different levels of QC and quality assembly. I'd avoid Glowforge and similar brands. They are a great for first time users at home; but I'd be more worried about work volume, laser power, and future spare part availability. I know quite a few people that outgrew their Glowforge-like machines and went with bigger more industrial imported lasers.

Purchase Lightburn for sending jobs from the PC to the laser.

Routers: I have a few friends that have a larger Shapeko-like machine that works well for Etsy-like products. Think embossed cutting boards, wall art etc. Granted there part is size like 12"x 24". The machine isn't stiff (compared to industrial units); but it gets the job done. Whatever machine you purchase you'd want to stick to linear guide rail motion system for all the axis; avoid the roller-wheel style motion systems if possible. If I personally was going to buy a new router in that price range; I'd buy either an Avid Pro machine or import a chinese machine straight from Alibaba. Start with checking out OMNICNC in China. Depending on your location; check out classifed ads, or talk to some local signmakers. They might be selling their older US-made Multicam machines for cheap.

I can give much more specific recommendations based on your production rate and part size/features/materials.

EDIT: Mirroring what other people are saying. If you just need some parts just to get brewery going, then I'd sub this out to someone else. If you have time, money, desire, love to spend many hours and thousands of dollars to learn the art of manufacturing, then go for it. But I'd recommend to solve the goal at hand first. Your brewery.

OP here: I was looking for a way to search people's interesting personal projects on the internet. I realized the a huge common denominator was many people hosted it with Blogspot. I found this random site that provides search on Blogspot. Seems like a gem of a search tool to view the internet from the pre-SEO stuffing internet era.

I'm happy to do so!

I also heavily participated in Formula SAE. Don't get too distracted by the Auto sport connectors. They are top-tier hardware, but use Deutsch DT/DTM connectors for everything. They work quite well if you're on a budget, but still want very high quality interconnects.Our team used DTM connectors for a while, before eventually migrating to Autosports.

Not necessarily modern technique based hacks, but my personal hacks boil down to 3 things.

1. Composition matters. Alot. Move yourself,your camera, and your subjects around.

2. Control your light naturally. Pay attention to directionality of light, shadows and softness. You can cover a light with your body or change directions.

3. Slow down and take your time. Imagine your shooting film, make the shots count.

Using these techniques, I stopped carrying my DSLR and only shoot with my phone. The photos from the phone have far higher quality and value since I have it in my pocket at all times. Making "capturing the moment" a very accessible thing.

Exactly like jogging with CNC controls; but with force feedback. I have done some cuts with the Haas in handjog mode. I always feel iffy doing it, since I can't feel the feedback from the cutting forces.

I was about to recommend the exact same thing. Epoxy granite is used by some world class machine tool builders.

I think an interesting modern take on the Gingery machine, would be a homebuilt machine tool; built from epoxy-concrete. Linear rails, servos, fly-by-wire controls with force feedback.

Machine form factor would be a la Mazak Integrex but much smaller and fits in a two car garage. Total envelope would be 8"x8"x8"; high spindle speeds w/ low depth of cuts. That should compensate for the total lack of stiffness :)

You have a very valid point. I'm 5'11" so getting my foot down at stop lights has never been an issue; unless I'm dealing with a 450cc dirt bike.

My GF is about 5'2 and stopping a motorcycle in traffic is actually a pretty big deal. She isn't tall enough to put her foot down, without awkwardly angling the bike. The only bike should could ride of mine is a 250cc ninja. The 400cc+ supermoto was way to awkward/tall for her to comfortably ride.

I'm totally open to the self-balancing idea now.

Really interesting concept. I like the general lines of the bike. Looks really well thought out/clean sheet mechanical design.

Two points of contention:

1) After putting over 40k+ miles on various bikes, I never once wished there was a self balancing feature to make it stand up on its own. At a stop light; I put my foot down; at a really long stop light, I put the kickstand down.

If the self balancing feature can function as form of driver aid. Meaning you declarivity tell the bike "I want this line with this apex"; then the bike will follow through with it's active systems. That would be wild.

2) The device right behind the rider gives me goosebumps. In a crash; seems like a great way to break your back and be pinned in the bike.

Background data:MechE background. Owned and crashed high CC several motorcycles. I'm doing totally fine healthwise; just really lucky.

Thank you for the extra photos and info! It really helps. Especially with the notes on the motor mounting. Have you spec'd out Linear motors for your machine? I have been eyeing Yaksawas's product. But I'm scared to RFQ the product due to sticker shock :)

The machine you built is very professional looking! Great job! I wish more people did high quality "last mile" integration" on their machines. Energy chain and powder coat makes a huge difference in perceived quality.

Couple of questions:

1. Why did you chose to mount the Y axis servos in that orientation instead of where the axis of the servo is perpendicular to the floor?

2. Since you welded the frame, how did you manage the mount the linear rails square? Shims? Post weld machining the frame via bridge gantry?

3. What was the total cost of this vs buying an OTS machine? I'm guessing around $12k.

The most simple answer:

Because people are willing to buy them.

It also supports typical human behavior. If you're neighbor has a 400hp car, to "one up" them you buy a 500hp car.

Childish? Yes. Are adults with childish behavior new to the human race? Not at all :)

Hi Dave, I'd love to hear your thoughts. I'll send an email now.

It boils down to two things currently.

1.I spent all the money so far on buying machines and hardware infrastructure. I can slow down on hardware purchases now.

2. The fulfillment side of things is a bit of a mess, its a mix of my SQL database and my email functioning as the ground truth datastore. It makes it tough to hand off to an employee for the fulfillment. I think I really just need to do a low code Trello based workflow for fulfillment in the meantime. Then build those features later.

Hi Rob,

Thanks for chiming in. I can keep my dayjob for a while. Which in theory gives me unlimited burn rate. But I'm worried that I should just cashout everything. I won't lose any money. Then move to Saas,where the overhead and unit economics are incredible. Or just do software engineering as an employee with much less stress.

Thank you for mentioning Arvid, I'll keep him in mind!

I ordered from Sparkfun multiple times and everything worked out great. I wonder if they are slowly going to expand their reach to be a la protolabs for all the B2B PCB design/prototype deals. Protolabs did a quarter billion sales last year I think.

Clearing $1k from semi automated work sounds a lot less tedious than packing 500 bags with a $2 sensor. I always feel like they lose money on sub $10 orders.

You're 100% right! There are a couple of pet peeves that McMaster Carr does that I wished they fixed. Cough, cough,public API access. But I assume most of these points are used to prevent scraping of their site.

Sometimes I spend close to $2k a month at McMaster for work things. They are literally a modern marvel. I order something at 6pm . It's there at 9am the next day.

I buy things from them even if it's double price, since I know there's no drama with their orders. It just works. Everytime.

They are the defacto standard that I have for a great e-commerce experience. When I need design /UI tips for my website. McMaster is always referenced. Never Amazon...

Thanks for digging that up! I actually remember reading that thread a while back. Most of that thread is "my computer is broken", but their are a couple nuggets of gold in there.

Got a link to that thread? A few of my buddies are machine shop owners, and I would love to see what they said.

Great points you brought up! Only have experience with the Prusa's and it's been great. I never operated the a belt bed printer.

It seems like another way to solve the problem is to stack hundreds of the steel sheets in a magazine behind an i3 styled machined. When the print is done, the Y axis goes all the way to front of the machine.

It will basically hit the steel sheet in a ramp and unload the sheet, through the front of the machine. (In a clean under actuated robotic style). Move the bed to the other extreme to reload a fresh plate. This would involve a machine with more Y axis range of motion.

If you're familiar with CNC machine tools, this would operate similar to a Brother Speedio Tool changer. Where the Z axis spindle motor actually indexes the tool changer, when Z axis is high enough to engage the tool changer.

Regardless, I'm excited to see how both style machines solve the next stage of lights out automation.

This is really interesting. In my opinion, Prusa is the best in class printer to purchase for a turnkey machine. If you want to scale up a printfarm. Add more Prusa's and some humans to tend them.

But there was never really a way to run the print farm lights out. As in queue up the job for 7 days. Then have humans show up at the end of the week and package ship orders.

If there was an OTS machine that could contend Prusa for production value per dollar. This looks to be it. I'll be buying one of these when they are out.

For the patent comment, I think Stratasys has a patent on it, but one of the claims is that the machine prints with orthogonal axes. Angle the print head and now its good? https://www.stratasys.com/3d-printers/continuous-build-3d-pr...

If you're targeting a browser based solution. Three JS and Babylon.js gives you the building blocks.

If you're looking more of a Constructive Solid Geometry based web solution jscad/CSG.js.

If you want something more turnkey with a BREP kernel look at what the folks at https://www.autodrop3d.com/parametric-cad.html and https://buerli.io/ are building. Super interesting.

Not affiliated with the above companies, I just do dev on related software.

Like what some people said 5/100 on the page speed score is pretty bad. My site used to have this. It boiled down to two things that caused it.

1) I wasn't code splitting by page, and had a couple heavy dependencies (which made things worse)

2) I wasn't pre-rendering my create-react app via react-snap. Meaning during the CI build react-snap runs Puppeteer visits each page on the site, and generates ready-to-go html versions of each page.

Those two changes took me from a 7/100 to a 95+/100 in short order. Makes logical sense too. Now a days the site hovers at 85/100. But I don't have the time right now to reinvestigate it.

With the options at the time, I'm happy with my tech choices. If I had to start with React again today. I would do Next.JS with a static build output. Would save me a bunch of time scaffolding things.

If you already know Ruby on Rails, do that.

If you already know Python well, do Django or Django REST.

If you don't have any prior language affinities, go Javascript/Typescript on the whole stack. Language switching is a real problem for me.

I'd stick with App Engine/Heroku for prod. I like Postgres alot, especially with JSONB. Since it functions as a MongoDB-like database if needed.

If App Engine fails you, you should be able to switch to another provider very easily, or a standard VM without a code rewrite only a CI/CD rewrite

With that being said,I run a monorepo. /api folder has Feathers JS with Postgres /admin folder has react admin website /website folder has public facing react website.

Everything can run local during dev. Everything is CI/CD piped 12 factor styled One git push and merge to master, deploys everything. I love it.

If a single provider fails me, I can switch pretty easily.

I think there is a pretty large interest in something like this. But I don't think it will as simple/automated as you would like.

At the minimum, you would need to test/review each product. Provide truly valuable insights, and touch/feel each product side by side.

There are several people that do this in their respective niches, and make a quite a good living for themselves. But keep in mind it took them years of doing it for fun and acquiring taste before anything worked out.

Based on your original post, if I visited the site.It would feel like one of those low quality affiliate link stuff websites.

But if you add some truly valuable insight,data etc. It could be worth it.