Ask HN: What is your hardware/garage side-project? | Hacker News Reader

Ask HN: What is your hardware/garage side-project?

63 pointsselmat80 comments

If have personal blog about it please share it.

I'm working on a headphone stand that has a USB DAC/headphone amplifier built in, along with circuitry to detect when the headphones have been hung up, thereby switching the DAC output to a line out (to drive speakers, or whatever).

It's a shameless amalgamation of a DIY post on Reddit[1] and some design aesthetics from an existing DAC/headphone amp[2] that I really like. I'm currently at the point where the internals are close to finalized enough to order boards and do integration testing, and starting work on modeling the enclosure, which will involve some mix of CNC milling, mandrel bending and welding.

I really like throwing away money, apparently.

Blog here: https://nuclearfurnace.wordpress.com/

[1] - https://www.reddit.com/r/3Dprinting/comments/529qyp/smart_he...

[2] - https://www.jdslabs.com/products/151/the-element/

That's really cool. I've always wanted something that would seamlessly switch between headphones and my monitors.

Yeah, right now, I have some dinky 2.1 speaker setup, and there's a remote control box that feeds back to the subwoofer to control the volume and expose a headphone port. I have to plug/unplug when I want to switch, and the volume control button/wheel has low resolution because it's so small.

The Reddit link uses a Windows program to do it, but I wanted it to be universal because I dual-boot, so doing it in the DAC/amplifier seemed like a better approach. :)

I've been developing a Motorola 68K-based retro computer. The goal is somewhere around Atari ST-era sophistication, but it isn't quite there yet. Currently has a UART, some switches, some LEDs, and a periodic timer. Still needs keyboard input, video output, removable storage, and a slightly smarter interrupt controller.

http://coderjames.bitbucket.org/

The website has a bunch of broken links to missing material, as I haven't gotten around to writing it yet. All the schematics, gerbers, and firmware source code are up though, so it should be reproducible.

I'm building a two person research submarine with some friends. Targeted diving depth: 100m

I have an electronics/IT background so building something so mechanical is a fun learning experience.

http://ubaad15.org

Now that's cool. Do you have some automated way of testing it or will you be in it on the first test at that depth?

I imagine that the first test pilot will be a computer. Submarines are really nicely simple systems, which is why it is possible for a bunch of amateurs to build one.

Love it !!

that's awesome, how do you plan on waterproofing all the openings?

The port holes will be acrylic windows in a window frame with O-ring seals. The O-ring will be placed on the side facing away from the water such that when the pressure rises on the window, the pressure on the seal rises as well, making it even more water proof. Same idea for the hatch.

I'm building my PiPlay Advance. A Pi Zero game console with all off the shelf components (plus our PCB) and only a few solder points.

Our BOM is about $35 - $45 right now before bulk discounts.

http://blog.sheasilverman.com/wp-content/uploads/2016/10/778...

https://docs.google.com/forms/d/e/1FAIpQLSerQdiZrYwMojO9M8Hi...

I'm currently working on trying to crack an electronic safe lock based on the awesome DEFCON talk by Plore ( https://media.defcon.org/DEF%20CON%2024/DEF%20CON%2024%20pre... )

I don't think the photos are exactly up-to-date, but I currently have an STM32F429 DISCO dev board, configured with a ladder DAC, so I can generate the analog keypresses for the safelock via my laptop, the board also sets a pin high to force my oscilloscope to trigger and capture the voltage from my ucurrent, which is hooked up to the battery of the safe-lock.

https://www.anfractuosity.com/projects/safe/

I'm also trying to get some iron nanoparticles, if I can find them cheap enough, in order to try to optically image a floppy disk.

I've got some images from one in the link below, but that was using a magnetic developer, with much larger particles.

https://www.anfractuosity.com/projects/optical-magnetic-stri...

A 40 digit calculator using a Ras Pi, Java, 7-segment readouts, MAX7219 chips driven slightly out of spec (ab)using the SPI pins. Two 4x4 keypads to form an 8x4 keypad so it will be a scientific calculator. Found a Java library to do transcendental functions using BigDecimal. All of this on a breadboard. (I'm a software guy who doesn't know which end of a soldering iron to pick up.)

Other project. A 64x64 RGB LED display driven by a Teensy 3.2, in C++. A friend who does hardware but not so much software sent a beautifully assembled rig in a very nice enclosure.

No blog.

I've been working on my generative adversarial network. It's a fascinating type of ML that can learn from both supervised and unsupervised data.

In essence, given data it will create a generator of that data. Right now my datasets are mostly images. I have tried on audio data and that does start to work - then diverges.

I'm still trying to find a configuration that works really well:

https://github.com/255bits/hypergan *

* needs modern GPU to run

This is really interesting. How much data does it require to build a generative model with say 60% accuracy for story type data?

GANs work on real-valued data. The problem with sequence data, like text, is that the error gradient can't flow through the generator.

There's some really great work being done called "Sequence GAN" that turns sequence generation into a RL problem. I'm not sure how successful they are, but it seems potentially powerful. https://arxiv.org/abs/1609.05473

I'm currently converting a cargo van into a camper van suitable for full-time living. I bootstrapped the project out of a garage in two weeks and will now be doing most of the work in SF out of the van itself. It's an exercise in woodworking, metal working, solar/electronics, and interior design. I'm documenting the build on Instagram (same username) if anyone wants to follow along.

https://www.instagram.com/bloudermilk/ for the lazy

I develop Open-Source software for CNC simulation.

https://blog.buildbotics.com

If you're in the bay and need someone with a simple 4'x8' CNC router running mach 3 to test with, I'm game.

I am. I don't really have anything to test but it would be fun to meet up. My email address is linked in the last paragraph this blog post: https://blog.buildbotics.com/is-camotics-alive-3be83275493#....

You're doing some much needed work and I really hope you not only stick to this but win out in every way.

I cannot understate how much I have been waiting for a good UX/UI on a CAM that doesn't require running HUGE closed-source packages.

That is exactly my goal. I'm trying to figure out how to make it more profitable so I can afford to focus on this software even more.

I'm working on a device that mounts to a baseball cap and uses infared LEDs to block your image from digital cameras (at least the ones that don't have infared filtering, i.e. most surviellance cameras, which need to see at night). First attempt was underwhelming: https://sr.ht/SBGV.jpg

Currently obtaining much more powerful LEDs. Next version is going to need cooling and an external power supply :D

Does wetware count?

I'm rebuilding engineered bacteria which can take pictures [1]. I'm using new proteins which will make the bacteria more reliable and use cheaper materials to take pictures. I hope to create some pretty sweet looking images to hang on the wall.

[1] - http://www.nature.com/news/2005/051121/full/news051121-8.htm...

I'm trying to design and build a small direct-drive robot arm, inspired by this cool post:

https://news.ycombinator.com/item?id=12536041

I haven't worked with electronics since college, so I was trying to find a suitable project.

I unconscioussly bought some components and a couple of CHIP computers and I wanted to do something that interacts with the real world (sensors and actuators), but servo-based actuators seemed a bit lame.

Initially, I was going to use two motors and a spring per DOF in order to simulate a muscle ("variable stiffness joint", according to the previous post - I was going by intuition).

But then I read that direct drive was possible (the motor's magnetic field acts as the "spring") and that once that you have a suitable motor controller built all the rest is software defined, and I found the sweetspot for a project.

Such a motor controller is sufficient to keep me researching and prototyping for some time!

If I succeed with this, I'm planning on researching how to do some sort of artificial skin for robots. For the moment the idea revolves around connecting a USB camera to a bunch of fiber optic threads from a cheap lamp such as http://prosites-lottofun9.homestead.com/files/fiberlamp2.jpg and try to detect pressure on loops of the fiber by changes in the light source. That way you can cheaply and efficiently have a huge number of data points, with off-the-shelf components... but if someone does it beforehand I'll also be happy!

I'm adding flow meters to my keggerator so I can detect when someone pours themselves a beer from one of the taps.

The plan is to have a camera take a photo, do some facial recognition to see if its either one of my flatmates or myself, then optionally post the pic to #beertheif on twitter or start playing some music based on the chosen beer.

That sounds really fun :) I've heard good things about the Swissflow flow meters, for monitoring beer flow to the tap, but they're _really_ expensive :(

Are you building your own flow meters? That's a fun project in magnets/magnetometers.

I've been looking at which flow meters to get but as deutronium says swissflow meters are quite expensive, however at the same time I don't really trust cheap meters to be in contact with my beer. Making them myself might be a good project!

My buddy has something going along the same lines but is using temp sensors to detect a pint being pulled:

https://mobile.twitter.com/toms_beer_keg/status/799323771043...

Two other non-contact methods I've seen are direct mass measurement (probably ebay the scales, it'll be like $500 new) or the most innovative thing I've seen is flow rate meters watching the CO2 source. See you can measure flow rate of the beer but that can get weird with bubble formation and sterilization but CO2 is clean and biologically boring and single phase. You can buy a flow rate meter with digital output for immense expense but its easier to use an analog one with a webcam. Big fun!

Measuring the co2 flow rate is an idea I hadn't considered! I only want to know when beer is being poured and not actually how much has been poured so measuring co2 may be a bit simpler for cleaning and actually wiring up too (although the wiring issue is specific to my setup)

Presently, none, but I have had a few ideas I've been wanting to get started on when I have more time for a hardware side project.

The first one to come to mind is an LED lighting setup that shift's it's color temperature according to the time of day, à la f.lux[1] or Redshift[2]. LED continue to improve for general lighting purposes, but it can't be good for my sleep patterns to have daylight temp lighting in my home right up until bedtime. It would be really nifty to the lighting in my home adjust to warmer temperatures with the solar day like my computer screens do.

1. https://justgetflux.com/

2. http://jonls.dk/redshift/

Does full time but fun count? We're building a network of automated food preparation and retail service locations based upon robotics, to initially launch in mainland China: http://8-food.com/

The idea is the consumer can buy from their phone and pick up customized meals. Currently we're iterating functional prototypes toward ready for manufacturing while tweaking our form factor and components and increasing our ingredient volume and breadth.

We'd consider bringing on additional founders with strong mechanical engineering (CAM/materials science/foodtech) experience.

deltabot - https://www.youtube.com/watch?v=cbNiJKSRCpA

digital theramin - https://www.youtube.com/watch?v=bZxVEUxaw1w

3d printer - https://www.youtube.com/watch?v=QFOR1YPwFMM

stacked focus camera rig - https://www.youtube.com/watch?v=6ovWjnzBGUw

more stacked focus rig - http://imgur.com/a/SMLx9

more deltabot - http://imgur.com/a/05G1L

more theramin - http://imgur.com/a/lvLE4

slayer exciter - http://imgur.com/a/5Rwtm

mame cabinet - http://imgur.com/a/hzE1x

polywell - http://imgur.com/a/Tsg40

I have a friend who builds kite buggies out of carbon fiber. Amazing, gifted inventor and craftsman, zero love for computers. Going to build a raspi based automatic tempering and ventilation system with him as well as seeing to getting him some crowdfunding.

Other than that, I really want to turn my, ahem, indoor flower growing project into a test environment to familiarise myself with sensor based programming. Thinking something like a bubbleponics setup with automatic nutritient dispensing and stuff. Zero human input being the goal.

CNC machine shop. I rigged a Sherline mill and lathe with stepper motors, limit switches, electronic touch-off and emergency stop button. I multiplexed the limit switches with a custom designed PCB with Schmitt triggers and an Arduino. I hooked this all up to a stepper motor controller board and power supply, and I placed all that into a custom built enclosure. I connected the controller board to a PC using PCIe parallel port cards. The PC runs LinuxCNC. This lets me machine alumninum into parts modelled in 3D CAD. Then I've machines various tools needed around the shop. I've also made a Turner's cube using the lathe, and again using the mill.

I've also built a custom PS3 style controllers for manhually controlling the CNC machines. This was again done with custom PCBs, joystick and encoder parts, and Arduino.

I'm working towards machining a Stirling engine with my CNC machines. I've also been working on and off on a robotic arm (which is what got me into machining metal parts in the first place).

Unrelated to CNC machining, I've done a small, beginner SDR project: ADS-B receiver built with R820T2 USB SDR and custom built antenna from coax cable and PVC pipes, lets me to track commercial aircraft that frequently fly over my house.

Not yet, I'm in research / learning / pooling resources for pcb fab on a desk, all-in-one. I've done CNCs before as a hobby, film scanners and stuff like that and I have some concepts which I think could yield an 'affordable' all-in-one pcb. You give the machine gerbers, you get (a few layered) pcb out. Maybe even a pick and place inside, but that's further down in concept. One step at a time, as time permits.

Out of curiosity, how are you planning on handling plating for vias and through-holes?

The desire to stop handling chemicals (I have no convenient means of proper disposal) has got me using the on-line prototype services for anything multi-layer (though I've thought about using my CNC mill for some simple single layer stuff).

That's the thing. I have to build several prototypes first in order to see which will be less hassle. Initial idea was to have the machine handle chemicals and a contained disposal system. On the other hand I could do it with hollow 'jackets' (that's how we call it, grommet is probably the proper name).

I have to balance two ideas into a concept first and then prototype. Those two primary ideas are:

1) have end-user buy as basic things as possible in order to produce PCBs (not rely on my chain of supply / custom shit). This is in order for machine to be as usable as possible, no matter where you live and to keep costs for you down as possible.

2) Have the workflow as hassle-free as possible with the best possible result.

Sometimes, those two ideas are clashing. Ultimately, it would be a machine built out of commonly available parts. For most parts, at least. And, I would provide pre-built machines as convenience. BOM for protos is already at around 7k-10k Euros, and it will take some time to build and refine on what I'll see from protos. Ultimately, I want a sub 2k Euros machine ready, out of the box, experience for end user. Maybe even less if I opt for smaller FR4 boards. From everything I've pooled so far together, on paper, it's doable. I just need to find a bit of time. I already set some time and cash after May next year just for this (after I finish a TV series I'm working on - my day to day job).

In my opinion, I think an affordable desktop PCB factory could open a lot of possibilities for people. If you had a machine and only things you would need to buy are FR4 boards and some commonly available and cheap chemicals and have the machine 'print out' a pcb for you, maybe even with parts placed, and have the exhaust for fumes and disposing system safe and easy and cheap... We need this. I need this.

I'm developing an open hardware microwave vector network analyzer [1]. A previous switched single receiver prototype have worked for measuring S11 (reflected power, useful for antennas) up to about 4 GHz. I'm currently working on bringing up a full two port four receiver version that should work up to about 10 GHz.

[1] - https://github.com/loxodes/vna

My garage project is cleaning out my garage so I can do a garage project. :-( In this case, a little library for the inlaws and cabinet carcasses for our kitchen.

DIY Home Automation, using a mix of Raspberry Pis (for "bigger things") and Atmega328p micro controllers (same chip used in the Arduino Uno's) - Goal #1 is to have heating controlled by this lot..

So far, I've the heating relay control board done (16x 240V channels, controlling boiler activation and zone actuators to open the flow to different rooms). The on-wall thermostat PCB's are currently being manufactured, once they arrive I'll solder all the components in place and get a crude mounting done.. Then it's time to design the enclosure.

Once all this is done, I'll move onto the lighting as Goal #2. For these, it'll be the ATMega MC's, some RFM69 RF modules, and either a simple relay (for on/off lighting) or some as-yet-to-be-determined dimmer circuit. The challenge for these is, there's A) no neutral wire in at the light switch in EU (or maybe just Ireland/UK?), and B) it's 240V supply or nothing - unlike the thermostats, where each is wired separately back to a central point, and can be swapped over to a 5V supply, these will need to drop the voltage on board using minimal space (much less than a typical 5V wall-wart).

Home automation with old industrial hardware: an Allen Bradley SLC 503.

I read and write data to the SLC with a python script on a RaspberryPI, which also serves up an interface to the SLC and historical data using Flask and MySQL. I've got the thermostat, a door switch, and a few lights hooked up.

I VPN in when I'm on the road just to keep tabs on things. I've never enjoyed how 'closed' commercial home automation is, and this was the result.

A very trivial implementation of TCP running atop RS232<->USB converters connected back-to-back so you can run SSH over USB (or RS232) at around 3Mbps. For tunnels and stuff.

5000 LEDs drilled into 30m^2 worth of plywood, wrote a kernel driver to present the LEDs as a frame buffer on an RPi with Chromium, we (small team) mounted whole thing onto concrete slabs standing in a lake. Another RPi receives SMS text messages using a JS script to decode GSM PDUs with Unicode that are ultimately rendered on the LEDs. Nice graphics are added on top if you write something sweet for example.

A device to virtually switch an SD card out of an RPi and onto a host computer so you can remotely format the SD card and reboot the RPi.

A crawler versus human detector that is a bit simpler than those normally are.

Lots of projects that get tossed because they require expertise outside of my own area. I'm quite bad at finding other tinkerers to cooperate with, unfortunately.

Have you seen Linux's own IP over USB support? No need for an RS232 layer. http://superuser.com/questions/593757/the-best-way-to-do-tcp...

You could even combine this with multiple cables to multiple ports (eg. 2x cables to 2x ports on each node) using the bonding driver.

Nice projects! Do you have any photos/blog post/etc of your LED stand?

A lighting desk for lighting live bands. The other desks on the market don't support the busking that this is designed for. The aim is 'no friction between the music and your lights'.

36 motorised faders, 2 DMX universes built in, up to 8 more with ethernet adapters. uses two Beaglebone Blacks and a Raspberry Pi 2, as well as 42 Atmel ATTiny828s. GUI code written in Python using pyside, and the other stuff written in Go.

No blog, just a single facebook page: https://www.facebook.com/ManyHandsLighting/photos/a.79431027....

Getting a drone to play ping pong.

More specifically using a tiny indoor quad to intercept a ping pong ball off the bounce. While I doubt it will be able to return the ball reliably, it could be used for beer-pong to defend your cups from a bounce shot.

The high-level idea is that something like Reinforcement-learning / openAI trained on a physics simulator should then be able to learn how to inctercept the ball, then be applied to the drone IRL. Right now I'm stuck on the hardware hacks: using an arudino to input control commands into a the hand-held controller, and using a mirror on the table to get a 3d position on the drone/ball.

Reply if you're interested, have ideas.

Cool!

How does it hold the paddle? I imagine a simple string on a winch might be sufficient to create a pendulum effect, so the drone can move side to side to swing the paddle, or reel in the winch to change height.

I've been making little speakers out of old suitcases and boxes. Similar to what www.theboomcase.com does but not quite as fancy/powerful. I'm not trying to sell them but its fun to give them to friends as gifts and its enjoyable to work on at our makerspace with the kids.

I like finding sale components on www.parts-express.com and amazon. You can get small amps for $10-20. I've made them with a rechargeable lead-acid battery. Next goal is to figure out how to get a bluetooth module in the mix (right now they just have an aux port). I'm admittedly bad at audio electronics but I'd like to figure out how to use more powerful amps with a battery.

Sorry, no pictures on me at the moment.

I am working on programmable J.A.R.V.I.S.

Currently, it could help with stress reduction, recognizing heart illness or overtraining detection. It's also an "open" device. Open to communicate with, and open to create apps full of health information for anything that supports Bluetooth 4.0+. Soon, I'll extend Aidlab with simple A.I module and gesture recognition, so mobile phones or the smartwatches will not be mandatory.

Site: http://www.aidlab.com (`For Developers` might be very interesting for YC community)

Blog: https://medium.com/@Aidlab

I've put together illustrations for each habit I'd like to keep, and then printed them on archival paper. I also built a LED calendar to keep track of things and combined it all into a wall frame. I occasionally write project logs on Hackaday[1].

It's fun to work on and I've noticed I'm already keeping up on these habits since they are always on my mind when building. Also the sunk cost has an effect on me to want to get as much out of this project as I can.. even more incentive!

[1] - https://hackaday.io/project/18304-21days-habit-tracker

Tickerstorm.io - An algo trading platform in Java (pretty early still).

Difference from other platforms I've found: integrated machine learning phase which reuses the same data pipelines used by the algo when backtesting or trading live. Trying to reuse as many open source frameworks as possible to reduce the amount of new things a user has to learn. Based on Apache Storm, H2O machine learning, guava, Spring Framework etc. Not HF focused.

Most other platforms start off with you coding up your rules and allow you to back test those assumptions. However, this platform also incorporates the discovery phase of these rules rather than just assume you already have a set of rules.

I recently made a tiny laser turret that's controlled by a video game controller. https://www.youtube.com/watch?v=wlMuXKNZ748

I'm building a USB-connected OPL3 synthesizer. I'm using an Arduino Nano as a command processor, so that the host computer sends register change commands, and the Arduino manipulates its output pins accordingly to actually send the data to the hardware.

When I've got that working, I'll make some modifications to Dosbox, to make it pass through Adlib/Adlib Gold commands over serial to the device.

Minimum Viable Synth. https://github.com/kbob/MinimumViableSynth

Today's progress was (a) receiving my SMD rework station from Amazon, (b) continuing to wait for OSH Park to send PCBs, (c) writing graphics primitives for the LCD display (which I can't see until I make the PCBs).

Using a BeagleBone Black to read current sensors and send MQTT messages whenever an appliance (washer/dryer) is turned on/off.

Which current sensors are you using? This sounds interesting.

I did something similar a couple years back except thermally by monitoring the dryer exhaust temp and ran into a problem where apparently "permanent press" means cold air for the last 10-15 minutes. Ok then that meant a timer initiating when the temp drops below whatever. I would imagine this will show up in your current monitoring where the heating elements turn off but the tub keeps spinning. Of course you may be monitoring a gas dryer LOL. Good luck it'll be fun.

I have been working on my Anybar implementation on real hardware, so it works on everything and I don't have to worry about any particular DE or OS:

http://www.tnhh.net/mobile/posts/led-strip-for-your-computer...

Makes a good swag on my desk, too.

I've got one of those IKEA Dioder strips, and I've got it hooked up to a RasPi. There's a node library (of course) that lets you set it to any RGB color and even supports transitions.

Next step would be to make it flash when messages come in - different color for Facebook Messenger, WhatsApp, etc. - and having it integrate with the Philips Hue and Alexa.

I built a 5-octave, 31-note-per-octave just intonation keyboard with pressure sensitive keys.

I've been working on putting together a thorough writeup, but for now there's just this video which I uploaded today:

https://www.youtube.com/watch?v=Ep52Vh6oAOE

A ultrasound imaging dev kit for academics, teachers, researchers, and hackers : https://kelu124.gitbooks.io/echomods/content/

Just recently added an ultrasonic presence sensor to my computer. http://www.michevan.id.au/content/are-you-there/

I'm trying to build a camera that uses the sensors on your phone to help you take the perfect picture. Your mobile phone knows so much about it's environment, yet most photography apps only take into account what the lens sees.

A Flying Car. Anyone with CAD, CFD, aerodynamics skills wanna help?

Ambitious! What kind of propulsion system will it have? How far did you get at this point? Where are you based?

Electric propulsion. I have some very promising aero designs that I hope to get into a (virtual) wind tunnel. Bay area.

me and my friend are working on something similar. We prefer to call it a drone though :$

My friend was looking into the material strength and feasibility while I was looking into the software.

We intend to use a 3d printer to build a small scale model and progressively keep making it bigger.

We would love to collaborate if that's ok with you.

Based in Canada (Toronto) and we put ~10 hrs/week max on this project.

The only person I ever saw get a human off the ground was this guy https://www.youtube.com/watch?v=soxxPyaAT1k and several build vids. He would have really benefited with some fly by wire help

that + [0] + [1] inspired us to get started...

[0] - https://www.youtube.com/watch?v=0Yn2uyQJ1jc

[1] - https://www.youtube.com/watch?v=PAdCt6qgj9k

Terrafugia is awesome. They have tremendous respect from me.

But I dislike any type of folding wings. They will need to make a massively strong hinge/joint for this type of machine since the entire car hangs off that hinge. This makes is heavy, inefficient, painful to maintain, vulnerable in fender benders, etc, etc.

Sounds good to me - call it what you like. Perhaps we can start a conversation. I'm skanga at google's email thingy.

I learned some FPGA on my graduate course, while my mummy do my lunch I try to create an ADC to create a radar so I can jack off without my mummy catch me

An AOA (i.e. Angle Of Attack) indicator based on GPS and accelerometer data. Arduino-mounted.

I built a barn-door electric drive astrophotography rig for long exposure star photos.

A completely numeric orrery for lack of a better term.

Reply on news.ycombinator.com