What do you want?

I'm an electrical engineering student with an abundance of free time and no job. I have experience in electronics design and fabrication, and I can write firmware like nobody's business. I'm offering my time and effort to design PCBs and circuitry for any ideas you guys think might actually work or be useful, then distribute the schematics and parts lists freely.

So, what can I help make real?


  • One thing I could make that might be interesting would be a magnetic amplifier, something that used a three axis magnetometer and three coils, and just spit out the detected flux at a stronger level.
  • edited May 2015
    That would be sick. How big would such a device be?
  • Ideally, not big. I'd picture a glove with a lighter sized enclosure on the back. The biggest part would be the battery. Running the coils would take a good bit of juice.
  • I want to make that magnetic ring which would let you "feel" the music (already talked about it in different thread) Seems you just need magnetic wire coil and amplifier, so I would say it should be easy. It's basically one of these "invisible headphones", but smaller.
  • Industrial EE here. For a long time I've loved the idea of a device which would accept input from a variety of sensors and process the information to a variety of outputs. A plug-and-play set of toys which expand the senses.

    You buy a single processing unit and on the input side you attach a distance sensor and the distance sensor identifies itself to the processor during startup. On the output side you attach a solenoid and it identifies itself to the processor. Now you have a distance sensor which pulses a coil. Bottlenose.

    Later, you decide to switch to an infrared thermometer so you can feel temperature so you pull the battery and plug in your IR thermometer. Once again, during startup, the input identifies itself to the processor and information is processed to the pulsing coil. Even later your buddy, who doesn't have a magnetic implant wants to try so you shut down, replace the solenoid with a vibrator, start it up and your buddy can feel temperature through vibrations.

    This project could get big really fast and would be severely prone to scope creep. But if it is planned out in advance it might be a valuable tool. I can think of more than a dozen inputs and at least five outputs.
  • Okay, so current thought: build a base platform to expand upon.
    I want to use a set of 4 coils arranged in a tetrahedron that's embedded in a ring or glove or band of some kind that sits over your implant location and provides 3D spatial representation of fields. I would probably use a Teensy for the backbone with 4 decent DACs to drive the coils. The coils I would scavenge from headphones, I think. I need to know how much of a field it takes to get a response from an implant though. 

    The system would be fairly easy to make extensible, I could make a stackable boards idea where each board is basically a carrier for additional parts with a communal I2C bus and power lines. This would be pretty much plug and play with a huge amount of configurability. 

    Again my main concern is battery life. If I can drive the coils at a power level such as playing loud music through the headphone coils, I'm good for a few hours. If it takes a lot more, there might be some issues. 

    I need some details about how strongly you can feel magnetic fields. Can you feel static fields or just changing flux? Can you feel the field from an earbud placed right next to the implant? Can you feel the field from a computer fan, and at what distance?
  • I picked up a magnetic ring to play with, I'll use that as a baseline. If I can feel it on the ring, you'd definitely be able to feel it with an implanted magnet.
  • Got got to say the modular platform is a great idea. I dont have any hardware experience, but if you need a second set of hands for software I'm happy to help. 
  • If the battery life of an iPod (gen 3?) is anything to judge by, you can play music over headphones/built-in speakers for over 12 hours at full charge (tested using headphones with 30mm drivers at close to max volume). 

    Would running an inverted form of your signal output to 2 of the 4 coils opposite each other in the tetrahedron, with one inverted and one original signal feed to each coil "pair", enhance the sensation, or would that just cancel the signal out entirely? I'm thinking of it like a pair of speakers in a similar configuration. When one speaker cone is moving inward, the other is moving outward. Except in this case, there's only one object being moved (the magnet), so every signal gets "doubled".

  • It's a tetrahedron, not a cube, so none of the magnets are directly opposed to one another. If they were, you would want to run the signals with opposite polarity, so one pushes the magnet while the other pulls it. With the tetrahedron, I recreate a field by varying the field strength through each coil respectively, and I will run the opposing side at opposite polarity to increase output strength.

    @meanderingman How familiar are you with arduino?

    I was thinking that for each shield style board, it could carry an EEPROM with drivers to run the hardware, in a very plug and play manner. I haven't actually given much thought if this is possible, but it's an interesting thought.
  • That's more or less what I meant. Just was thinking of a square pyramid instead of a tetrahedron. -cue facepalm-

    If I understand the plug and play methodology, it's designed so that a central system can have accessories attached. From a hardware and processing standpoint, how do you intend to distribute the "data load"? of the system? Have the various "shields" process the data they recieve and convert it into a simpler data stream for the main board? Or feed most of the data to the main board and let it do most of the processing? 
  • The idea would be that each shield would have a storage device with the code it needs to run, and at startup the processor would aggregate this code into one main loop and run that.
  • I've been reading into it, and the best way it would work, imo, would be to have the eeproms loaded with what essentially amounts to a script. The processor would be running something like FreeRTOS and would check each shield and create process threads for each and proceed to execute. Not the most efficient method, but it is probably the most user friendly.
    Just plug in a shield, reboot, and it starts spitting out data.

    I'm thinking a thin base board that connects to a battery, and has either linear voltage regs or a buck-boost regulator that leads to the power rails, and a lipo charging chip so the device can be easily recharged off of a USB socket. It would also have a power switch.

    The next layer would be the main core chip, which would have the teensy, a status RGB LED and some buttons.
  • Other than that all the boards would be inputs or outputs. Vibe motors, inductor controllers, visual displays, sensors.

    My only concern would be stacking too many things, it would pretty quickly lose its portability. I think this is a good point to start with though, worry about minimization later.
  • Need an opinion here, would it be better to use readily available breakout boards (pricey, but easy to build) or a full custom layout (cheaper, but requires more skill to assemble)? If I did full custom and there was enough interest, I could build and sell the boards myself if I have the time.
  • image
    Autorouted a basic base board. Teensy, charging IC with LED, battery terminal, and two busses for shields.
  • EAGLE <3
  • Eagle is a wonderful software. Free for academic and hobby use? Yes please
  • I have a Haworth magnet and an M31 magnet. Recently I built something like a Bottlenose. Originally I had planned to flip the polarity on a 12V coil by switching a relay but I discovered that I can feel a 5V relay coil so instead of driving a large coil with a transistor I decided to scavenge the core and coil from a 5V relay.

    Sensation in the Haworth magnet was obvious but no so much in the M31. Another grinder who got an M31 six weeks ago said he can just barely feel it but enough to navigate his house. The power draw on that 5V coil is likely higher than earbuds at full volume.

    I am interested in your tetrahedron idea. Haworth magnets and M31s act differently. I can feel the Haworth magnet spinning when I wave a strong permanent magnet near it but the M31 wants to turn my finger.

    What if each component had its own processor? Instead of a base, build sensors which always output your protocol and each feedback device handles that protocol by activating the feedback mechanism appropriately. That protocol could be as simple as a tone.
  • So I'm going to take your idea and spin it a bit. No base board. Each board has a separate process or that hooks into an i2c bus as a slave, and the will provide data in a regulated fashion, only replying with the fashion it should be represented in. The output boards have a master processor that interprets the data and jiggles whatever doohickey that specific output board is meant to do. Not a bad idea, but there's still the issue of the power supply board. 
  • Each board has its own battery, but you have the ability to charge the whole assembly from a single board's charging port?

  • why all the batteries? it's not like each board do anything unless it's connected to the master.
    @ElectricFeel I bet you can do better than the autorouter, give it a try.
  • Oh I can CERTAINLY do  better than autorouter. I just wanted to make a picture. I'm not going to hand route a board until I'm sure whats going on the board and where on the board it will be.

    Do you want me to provide support for multiple outputs simultaneously?
  • @ThomasEgiThat way you can have less weight overall for each board, and you don't end up with a giant battery and higher overall weight in the baseline system. It also provides backups incase one of your batteries dies.
  • Multiple batteries means multiple charging and regulation circuits and more overall hardware. One consolidated power supply is better.
  • I suggest the battery be connected to the feedback device. If a device uses 12V for a coil you simply use a 12V battery. If a servo uses 6V you use AA or AAA. Eventually rechargeable batteries but why rush that?
  • It's simple enough either way. The question I'm on right now is do we want the circuit to be single-input/single-output, multi-input/multi-output, or half/half?

    Single-in/single-out would be very simple and small, basically two boards that snap together and go, everything would be well defined. The others make the layout more complex and require more user customization to work (what to display from what input, etc).

  • Could each receiver/feedback module have a feed-through port? That way a sensor can simultaneously feed information to as many devices as the user wants. This would make multiple batteries a bit tricky. But it might be nice to have an infrared thermometer with a vibrator AND a screen readout.
  • That depends on what communications protocol I use. If I use i2c (I like i2c) I can't have multiple masters on the same bus, but if I have a simple switch to toggle master/slave for the outputs, that's pretty doable. So one input, multiple outputs is fairly reasonable. 
  • I can agree on i2c. and maybe 3 power rails. 3.3, 5, 12v. 6 pin can be justified. Altho it would be easier to simply keep everything 3.3V.
    I also opt for a single battery. Many small batteries waste space as big parts are just the mechanics of the battery instead of the chemistry storing the energy. A single-master system should be perfectly fine.
Sign In or Register to comment.