Electronics Projects (by alaric)

So, my electronics workbench is a mess.

This is abundantly clear in the picture from my blog post on redesigning my workspace; the awkward layout is certainly part of the problem, but a deeper problem is that I don't do many electronics projects. So this big workbench is rarely used for its intended purpose, and thus accumulates junk, and thus isn't very inviting to start projects at, which adds to the fact that I'm a bit edgy about STARTING electronics projects, and a vicious cycle has set in...

The only electronic projects I did lately were the 12 volt DC power distribution system for the van and a 9:1 impedance transformer, but those were mainly mechanical builds; the electronics were trivial.

The problem

I tend to find that, with electronics projects, I have a sort of intermediate hump that I often get stuck on. I come up with an idea, I do some research, I sketch a circuit diagram - but then I have to order actual components and decide how to fit them together, and I hate choosing electronic components. There are often thousands of components that meet the main requirements, and I worry too much about picking the right one, and it's just tedious cross-referencing the web sites of distributors, PDF datasheets, whether I have a footprint for the part available in KiCad, etc. This stage is terribly boring, so I often stick there, with a schematic but no PCB layout.

This is BORING and this must END. I hereby resolve to:

1. Make my new desk setup. SOON. I have the money to buy the materials required but I'm holding back a bit until some more urgent spending has completed successfully without any surprise extras, to make sure I have the money. With that built, I will have a clear work space that is used for electronics AND computer work, and is laid out so that nothing other than the tools I'm currently using will need to be on it, and it will be easy to put away a current work in progress to do something else, so it will be easy to resist the temptation to dump junk there and forget about it.
2. Actually breadboard stuff. I fret about choosing the exact right components because I'm going to be soldering them to PCBs and everything has to go together properly, and that's a mental overhead I keep catching on. But if I prototype more things on breadboards I can:
   1. Check it all works, without fretting about double-checking everything on paper.
   2. Confirm my component choices are OK, THEN design a PCB to solder them to.
   3. Check the component footprints with a measuring tool to confirm the PCB design, rather than cross-referencing hateful PDF datasheets.
   4. Get some immediate "Hey, it works!!!" excitement to carry my enthusiasm from the initial "Hey, I have a cool idea!" on enough to last me through to actually finishing the damned thing.

So, for the record, here's a list of my "current TODO" electronics projects stuck in development hell that I mean to progress:

1. Make a little audio mixer for my new desk. I have a four-way HDMI/USB switch for two PCs, a laptop, and something under test (eg, raspberry pi project in development), but I have to separately crank the input selector on my audio amplifier to select an audio source. Many years ago, I had a small mixing desk next to my computer, with proper sliders for the channels. I had my PC, minidisc player, CD player, and a radio receiver plugged into the inputs, and it also had a microphone pre-amp and a line output that I fed back into the PC's line in, and separate headphone and line outputs, and it was pretty cool to be able to mix audio sources like that. Sadly, it was huge, and it eventually died of old age; my needs now are simpler (the PC is a fine CD player if I ever need one, I have no minidiscs any more, and so on). So I think my needs would be met nicely by a four-channel stereo mixer, but I want it as small as possible - ideally little larger than four knobs for the four input levels (there's no need for a master level output, as my amplifier has a nice big volume knob), with four compact 3.5mm stereo line ins and one line out on the back. This should be an easy project for five stereo op-amps, four stereo potentiometers, some resistors to set up the op-amps as a mixer, and some capacitors to decouple the inputs and outputs from any DC offset. I'll run it from 12v as I'm working towards an integrated 12v power system in the workshop. STATUS: DONE!
2. Start building a CAN bus in the workshop, beginning with thermocouples inside the rocket mass heater. When I made the RMH, I embedded four K-type thermocouples in the system - one at the top of the primary heat exchanger, one between the primary and secondary heat exchangers, one embedded in the middle of the thermal mass, and one in the flue just as it leaves through the roof. I can read them individually by hooking up a multimeter, but that's a pain; I want to have them read digitally, as part of an integrated workshop automation system that will also control lights and other stuff. So to get this started, I plan to just wire together some off-the-shelf modules (Arduino mini, CAN<->SPI bridge, thermocouple->I2C interface) to make two units. One will have a single thermocouple input, as it goes at the far end where the flue is (and where I'll want a small control panel by the door in future, so it can go in a box designed to become that in due course), and the other will have three thermocouples and possibly a small display, as it's by the business end of the heater. They will be joined by a CAN bus and 12v power distribution via a twisted-pair cable (UTP's 100 ohm impedance is close enough to CAN's 120 ohms, I think, but experiments will confirm that), which can then be extended to join other things too (such as a display/control panel embedded in the new desk's top, perhaps?). The CAN+12V cable network will be easy to route around the workshop, and enable everything to integrate together beautifully.
3. Finish the environment control system for the home mainframe. There are a bunch of temperature+humidity sensors inside the mainframe's airflow path, but there's nothing connected to them. I have designed, but never finished building, a little microcontroller board that will read them all, and also drive three RGB LEDs that will go into holes in the chassis that have lain empty for years now. I've written the firmware for the micro to report sensor readings back to the mainframe, and accept from it instructions as to what colours to illuminate the LEDs in order to summarise system status and provide some nice blinkenlights; and if the mainframe doesn't update the LEDs for more than a couple of seconds, it will flash them all in an angry red pattern to indicate that something's gone wrong (and I even have some pins spare to attach a small relay that could pulse the reset line on the mainframe's motherboard if there's no activity in too long).
4. Build the workshop DC power management panel. Currently, 12v power in the workshop for my lighting is provided by a mains-fed high current linear power supply I built for the purpose, but this was only meant to be temporary. I want to build a box with a power management system inside, capable of turning the mains-fed power on and off, while managing the charging and discharging of a lead-acid battery that actually runs the lights (and future electronics payloads via the CAN bus) - all under microprocessor control, and in future, also incorporating power input from solar panels on the workshop roof so the mains input is only needed in extreme circumstances. I have a nice wall-mounting box for this, already drilled with holes for the volt/amp meter modules I've purchased, which also has space for a raspberry pi that can bridge from the CAN bus (which the power management microcontroller will of course be part of, as it can report on the power system and do things like turn the lights on and off) to the world of TCP/IP, and drive a little speaker via a speech synthesiser for alarm noises and announcements.
5. Fix the raspberry pi in the living room, which USED to drive a speech synthesiser via a speaker, but I had to turn it off because the noise leakage from the Pi to the amplifier was too annoying.
6. Get back to work on wearable computer research, and ancient passion of mine! Starting with a hiking staff with an integral LiPo battery stack and lights, head-mounted computer display with chord keyer input, etc etc...

I've also got a few in-progress electronics kits to finish assembling, but that's separate - I don't get the same mental block around them, as they come with all the right components already laid out 🙂

Wish me luck!

1 Comment

• 

  By andyjpb, Mon 9th Nov 2020 @ 6:37 pm

  Good luck!

Other Links to this Post

RSS feed for comments on this post.

Leave a comment