Hi!
Just looking for assistance as you guys have helped before, and were SO much helpful than Reddit was, and without the condescension.
It was a giant wearable set of light up wings for anyone who saw it
Since re-registered with the account I buy stuff from core electronics with.
So the project I am working on is a fairy skeleton inside an old candle iron lantern (PLA), on a large hooked cane (that will also help me stand up on digistilts).
There is lights underneath that will flicker in the normal fire spectrum
And when someone whispers to her, the lights will change to blue/green faster moving lights, and a speaker will play back creepy things in Gaelic (I have a 7 year old eager to record)
Space is severely limited in this, which means the batteries will be about 50-75cm away by necessity, and I am using 18awg twisted wire for that.
Battery run time only needs to be about 4 hours, and battery changes are possible
In case labels are missing, 18650 batteries, MT3608 5v boost converter, Arduino nano, DFplayer, MAX9814, PAM8403, 10k ohm potentiometer.
It includes charging points for the batteries, but I will probably scrap those and design the prop so I can just replace batteries.
The speakers are 1w, so as far as I am aware multiple components running off the arduinos 5v should be fine. Plus, I want them to only be whisper quiet I shouldn’t be going over the 400 limit. Please correct me if I am wrong.
Does everything look okay?
Anything obvious I am missing or should be concerned about?
Looks good from a brief overview. There are a few places where you are assuming that the input and output grounds of the modules are not isolated: it is unlikely that they are, but worth a quick check. I’m not sure whether the LEDs will maintain their latest state when the MCU powers down - if they stay on they will drain the second battery. If you are removing the batteries for recharging when not in use, then that’s not a problem. The tricky bit might be getting the Nano to respond to a whisper, as distinct from all the other noises that could be happening, especially as the discrimination might require sampling over a relatively long period while the lights are still being driven. It might be necessary to mount it in a way that it won’t respond to vibrations in the staff (for instance) but will respond to a breath.
Because whispers take advantage of the resonant frequencies in the back of the throat and mouth they usually have really clear peaks in their audio spectrum. The exact frequencies of the peak change from person to person, but are statistically likely to be distinct from the wind. Not easy to get right, but maybe a direction to look.
I had a quick read of this short article from UNSW. I
I’m not an expert in phonetics but the ideas I wanted to communicate are there with much more precision.
It’s useful because it throws some numbers out there you might be able to use as starting points.
More a question then a comment. PLA do you mean the lantern is a 3D print using PLA ? If so, just keep an eye on the LED heat. I had a play with two small ws2812b strips about 16 leds in total that I just double sided taped to some PLA and it did warp my back plate a little. Im sure giving some separation and/or some stiffer support ridges would have helped. i.e. it was not so hot as to melt anything, just enough to allow the plastic to deform a little. Also if you light levels are low, then heat will be low, so ignore me
Thanks. Would you mind explain what you mean by checking their grounding. At the moment all grounds connect together…
Are you saying that some components I need to ground elsewhere as well?
With the batteries and the lights, the normal setup I use for most wearables with lights and something else (usually servos) I have a 9v for the Arduino and a block of 4 AA for the lights, wired like that, I have never noticed a battery drain and the lights turn off when the Arduino does. But also, as you said, I generally change every 4 hours or so and no batteries are in when not actively in use, so I probably wouldn’t notice.
What’s the best way to fix that?
@Pixmusix thanks for that! The voice activation is kinda of a play around for something I plan to do later.
Vibration from the staff is definitely an issue, I plan to weight the bottom of the staff with magnets, and have somewhere on the digistilts to “lock it in” when people talk to it.
The goal is to have it respond to it’s name spoken about 55db in a 75db room… If I end up with it responding to any deliberate input from 15cm away in a 60db room
And if the whole thing fails I have a motion sensor I can use to similar effect
@Michael99645 yeah, printed in PLA, and then resin smoothed (so like a 0.5mm thick coat of resin) the 6 LEDs won’t actually touch any of the part, it will be suspended, plus there will be airflow from underneath and through (because I made a tiny whoopsie while cleaning it up and making space for the Arduino)
Some modules, particularly buck or boost circuits, are built so that the potential of the ground at the input side is not the same as the output side - the two are isolated. This is useful in some installations, for instance where negative voltages are needed. It is not common, but not always documented. Just check that there is zero resistance between the ground on the input side and the ground on the output side, to confirm they are common.
The LEDs not turning off? Use a double pole switch and route the LED supply through that as well. But now I think about it, the time that the LEDS stay on is when the sketch is halted but the MCU is still powered up, so I think you are correct to say that the LEDs will not be lit when the MCU is not powered.
It wouldn’t work for this particular build, but I am dropping the charge module.
When I have used battery packs I always run into issues (a 12v that would only give 5v, or it reversing programmed light colours, other weird stuff)
Most importantly for wearables, (some costumes have 8 batteries in different places) nothing wearable has capacity to last 8-12 hours, and having a few 18650s or 9v in pockets for quick changes is way easier than having to be stuck to a wall for 2 hours.
The charge model was to best of both worlds it, but feels like more effort than 2 batteries in my pocket
