Group Actions
Engagement
  • Author Author: djfraz
  • Date Created: Date Created
  • Views 1206 views
  • Likes 5 likes
  • Comments 7 comments
Related
Recommended

Smart Home: Party Ball, Kind of

djfraz

So the I have gotten the code to function as intended. So now for the hardware.

 

I started by depopulating the pcb that came with the original product. My original plan was to make an acrylic version with just the mounting holes, but realise if i flipped the pcb over, it would have the same effect.

imageimageimage

The only component which gave me any trouble was the inductor, the contacts were soldered under the ferrite core, and in the end i just snapped it off, as i was not worried about keeping the PCB traces.

 

For the neopixels iI purchased some breadboard friendly ones from ebay (same price as most online retailers and no shipping cost) and simply hot glued them into place. The was to be able to replace them if needed.

 

And that is really all the hardware, the motor was connected through a transistor with diode protection. And it was ready for its first test.

 

image

Everything appeared to be working, so I turned off the lights, closed al the blinds ready to see it in the dark, and;

image

The picture has done it some justice here, but it is very dim in person. and for comparison, I purchased another of the original for comparison, and turned them on together. All the light is from the original. I have included a short video to greater illustrate the point, this one was with the lights on.

image

 

 

 

It is fair to say that I am a bit disappointed at the moment. I though that the Neopixels would be bright enough, any of Adafruit videos show them being very bright.

 

I think the next step is going to be looking for higher powered RGB Addressable LEDs, somewhere in the region of 1W, as i think that is the value of the ones in the original. If anyone know where i could get my hand on them it would be much appreciated. I was considering buying normal 1W RGB LEDs and the Driver IC WS2811, but i think they are surface mounted and I am not wanting to go down the road of getting a custom PCB made for this.

  • in reply to 14rhb

    It has definitely been V1.0, it was not so much a fail of the design, more of my expectations. The process of designing and programming this has been rewarding in the knowledge i have gained, which i will be using towards other projects. I will continue to work on the project, incorporating other features into the software to begin with and looking into ways of increasing the brightness of the leds.

     

    I had not though too much into the luminous flux density of the original leds or how they compare with the Neopixels. Looking at the original design, i would suspect that this is the case.

  • Duncan,

     

    There is no such thing as a failure, if you keep trying (then it is just an experiment).  You have made some great steps towards your goal, so keep moving forward.  There are some good suggestions  here (thanks to ,   and ), so step back and plan what to try next.

    Good luck!

    Gene

  • Hi Duncan,

     

    I wouldn't say your design is a fail as titled in your video. It looks good, but from reading your article you are disappointed with the brightness - maybe this is just version 1 in development, lessons learnt and therefore definitely not a fail? image.

     

    The single LED on the original board do look like they are higher power than the Neopixels - I see you mentioned that in your teardown, assuming they were 1W....which is probably correct. The thing I notice is that they have a clear lens and probably focus that power into a narrower beam angle than your Neopixels. So you probably loose on luminous flux density due to both the lower power and lower beam forming.

     

    You may have already mentioned this and I've just missed those comments - my apologies if that is the case.

     

    Rod

  • in reply to jc2048

    I was wanting to use addressable RGB Leds to be able to expand on the functionality of the original device. So, if I decide that I want all the lights to be purple, then all of them change to purple, instead of just seting the three Leds to their corresponding RGB value.

     

    It was also just to experiment. I may end up replacing the original LEDs and driving them together, as I plan to use and esp01 with only two gpio pins, and the motor control is a nice feature.

     

    I am still undecided how to proceed with this project, but it is for more of s learning experience than for the occasionally use the device will see.

  • in reply to fmilburn

    It looks (from Duncan's previous tear-down) like the case of that product is the heatsink. It would be much better with a metal-clad board, but it's nice that the heat is radiated off the exterior and not internally into the case.

     

    It's possible, looking at the board, that the person who originally designed it intended it to have an mcb and it's been subsequently cheapened (probably on the basis that something that runs for about an hour on three AA batteries probably won't be run for anything like 60,000 hours to test the LED lifetime).

     

    Once you get to 1W, there's a good case for a constant-current driver rather than resistors.