Project: Disco Ball
 
Notifications
Clear all

Project: Disco Ball

76 Posts
20 Users
0 Reactions
44.2 K Views
Wolfie
Posts: 2225
Topic starter
(@wolfie)
Noble Member
Joined: 11 years ago
[#42194]

I thought it time to spawn this into a separate thread to outline construction and building of a Halloween costume in the form of a giant "disco" ball. While its not likely to be a true disco ball (ie mirrored ball) since I am having issues locating inexpensive acrylic mirrors in 4" squares, its still going to be called a disco ball for the purposes of this thread.

Criteria:
1. Wearable
2. Fits through commercial doorways (36" assumption).
3. Completely self contained for power and sequencing.
4. Utilize scrap 5v bullet pixels from my old megatree.
5. Able to be transported and operated by my 12y granddaughter.
6. Outer shell capable of holding the bullet pixels
7. Granddaughter must be able to don and remove it by herself
8. She must be able to walk and navigate stairs carrying it and wearing it.

Options:
1. WiFi connectivity so it can perform as a disconnected show prop
2. Remote sequence ability
3. Possible conversion to ground prop (not worn) as a pumpkin or christmas ball for future use.

General construction plans:
Weld a wire frame from 12ga wire to support an outer mesh, most likely chicken wire. The initial brainstorm here was to make an outer shell from paper mache. However we are re-thinking that aspect and may go with duct tape (couple layers probably) with a final covering of aluminum duct sealing tape (real duct tape) to give it a shiny appearance.

Initial thoughts were toward a 32" ball but its now been downsized to 30" to conserve the amount of pixels needed and the power to run them.

We have settled on a final design of 30" diameter. LEDs arranged in 24 vertical columns containing up to 11 pixels each. By going with 30", it will fit through my home doorway easily. It lightens the whole rig. Reduces LED count from 300 to 260 or less. It reduces the skin area from 3217 in2 to 2827 in2 saving 380 in2 of shell area.


75 Replies
Wolfie
Posts: 2225
Topic starter
(@wolfie)
Noble Member
Joined: 11 years ago

Initial frame construction is complete. We leveraged her backpack from last year. She had blown out the zipper and it had some hole. We cannibalized it for the shoulder straps and I rigged up a frame to attach them to the ball.

Criteria 1, 2, 5, 8 satisfied.

This was the first time I have done any welding so it was a lot of seat flying and making stuff up as I went. Its not gonna win any welding prizes, thats for darn sure. But it will hold.

Last bits: Arm Holes. Need to make some small hoops and weld those in place now that we know how it will sit on her.


Reply
Wolfie
Posts: 2225
Topic starter
(@wolfie)
Noble Member
Joined: 11 years ago

Time for some electrical details. If we use the max of 260 pixels, that will end up with a total of 15.6a at full intensity white. If I reduce the lighting curve to 75% that will leave me with a max of 11.7a. During testing of some 12v pixels, 75% produced no noticeable difference in visible brightness. 5v pixels may differ. Still to be tested. So for now I am striving for 100% brightness but leaving the option for lighting curves open for experimentation.

The fact is, its unlikely that a full 11 rows will be present since there is now a neck hole and bottom body hole present and that removes some pixels from the mix. Optionally I can reduce the vertical spacing between and still maintain 4" horizontal spacing. Still toying with that.

Power at this time is possibly a couple of my drone LiPo batteries at 14v. Again, testing is required to see longevity of them under pixel load.

The 2 bucks I ordered off ebay arrived today:
http://www.ebay.com/itm/182036212948

They are decently sized. Bit more hefty than I had imagined. Looks like cast aluminum casing and fully potted with a rubber wire strain relief wedge. 16ga output and 18ga input wire sizes so its respectable wiring.


Guess its time to hook these bad boys up and see what kind of load I can put on them and see how much heat they dissipate. They will be anchored to the inside wall of the ball so there is little chance of skin contact but if they get too hot, then I will have to re-think them and either put more of them with less load or find another alternative. I also need to put an ohm meter on the grounds and case to see if they are in any way connected.

Since these appear completely waterproof, they lend themselves to use in the front yard and feeding them right from 12v supplies when she's not wearing the ball. Alternatively, I could run AC to it and put a 5v supply inside the ball for fixed display use. I would have to put some sort of disconnect in past these and before the LEDs though.


Reply
Wolfie
Posts: 2225
Topic starter
(@wolfie)
Noble Member
Joined: 11 years ago

Ok, I have one of the modules hooked up and am testing the actual pixels. Either specs on various sites are wrong or these pixels are really efficient.

I have two 50px strands in series for a total of 100px with power inhection between the strands. I also have another set of two strands of 50px in series for another 100px. I have the two 100px lengths conected to a common power, ground and data (each strand will do exactly as the other).

Driving them to 100% white, they draw 5.3a. Measured the voltage at input and with that and the current, that works out to about 26ma per node.I removed one length of 50px leaving 150px powered at full white. The draw is 4.26a for 28ma per pixel. With only a single strand of 100 running, the meter reads 2.87a or 28.7ma per node.

So, minor deviation but reasonable for smaller wire, etc.

However Ray's pixels say 45ma on some, others were a higher rating. Unfortunately Aliexpress refuses to show me any products. Keeps redirecting to login. So I guess I am done shopping there.

Anyway, Holiday coro says 60ma per pixel.

Todd doesn't seem to have any usable specs on his site.

At this point all I know is what these ARE drawing which is very much less than what I was expecting them to draw. After sitting here with the strings on full white for over 2 hours, the module running at 5.3a is cool as a cucumber. I mean I can't tell the difference between it and the metal workbench beside me as far as temps go. Its room temperature drawing slightly above its rated ouput of 5a.

I have tested it running direct from a 360w 12v PSU I got from Todd. And I have tested it plugged into my drone battery. Works just the same. Same current output.

At this point, I don't think I will need anything else but a pair of these jewels. This one is running 200px of the projected 266 max pixels I should need. So they will easily handle the pixel load. Good news indeed.

I will leave it running overnight on the 12v psu, recharge the drone battery and test it with 150px tomorrow to see what sort of runtime I get out of one battery with slightly more draw than I expect to need.


Reply
Wolfie
Posts: 2225
Topic starter
(@wolfie)
Noble Member
Joined: 11 years ago

Well, another decision to be pondered. I am still going with 24 vertical columns.

I took a strip out and laid it on the outside of the ball and trimmed it at the neck and leg hole.

At 4" spacing I can get 9px yielding 216 for the ball. at 3" spacing 12px fit on the strip yielding 288 total for the ball.

At the equator, the strands will be spaced 4" apart but obviously they will get closer to each other at the poles so the horizontal spacing is non-uniform.

Thinking 3" vertical spacing since its partway between the equitorial 4" spacing and the roughly 1" spacing at the neck hole. It would yield a denser matrix and is still within the power ability of the two buck converters (144 on each).


Reply
Page 1 / 16
Share: