LED Water Wall

Creator: tylercrumpton Status: Designing version v1.0 Born On: 00:52, 11 June 2013 (CDT) Last Updated: 15:36, 09 August 2013 (CDT)

Version v0.1 prototype of a smaller version of one of the panels.

Overview

A board member of Huntsville's Downtown 47 and a member of their investment committee approached us on May 28th, 2013 with an interesting proposition. They are "interested in a project to build a wall covered with LED lights that illuminate when the diodes come in contact with water. Our intent is to place the wall in Big Spring Park as a 'water graffiti' exhibit". Similar projects were undertaken by Antonin Fourneau in Poitiers, France (video) and by Jae Min Lee in South Korea (link).

Requirements

• The group is willing to spend a decent amount of money to get this built, they don't want to cheap out on the project. They are willing to provide the money for a quality production.
• Our contact thought that 6' by 18' is a good size for the wall, but it is open for negotiation.
• Timeline is flexible; they would prefer that it would be done before it starts getting cold again, but if it takes until next summer, that may be acceptable.
• The exhibit will be transportable, and would not need to be left outside 24/7. It will be pulled out and set up for events as necessary.
• We would likely be able to keep the wall at the shop (only if we want!) to show off when the city isn't using it. (I know there are mixed feelings about this; they will not try to force us to hold it if we don't want to)
• They are very excited about Makers Local and the hackerspace scene, and would love to include us in things like this in the future.
• They are willing to pay for a small demo board to show off how the wall would work, and give us a chance to determine a more accurate budget/timeline before ordering thousands of LEDs.
• Depending on the cost, they are interested in looking into extra features/improvements of the LED wall.
• It needs to be safe (obviously); I don't think this will be an issue, if we take the proper precautions.

Design Decisions

Type of LED

Through-hole, 5mm - LINK

Pros:

• Cheaper (around $0.15)
• Larger "dots" (5mm diameter)
• Can be made flush with surface without epoxy
• If no epoxy is used, LEDs may be easily replaceable

Cons:

• Requires grinding
• Soldering process requires bending two leads and soldering each lead
• Must be soldered by hand
• Some brightness lost due to grinding

SMD (Surface-mounted) - LINK

Pros:

• Assembly is simple and quick: apply solder paste with stencil, place LEDs, heat in an oven
• Brighter (? Depends on the effect of the epoxy and other factors)

Cons:

• Smaller "dots" (<2mm diameter)
• More expensive ($0.41)
• Requires some sort of epoxy to be flush with surface
• LED would only extend ~1/3 of the way through the PCB hole
• With epoxy, LEDs may be harder to replace
• Requires a stencil to assemble

Decision

TBD

Power Supply

Some measurements were made on the v0.1 prototype boards (with the eBay LEDs), and it was found that each 25-LED board would draw up to 25mA at 7V for a total of 175mW.

Cost Estimation

PCB Manufacture

PCB Option 1

40 boards, 500mm by 500mm, 1.6mm-thick FR4, 1oz ENIG plating, black soldermask, white silkscreen

• Quote: $2370 + shipping from Shenzhen (~$60 per board)
• I also got a quote for a limited run of just 5 boards: $684 + shipping (~$137 per board)

Time Estimation

In order to estimate the full time-scope of the project, we'll be measuring the amount of time required to assemble prototypes and early boards. This should give us a reasonable estimate for the total number of person-hours required to complete the project, so we can provide a more accurate completion date.

PCB Assembly

Assembler	Task	Time (mm:ss)	Speed (LEDs/hr)	Notes
RAMGarden	Grind LEDs	10:52	138	25 LEDs on v0.1 board. Diffused flat-top LEDs with sand paper and placed LEDs in board simultaneously. Careful stance.
	Place LEDs
	Fold Legs	6:07	245
	Solder	21:00	71
	Cut Excess	5:42	263
	Total	43:41	34
tylercrumpton	Place LEDs	2:00	750	25 LEDs on v0.1 board. Flat-top LEDs. Mildly aggressive stance.
	Fold Legs	3:20	450
	Solder	8:00	188
	Cut Excess	5:00	300
	Total	18:20	82
tylercrumpton	Grind Flat	0:40	2,250	Ground down the LEDs while still on the tape on belt sander. Was able to grind all 25 at the same time.
	Remove LEDs from Tape	1:50	818	25 LEDs on v0.1 board. Ground-down Cree LEDs. Mildly aggressive stance.
	Place LEDs and Fold Legs	5:00	300
	Solder	9:51	152
	Cut Excess	1:43	874
	Total	19:03	79
RAMGarden	Grind Flat	3:22	446	Ground down the LEDs while still on the tape with grinding wheel. Was able to grind two at a time same time.
	Sand Tops	0:22	4,090
	Remove LEDs from Tape	1:29	1,011	25 LEDs on v0.1 board. Ground-down Cree LEDs. Mildly aggressive stance.
	Place LEDs and Fold Legs	10:00	150
	Solder	16:00	94
	Cut Excess	2:00	750
	Total	33:13	45
tylercrumpton	Place LEDs	2:19	647	25 LEDs on v0.1 board. Flat-top LEDs. Mildly aggressive stance.
	Fold Legs	2:21	638
	Solder	6:05	247
	Cut Excess	4:35	327	Used the not-so-fun magnetic clippers. Non-magnetic clippers will speed this up significantly.
	Total	15:20	98

Prototypes

Version v0.0 - complete

Initial version v0.0 homemade prototype to test the overall theory of operation and various sensor shapes.

This 5"x3" 6-LED prototype was made completely in-house. The single-sided PCB was created using an etch-resist mask cut by the vinyl cutter, and the holes were drilled with the drill press. Six blue LEDs and two 3V coin-cell batteries were used to populate the board.

Verdict: Pad shape did not seem to make a difference on water conductivity. Basic circuit idea works very well.

Cost: $0 (used materials that were on-hand)

Version v0.1 - complete

This prototype will consist of 10 6"x6" double-sided PCBs with black soldermask and solder-tinned pads. This prototype will give a better feel for the finished product, and can be used as a sample/demonstration to show off to others. Features that we'd like to see in this prototype are inter-connectivity*, water resistance, and overall aesthetics. This prototype will also help provide a better estimate for the time required to assemble the boards by hand. I'm going to try to work in some sort of music controller as well, using parts I have on-hand.

* Easy inter-connections were not included in this version.

Estimated cost: ~$200 (~$230 with expedited PCB shipping)

• PCBs: Ten double-sided 6"x6" boards, with black soldermask, shipped = $100
• LEDs: 300 C512A-WNS Cree 5mm LEDs from Mouser, white, shipped = $65
• Alternative LEDs: 300 5mm Ultra-bright wide-angle flat-top LEDs, eBay, US seller, shipped = $20
• Connectors: 20 pairs Molex Right-angle PCB connectors, Mouser, shipped with the LEDs = $8
• Other: Optional expedited shipping for PCBs = $30-40

Actual cost: $227.84

• PCBs: $139.65
• LEDs: $55.75 + $4.99 shipping
• Alternative LEDs: $19.65
• Connectors: $7.80

Volunteering

This is a rather large undertaking, so we'll need as many people participating as we can get! If you're interested in helping out with any of the tasks below, feel free to add yourself. If you want to help, but don't know what to do, add yourself to the "Helper" section. More than one person can be in each section, and one person can be in more than one section.

PCB Design

The wall will consist of many interconnected LED PCB boards that contain built-in water sensors. This team will handle designing and manufacturing these PCB boards.

• tylercrumpton

Power Supply System

The wall will require a moderately hefty power supply system to power all of the LEDs and a way to get this power to the boards. We also need to ensure that the system is safe, since we're dealing with water and electricity. This team will need to choose a power supply (or multiple supplies), and determine the best way to route power safely from the supply to the LED boards.

Wall Structure Design/Assembly

Building the physical wall using the PCBs is non-trivial and will need some careful design and thought. This team will help decide on materials and structural components, and will help build the wall's physical supports.

Auxiliary Output

Some "bonus" ideas have been floating around to give the wall more unique and amazing features. The idea of adding some sort of sound output from the wall is the proposed feature with the most attention at the moment. This team will design hardware and software that will receive input from the wall and produce some sort of auxiliary output on a laptop, mobile device, speakers, etc.

Here's the git repo for any auxiliary output related software: https://github.com/ramgarden/LEDWaterWall

• RAMGarden

PCB Soldering/Assembly

Once the PCBs are made and shipped, they will need to be assembled. The wall may exhibit over 15,000 LEDs so as many hands as possible will be needed to solder them all onto the boards. This team will assemble the final PCBs by soldering LEDs to each board and test the completed boards.

• tylercrumpton
• omegix
• JimShoe
• Candleknight
• thematthewknot
• ki4btj
• Brian Gray (Moose)

Photography and Documentation

One of the main goals of the project is to show off what Makers Local 256 is capable of doing. We'll need photos, videos, write-ups, and general documentation of the process to record our progress and successes. This team will be in charge of taking photos and videos of soldering events, wall construction, prototype assembly, circuit testing, and pretty much any other aspect of the project.

• RAMGarden

Helpers

If you don't know what to do, add yourself here, and we can help you find something!

• Kirstin - Snacks and food

Progress Log

00:56, 11 June 2013 (CDT): Started the wiki page with some PCB cost information. We have also created a small sample prototype to prove the viability of the overall idea. The prototype worked very well, and it seems that the water-sensor shape does not really have much effect on the LED brightness or conductivity of the water.

13:58, 19 June 2013 (CDT): We demonstrated the board and got some great feedback and ideas. We were given the go-ahead to draw up a design for more "finished" prototype in order to give a better feel for the final product and the time it will require to build it. One possibly feasible idea that merits investigation is using the same water pads to control a MIDI (or similar) instrument when water hits the pads.

12:44, 15 July 2013 (CDT): Prototype v0.1 PCB design has been completed and sent off to the fab house, and we're awaiting its completion. All of the parts for this prototype have arrived, except the PCB, so assembly of the prototype will begin when the boards arrive.

10:48, 24 July 2013 (CDT): Prototype v0.1 PCBs arrived and we began assembling and testing a few of the boards. We assembled two boards, one with the eBay LEDs as purchased, and one with the eBay LEDs sanded down for diffusion. Both boards were tested on a power supply, and all of the connections and sensors worked very well. We also timed various assembly steps in order to help estimate the overall assembly time of the wall.

21:20, 28 July 2013 (CDT): Two more boards were assembled, with both types of Cree LEDs. There is no noticeable light output difference between the Cree LEDs and the LEDs from eBay. Some wires were also soldered to the boards to allow for easy demonstrations.

11:13, 7 August 2013 (CDT): Some measurements were made on the power consumption and current draw of the v0.1 prototype boards. We're looking at around 200W for a wall of 27,000 LEDs.