«

»

LEDs – How To

This is the last portion of the series on LEDs (and is probably the most useful).  Below is the step-by-step build for my LED lighting system for my 150g SPS-dominant reef.  As with any DIY (especially electrical), great care must be taken for personal and material safety.

CHAPTER I – INTRODUCTION

CHAPTER II – LITERATURE REVIEW

CHAPTER III – DATA COLLECTION

CHAPTER IV – METHODOLOGY

CHAPTER V – ANALYSIS AND RESULTS

Chapter VI – Conclusion

Fixture Build Documentation

Before starting the build, all the necessary supplies and equipment were gathered, including the items in Table 20, Table 21, and Table 22.  Other supplies included a Sparkfun Electronics 937b electrostatic discharge (ESD) safe soldering iron, various wire strippers, nippers, pliers, screwdrivers, a soldering wire holder, and a heat gun (Figure 54, Figure 55, and Figure 56).

Figure 54:  LED Build Workstation

Figure 54: LED Build Workstation

Figure 55:  LED Build Supplies

Figure 55:  LED Build Supplies

Figure 56:  Nippers and Wire Strippers

Figure 56:  Nippers and Wire Strippers

The four heat sinks in Figure 57 are MakersLED heat sinks from RapidLED.com.  These heat sinks had the best safety factor out of the ones analyzed.  The faces shown are where the LEDs will be mounted.

Figure 57:  MakersLED Heat Sinks

Figure 57:  MakersLED Heat Sinks

The LEDs were joined together and were separated up front to speed production later (Figure 58, Figure 59, and Figure 60).  All LEDs were kept in their packaging as much as possible to prevent damage.

Figure 58:  Royal Blue LEDs

Figure 58:  Royal Blue LEDs

Figure 59:  430 nm Violet LEDs

Figure 59:  430 nm Violet LEDs

Figure 60:  Red LEDs

Figure 60:  Red LEDs

In order to pre-tin the LEDs, the soldering iron was turned on to 350 degrees Celcius and preheated (Figure 61).  The LED was place in the soldering wire holder for easier access to the soldering pads (Figure 62).

Figure 61:  Sparkfun Electronics 937b ESD-Safe Soldering Iron

Figure 61:  Sparkfun Electronics 937b ESD-Safe Soldering Iron

Figure 62:  LED in the Soldering Wire Holder

Figure 62:  LED in the Soldering Wire Holder

The tip of the soldering iron was placed on each LED pad for two seconds before the solder was touched to the iron.  This allowed the solder to flow properly on to each LED pad with no overruns or splatters (Figure 63).

Figure 63:  Tinned LED Solder Points

Figure 63:  Tinned LED Solder Points

After the LEDs were pre-tinned, they were laid out according to the design in 1.5 inch increments (Figure 64).  They were also tested with a multi-meter to ensure connectivity.

Figure 64:  LED Placement on the Heat Sink

Figure 64:  LED Placement on the Heat Sink

Next, 144 bolt sets were made by sliding the nylon washer onto the bolt and barely threading on the nut (Figure 65).

Figure 65:  Adding Nylon Washers and Nuts to the Bolts

Figure 65:  Adding Nylon Washers and Nuts to the Bolts

Two bolts were required to secure each LED, so four bolts were added to the outside races and seven bolts to each of the four inside races (Figure 66 and Figure 67).

Figure 66:  Adding Bolts to the LED Layout

Figure 66:  Adding Bolts to the LED Layout

Figure 67:  Bolts in Rough Position

Figure 67:  Bolts in Rough Position

Arctic Alumina Thermal Adhesive was also used in conjunction with the bolts to secure the LEDs, as well as to help conduct heat away from the LEDs (Figure 68).

Figure 68:  Arctic Alumina Thermal Adhesive

Figure 68:  Arctic Alumina Thermal Adhesive

This adhesive is a two-part mixture.  Each part (A and B) is added together in equal amounts (Figure 69) and a small drop is added to the back of each LED (Figure 70).  When the LED is placed on the heat sink, the thermal adhesive will spread to coat the back LED surface.

Figure 69:  Mixing Thermal Adhesive, Part A and B

Figure 69:  Mixing Thermal Adhesive, Part A and B

Figure 70:  Thermal Adhesive on LED Back

Figure 70:  Thermal Adhesive on LED Back

With the thermal adhesive on the back of the LED, it is placed into its proper position on the heatsink.  Then, the bolts are moved into their final position and tightened down.  The thermal adhesive should appear in a thin line around the LED if the correct amount was applied (Figure 71).

Figure 71:  LED in Permanent Position

Figure 71:  LED in Permanent Position

After all the LEDs are secured, wiring preparation was required.  The wiring diagram was necessary, along with pre-tinned wire connections (Figure 72).

Figure 72:  Wire Preparation

Figure 72: Wire Preparation

Each wire was cut to the proper length for each connection, the ends were stripped, and the stranded wires were twisted together.  The ends were pre-tinned with the soldering iron.  To make the LED-to-wire connection, the soldering iron tip was touched to the LED solder pad until the solder melted.  The pre-tinned wire was then inserted into the LED pad solder (Figure 73).  Each joint was inspected and tested with a multi-meter for connectivity.  Figure 74 shows the LEDs wired in series.

Figure 73:  Adding Wire to the LEDs

Figure 73:  Adding Wire to the LEDs

Figure 74:  Wiring Completed with Royal Blue Circuit Highlighted

Figure 74: Wiring Completed with Royal Blue Circuit Highlighted

Occasionally a mistake was made or an LED prematurely died (due to overheating with the soldering iron).  These LEDs were removed by unscrewing the bolts, applying light heat with a heat gun (Figure 75), applying light pressure with a screwdriver to pop it up, and then scraping the thermal adhesive off with a razor blade (Figure 76).  If the LED needed repositioning, it was retested with the multi-meter (to ensure it did not burn out during removal with the heat gun) and moved.  If the LED was burnt out, it was replaced with a functional LED.

Figure 75:  Heat Gun

Figure 75:  Heat Gun

Figure 76:  Remnant Thermal Adhesive

Figure 76:  Remnant Thermal Adhesive

Once all the LEDs were wired together (in series), the drivers were tuned.  Each Meanwell driver was opened up (Figure 77), and the potentiometer was located.  A small screwdriver was used to turn the driver current down as far as possible (turned the SVR2 potentiometer counterclockwise) (Figure 78).  The current was  measured with a multi-meter.  This step was required to ensure excess current did not destroy the LEDs during testing.  Additionally, the Meanwell drivers required the addition of an AC power cord (Figure 79).  As with all wire connections, the ends were soldered together and heat shrink tape covered the connection.  The RapidLED Nano drivers did not require the tuning or power cord steps.

Figure 77:  Meanwell Driver Internals

Figure 77:  Meanwell Driver Internals

Figure 78:  Turning down the Potentiometer

Figure 78: Turning down the Potentiometer

Figure 79:  Adding the AC Power Cord to the Meanwell Drivers

Figure 79:  Adding the AC Power Cord to the Meanwell Drivers

In addition to the LED system build, a hanging fixture was designed and built.  Figure 80 shows the base structure as viewed from above.  It was built with drawer rails and rollers for easy installation and removal into the prebuilt aquarium canopy.  The horizontal gaps were cut for the heat sink hangers, and the circular holes were for wires and some heat dissipation.

Figure 80:  LED Light System Fixture

Figure 80:  LED Light System Fixture

In order to tune the LED drivers, a secondary potentiometer was built (Figure 81 and Figure 82).  It allows the voltage to vary from 0-10 volts.  At 10 volts, the driver current is adjusted to 1.3 amps with the driver potentiometer (turning the potentiometer counterclockwise and monitoring the current with a multi-meter).

Figure 81:  Potentiometer Circuit Board

Figure 81:  Potentiometer Circuit Board

Figure 82:  Potentiometer Internals

Figure 82:  Potentiometer Internals

The test drive circuit (Figure 83 and Figure 84) was set up with the Meanwell driver potentiometer turned down, the secondary potentiometer turned down to zero volts, and all power was off.  The negative dimming wire (white) on the driver was connected to the negative (white) wire of the secondary potentiometer.  The negative DC output wire from the driver (black) was connected to the negative (black) lead of the LED string under test.  The positive secondary potentiometer wire (black) was connected to the positive driver dimming wire (blue).  The driver positive DC wire (red) was connected to the positive multi-meter lead.  Lastly, the negative multi-meter lead was connected to the positive (green) lead of the LED string under test.  This setup tests the LED string in series and monitors the current through the system at a maximum of 10 volts.

Figure 83:  LED Test and Adjustment Circuit

Figure 83:  LED Test and Adjustment Circuit

Figure 84:  Test Circuit Detail

Figure 84:  Test Circuit Detail

Unfortunately, even with a good design, not all projects go according to plan.  At least there was encouragement within view (Figure 85).  The LEDs would turn on, but they would only run at a maximum of 0.3 amps instead of 1.3 amps at 10 volts.  Troubleshooting took several hours over several days, which nearly impacted the schedule.  The problem was exacerbated by inconclusive test results from two secondary potentiometers.  Eventually a 9-volt “wall wart” was used to eliminate the potentiometer variable.  The root cause was a lower voltage driver capacity than marketed on the specification sheets.  The initial LED design required each driver to run 16 LEDs, but the drivers could only handle 14 LEDs.

Figure 85:  Troubleshooting the LED System

Figure 85:  Troubleshooting the LED System

Once two LEDs were removed from each of the driver LED strings, the current reached 1.30 amps at 10 volts (Figure 86).  Each driver was then re-tuned to run at 1.3 amps.  Unfortunately, the violet LED driver was also accidentally tuned to run at 1.3 amps instead of the required 700 mA (violet LEDs cannot handle high current), so all of the violet LEDs burnt out and had to be replaced.

Figure 86:  Correct Current Obtained

Figure 86:  Correct Current Obtained

The royal blue LEDs were placed on two drivers.  One driver controls the left side, and the other driver controls the right side.  This allows a minimalistic sunrise-to-sunset movement if desired.  Figure 87 shows the right side royal blue LED string with only 14 LEDs.  Two LEDs were taken off the string in order to run the string at full current.  The left side is identical.

Figure 87:  Right Side Royal Blue LEDs

Figure 87:  Right Side Royal Blue LEDs

A single driver controls the 14 neutral white LEDs.  Again, two LEDs were removed from the string in order to run at full current.  Figure 88 shows the neutral white LED string.  The intensity variation appearance is due to the LED optic directivity and relative heat sink and camera angles.  The intensity output across each LED is the same, regardless of visual perception.

Figure 88:  Neutral White LED String

Figure 88:  Neutral White LED String

Both the 405 nm and 430 nm violet LEDs were run on one driver.  The 430 nm LEDs (eight) line the top and bottom while the 405 nm LEDs (four) line the center (Figure 89).  Again, perceived intensity differences are not representative of actual LED output.  Unfortunately, this LED string burnt out quickly after this photo was taken since the driver was accidentally set to 1.3 amps.  The driver was re-tuned to 700 mA and the LEDs were replaced.

Figure 89:  405 and 430 nm Violet LED String

Figure 89:  405 and 430 nm Violet LED String

The cyan/turquoise, red, and cool blue strings all run on the RapidLED Nano drivers that do not require tuning.  However, each string was tested regardless.  Figure 90 shows the cyan/turquoise LED string, Figure 91 shows the red LED string, and Figure 92 shows the cool blue LED string.

Figure 90:  Cyan/Turquoise LED String

Figure 90:  Cyan/Turquoise LED String

Figure 91:  Red LED String

Figure 91:  Red LED String

Figure 92:  Cool Blue LED String

Figure 92:  Cool Blue LED String

Next, each driver was mounted to the overall fixture, and the heat sink fixtures were hung under the fixture.  Meanwell driver DC power was connected permanently to its corresponding LED string (positive-to-positive, negative-to-negative) in a screw terminal box (Figure 93 and Figure 94).

Figure 93:  Meanwell Drivers-to-LEDs Connections

Figure 93:  Meanwell Drivers-to-LEDs Connections

Figure 94:  Meanwell Drivers-to-LEDs Connections Full View

Figure 94:  Meanwell Drivers-to-LEDs Connections Full View

Inside of a project box (Figure 95), the RapidLED Nano drivers were also connected to their LED strings on a screw terminal box.  All dimming wires were fed into the project box as well and were soldered onto a pinned circuit board.  A computer connector was salvaged and used in conjunction with the pins.

Figure 95:  DC Power and Dimming Control

Figure 95: DC Power and Dimming Control

To run six fans (one on each heat sink and one on each side of the overall fixture) in parallel, the Molex connectors were cut off (Figure 96 and Figure 97).  Additional lengths of wire were added and fed to a potentiometer.  This potentiometer controls the fan speed (and therefore, noise level).

Figure 96:  Fan with Molex Connectors

Figure 96:  Fan with Molex Connectors

Figure 97:  Fan with Molex Connectors Removed

Figure 97:  Fan with Molex Connectors Removed

After all the wiring connections were completed, the entire system was tested one last time before the final cleanup (Figure 98).

Figure 98:  Final Developmental Test

Figure 98:  Final Developmental Test

Arguably the most frustrating part of the build was adding the optics.  Super glue cannot be used to secure the optics since it will emit fumes at high temperatures, which will cloud the optics.  So, room temperature vulcanizing (RTV) silicone is used to secure the optics (Figure 99).  Needless to say, it is not the best adhesive.

Figure 99:  RTV Silicone with Optic

Figure 99:  RTV Silicone with Optic

A small amount of RTV was dabbed onto the backside of the optic and allowed to sit for several seconds to partially gel (Figure 100).  Then, the optic was placed over the LED and held for approximately 30 seconds (Figure 101).  It was not uncommon for the optics to fall off a few minutes later.  Additionally, the Exotic LEDs have a slightly larger LED base, so the optics would not properly fit over them.  Two tabs were clipped off the optics to make them fit (Figure 102).  The entire process took several hours (Figure 103).

Figure 100:  Optic Coated with RTV

Figure 100:  Optic Coated with RTV

Figure 101:  Optics on LEDs

Figure 101:  Optics on LEDs

Figure 102:  Optic with Two Tabs Clipped and Coated in RTV

Figure 102:  Optic with Two Tabs Clipped and Coated in RTV

Figure 103:  LEDs Covered in Optics

Figure 103:  LEDs Covered in Optics

To make the connection between the driver dimming wires (attached to the computer connector) and the Neptune AquaController Apex, three long CAT5e cables were cut in half to make six cables (for six dimming channels, one for each color).  Shorter cables could have been used if one of the male ends was removed, but it was more economical to purchase three long ones in this design.

The Neptune AquaController Apex Base Module (and the Variable Dimming Module, VDM) have two ports with the ability to each control two channels (V1/V2 and V3/V4) (Figure 104).  Subsequently, the CAT5e cable can be used instead of purchasing a proprietary cable.  With the clip side up, the pins are, from left to right, 10-volt DC, ground, not used, not used, 10-volt DC, ground, not used, and not used (Figure 105).

Figure 104:  Neptune AquaController Apex Base Module

Figure 104:  Neptune AquaController Apex Base Module

Figure 105:  CAT5e Dimming Cable

Figure 105:  CAT5e Dimming Cable

As shown in Figure 106, the colored computer connectors (positive and negative pairs) were connected to each pair of the CAT5e cable.  Two blue CAT5e cables were used along with one black cable.  These cables were then plugged into the Neptune AquaController Apex variable dimming ports.  Since the base module only has four dimming channels available, a secondary VDM was installed as well.

Figure 106:  Dimming Wire Connections

Figure 106:  Dimming Wire Connections

The finishing touches on the light system (Figure 107) included cleaning up the wiring with cable covers, cable ties, and securing all units with Velcro.  The fans were permanently mounted onto the heat sinks, the acrylic heat sink covers were installed, and the heat sink end caps were also screwed on to prevent injury.

Figure 107:  LED Light System Complete

Figure 107:  LED Light System Complete

With the light system complete, programming the Neptune AquaController Apex was next (Figure 108).  Keep in mind that running Apex Fusion changes or eliminates some of the following steps.

Figure 108:  Neptune AquaController Apex

Figure 108:  Neptune AquaController Apex

A network cable (also CAT5e) was connected directly between the network router and the Apex.  To determine the internet protocol (IP) address of the network, the Disk Operating System (DOS) Command Prompt was used with “ipconfig/all” (Figure 109).   This command returns adapter IP addresses (Figure 110).

Figure 109:  DOS Command Prompt

Figure 109:  DOS Command Prompt

Figure 110:  Adapter IP Addresses

Figure 110:  Adapter IP Addresses

The IP addresses in Figure 110 are used to setup the Apex Ethernet at http://apex (Figure 111).  In order to keep the Apex connected to the wireless router, a static Dynamic Name Service (DNS) was required (Figure 112).  DynDNS was used for the service that constantly monitors the external IP address and updates the server.

Figure 111:  Apex Ethernet Setup

Figure 111:  Apex Ethernet Setup

Figure 112:  Dynamic Name Service

Figure 112:  Dynamic Name Service

Email and text messaging alerts were setup in the AquaController Email Setup screen (Figure 113).  This required setting up an email account without Secure Socket Layer (SSL).  GMX was used for this service, as it was free.  The “Alt To:  address” line was filled in with the text messaging email address (for Sprint it is phonenumber@messaging.sprintpcs.com).

Figure 113:  AquaController Email Setup

Figure 113:  AquaController Email Setup

Additional wireless internet settings were required for final configuration on the router (Figure 114 and Figure 115).

Figure 114:  Virtual Server Settings

Figure 114:  Virtual Server Settings

Figure 115:  Wi-Fi Security Setup

Figure 115:  Wi-Fi Security Setup

After all the Neptune AquaController Apex and home network setup, the Apex home screen is accessible from internet-ready devices (Figure 35).  In other words, remote devices can be used to control the intensity, duration, and spectrum of the LEDs through programming.

The lighting system was installed into the canopy and was connected to power and the dimming modules.  There were no problems as the system ran correctly (Figure 116).  Thus, operational test and evaluation started immediately.

Figure 116:  LED Light System Installed and Functioning

Figure 116:  LED Light System Installed and Functioning

Disclaimer

The views and opinions expressed or implied in this paper are those of the author and should not be construed as carrying the official sanction of the University of Dayton, the Engineering Department, or of individuals/groups mentioned in this paper.  This project is for informational purposes only, and it should not be used replace proper electrical engineering training before attempting such a project.  Any projects arising from this paper are at the reader’s own risk.  Additionally, this report and analysis shall not be used for commercial and/or profit without the author’s explicit written permission and any permission required from the University of Dayton.

References

Leave a Reply