Another DIY LED lighting

[RockyBoy]

Wow that is great news for those who still wonder to go green or not.

Ya bro.. the LEDs are really good on the corals too. I have programmed to switch on about 60-70% royal blue first, then combination of blues and white for total of 70%, then the final timing is to switch on 100%.

When the royal blues are switched on, the corals really glow, especially the hammers, purple anemone, some sps, but of course the fishes look so dull.

When all the lights on, the tank colour looks like 14K. I am still not accustom to the percived lower lighting levels viewing from the front glass. But from the top, damn bright, like my 2 MH lights.

But overall, I am happy that the chiller doesn't have to work so hard now, and I have also cut down the total lights consumption by 15% and I could see the electric bills will come down, even when the tariffs are up

Also save on top-up water since no more MH to boil the water.

[RockyBoy]

Thanks Bro

Is yuor aluminium sheet use as heat sink? What is the wattage per LED you used? I am also thinking of DIY a LED set too.

Need to use aluminium sheet which acts like a heat sink, preferable more than 3mm thick. If you can use those heat sink which looks like butterfly , is better because the greater surface area to conduct the heat away from the LED, but expensive. Also, need some sort of heat grease to have excellent contact with the LED and the aluminium sheet. Then of course you need fan with suitable CFM to extract the heat out.

[Jameshong]

After running for 20 hours... it's time to take the temperature !

I have pointed the digital thermomter directly to the black portion of the star PCB near to the soldering point of the emitter and the reading is 38~39 degree within acceptable range.

I have make another 2 tests; one is pointed directly to the aluminum plate...reading is 35~36 degree and the other pointed to the aluminum plate with splash glass and the reading is as low as 32~33 degree.

This test will be carry on for 3 days to further confirm the average reading then follows by another test without fan but this one, I need to take leave to monitor closely !

Close shot

[Jameshong]

After running for 20 hours... it's time to take the temperature !

I have pointed the digital thermomter directly to the black portion of the star PCB near to the soldering point of the emitter and the reading is 38~39 degree within acceptable range. Distance is 2.5cm away from the point measured.

I have make another 2 tests; one is pointed directly to the aluminum plate...reading is 35~36 degree and the other pointed to the aluminum plate with splash glass and the reading is as low as 32~33 degree.

This test will be carry on for 3 days to further confirm the average reading then follows by another test without fan but this one, I need to take leave to monitor closely !

Close shot

[comycus]

bro v. informative indeed! Have provided us with lotsa information needed to diy. Just wondering, can any newt get the equipment and solder them tog or some skill involved... Really would love to try. Save $ and get the sense of satisfaction of diy.

btw, cool project!

[Jameshong]

bro v. informative indeed! Have provided us with lotsa information needed to diy. Just wondering, can any newt get the equipment and solder them tog or some skill involved... Really would love to try. Save $ and get the sense of satisfaction of diy.

btw, cool project!

Hi bro,

You need proper soldering tool with ESI and some soldering skill in order to secure the LEDs.

To go green further...I would advice all DIY guru or newbies to use LEAD FREE soldering lead. Althougth It's more expensive but It can withstand 40 degree higher than normal leaded soldering lead and non-toxic too.

[comycus]

Hi bro,

You need proper soldering tool with ESI and some soldering skill in order to secure the LEDs.

To go green further...I would advice all DIY guru or newbies to use LEAD FREE soldering lead. Althougth It's more expensive but It can withstand 40 degree higher than normal leaded soldering lead and non-toxic too.

no experience in anything electrical... Guess have to wait for mainstream product prices to drop before forking out the $.

Or some DIY guru lend me a helping hand

[albinosage]

Thank you bro, for making this thread so detailed and informative. I can see you are very excited and satisfied with your efforts haha. Thanks again for sharing!

[BarraCudaTM]

After running for 20 hours... it's time to take the temperature !

I have pointed the digital thermomter directly to the black portion of the star PCB near to the soldering point of the emitter and the reading is 38~39 degree within acceptable range. Distance is 2.5cm away from the point measured.

I have make another 2 tests; one is pointed directly to the aluminum plate...reading is 35~36 degree and the other pointed to the aluminum plate with splash glass and the reading is as low as 32~33 degree.

This test will be carry on for 3 days to further confirm the average reading then follows by another test without fan but this one, I need to take leave to monitor closely !

Close shot

A contact thermometer would be a better choice in this case.

There are several things to ensure good accuracy for infrared thermometers.

First of all, the spot size

The distance between the thermometer and object determines the measurement spot size. Look at the top of yr thermometer, there should be a simple distance to spot size ratio. Basically, the further the distance, the bigger the spot size. Make sure the spot size(the area of temperature measurement) is smaller than yr object.

2nd, infrared thermometers are designed to work best with black body. Reflective objects will severely skew the temperature measurement. Lens/solder/aluminum anything reflective will show a much lower temperature.

3rd, infrared thermometers cannot measure through glass/plastic/food wraps.

You can try pasting a black tape(spot size must be smaller than the tape) over the bare aluminum and try to take the temperature difference between the bare aluminum and black tape. The black tape will show a much higher temperature

[Jameshong]

A contact thermometer would be a better choice in this case.

There are several things to ensure good accuracy for infrared thermometers.

First of all, the spot size

The distance between the thermometer and object determines the measurement spot size. Look at the top of yr thermometer, there should be a simple distance to spot size ratio. Basically, the further the distance, the bigger the spot size. Make sure the spot size(the area of temperature measurement) is smaller than yr object.

2nd, infrared thermometers are designed to work best with black body. Reflective objects will severely skew the temperature measurement. Lens/solder/aluminum anything reflective will show a much lower temperature.

You can try heating up a shiny stainless steel pot and compare the results from a contact thermometer and infrared thermometer, you will be surprised at the great difference

Hi BarraCuda,

Thanks for your comments .

As I have mentioned that 3 points measurement taken with 3 different reading . One of it is pointing IR(red light) directly to the black paint/portion(not reflective) on the star metal core PCB beside the emitter less than 2.5 cm . Although contact thermometer is more accurate but I do not need to go to the extend till 100% as long the different is around +-5% on actual target.

FYI, this fluke digital thermometer that I'm using is mainly designed for commerical uses. And I have been using it in my daily work for years without fails and the accurancy is +-1%.

IMHO, the temperature for 35 lens with 3mm aluminum plate and 2 fans should not exist 45 degree unless we are talking about 100 pcs lor !

[BarraCudaTM]

LEDs maximum temperature is rated at the LED junction which cannot be measured directly. Its not as simple as just taking the measurement of the star.

First of all, we need to consider the thermal resistance of the LED. For SSC P4, its 8.8C/watt. Assuming that there is a thermal resistance of 1.2C/watt between the LED slug and aluminum star. So total thermal resistance will be 10C/watt

The formula for calculating junction temperature will be

Junction Temp = Thermal Resistance X Wattage + Temperature of star

Assuming, temperature of star = 40C, wattage = 4watt, thermal resistance = 10C/watt

Junction temperature = 10 X 4 + 40 = 80C.

There should be a life span chart telling you the estimated lifespan at a specific junction temperature.

I burned hundreds of $$ to realise all these, I certainly hope noone will follow my footsteps.

[BarraCudaTM]

Hi BarraCuda,

Thanks for your comments .

As I have mentioned that 3 points measurement taken with 3 different reading . One of it is pointing IR(red light) directly to the black paint/portion(not reflective) on the star metal core PCB beside the emitter less than 2.5 cm . Although contact thermometer is more accurate but I do not need to go to the extend till 100% as long the different is around +-5% on actual target.

FYI, this fluke digital thermometer that I'm using is mainly designed for commerical uses. And I have been using it in my daily work for years without fails and the accurancy is +-1%.

IMHO, the temperature for 35 lens with 3mm aluminum plate and 2 fans should not exist 45 degree unless we are talking about 100 pcs lor !

Initially I thought infrared thermometers are accurate until you use with reflective surfaces.

I'm involved in temperature control of special steel for heat treatment purpose. I tried measuring the shiny surface but kept getting very low temperature but the contact thermocouple shows that the temperature is just right. Then I tried measuring the mill scale(black color) on the steel and finally the temperature tallies with the contact thermocouple. The initial temperature of the infrared thermometer is ~140C and the contact thermocouple showed 200C, that is a difference of 60C!

You can do a simple search on the internet regarding the limitation of infrared thermometers. Typically infrared thermometers are calibrated at 0.95 emissivity, with a black body emissivity at 1.0. The more reflective the surface, the higher the error.

FYI, I'm also using the Fluke 62 because its the cheapest fluke I can find ... around $100+

[Jameshong]

LEDs maximum temperature is rated at the LED junction which cannot be measured directly. Its not as simple as just taking the measurement of the star.

First of all, we need to consider the thermal resistance of the LED. For SSC P4, its 8.8C/watt. Assuming that there is a thermal resistance of 1.2C/watt between the LED slug and aluminum star. So total thermal resistance will be 10C/watt

The formula for calculating junction temperature will be

Junction Temp = Thermal Resistance X Wattage + Temperature of star

Assuming, temperature of star = 40C, wattage = 4watt, thermal resistance = 10C/watt

Junction temperature = 10 X 4 + 40 = 80C.

There should be a life span chart telling you the estimated lifespan at a specific junction temperature.

I burned hundreds of $$ to realise all these, I certainly hope noone will follow my footsteps.

Ok noted.

Is my below statement correct ?

SSC P4 spec:

Junction Temperature, TJ : 145C @ <=700mA

RθJ-Bis measured with a SSC metal core pcb.(25 ºC≤TJ≤110 ºC)

RθJ-C is measured with only emitter. .(25 ºC≤TJ≤110 ºC)

And my current reading is 39C

So my junction temperature is 10 * 4Watt * 39 = 79C

as such It's still within good range! Am I right to say that ?

[Terryz_]

Hi bro,

Attached is my LED layout. The top right corner is where my overflow box is, so no need to put any LEDs.

Since my tank is 2.5feet deep, to penetrate deeper, I use the 60deg lens for those areas that need the extra reach, which are at the fronts

Cool, thanks bro.. For the attachment... Great reference material when i diy one for my tank.. Thanks..

[nobody]

There is a limit to how efficient a fan or even a dozen of fans can cool the light. Otherwise PC makers will just screw in a 120mm fan blowing at the CPU instead of using massive heatsink/fan combination. A beefy heatsink is more or less a MUST for high power LED driven at spec or overdriven......unless you incorporate a temperature sensor/thermal fuse to cut off/lower the LED current in the event of overheating. A beefy heatsink also act as a buffer in the event of fan failure. Dosen't really matters in this case since you're using an acrylic casing..........

When the temperature of the heatsink has reached 80deg. The junction temperature had probably exceeded the safe limit, depending on how good is the thermal interface between the junction/metal core PCB and the interface between metal core PCB/aluminium heatsink. Even if it dosen't exceed the max allowed temperature. Why not use a proper heatsink and further reduce the junction temperature? It not only will improve efficiency(LED efficiency increase with decrease in temperature), it will also prolong the lifespan of your LED.

Some people learn from their mistake. Others learn from people's mistake. But there are also those who insist on making mistake before they learn.......

[BarraCudaTM]

Ok noted.

Is my below statement correct ?

SSC P4 spec:

Junction Temperature, TJ : 145C @ <=700mA

RθJ-Bis measured with a SSC metal core pcb.(25 ºC≤TJ≤110 ºC)

RθJ-C is measured with only emitter. .(25 ºC≤TJ≤110 ºC)

And my current reading is 39C

So my junction temperature is 10 * 4Watt * 39 = 79C

as such It's still within good range! Am I right to say that ?

Correct, provided the temperature measurement is accurate.

If the temperature reading is skewed by 20C then the LEDs will start to dim in about 6months or not.

[sherman]

Wow its great to see all the DIY guru coming in to share their experience.

[Jameshong]

Correct, provided the temperature measurement is accurate.

If the temperature reading is skewed by 20C then the LEDs will start to dim in about 6months or not.

Thanks for sharing the knowledge !

[Jameshong]

There is a limit to how efficient a fan or even a dozen of fans can cool the light. Otherwise PC makers will just screw in a 120mm fan blowing at the CPU instead of using massive heatsink/fan combination. A beefy heatsink is more or less a MUST for high power LED driven at spec or overdriven......unless you incorporate a temperature sensor/thermal fuse to cut off/lower the LED current in the event of overheating. A beefy heatsink also act as a buffer in the event of fan failure. Dosen't really matters in this case since you're using an acrylic casing..........

When the temperature of the heatsink has reached 80deg. The junction temperature had probably exceeded the safe limit, depending on how good is the thermal interface between the junction/metal core PCB and the interface between metal core PCB/aluminium heatsink. Even if it dosen't exceed the max allowed temperature. Why not use a proper heatsink and further reduce the junction temperature? It not only will improve efficiency(LED efficiency increase with decrease in temperature), it will also prolong the lifespan of your LED.

Some people learn from their mistake. Others learn from people's mistake. But there are also those who insist on making mistake before they learn.......

Thanks for your suggestion and comment.

At this moment, I have no intention to switch to beefy heatsink yet till I do more testing and confirmation.

Btw, do you have any success diying project to share ?

Thanks .

[Gouldian]

Very informative thread. Thanks!

[BarraCudaTM]

There is a limit to how efficient a fan or even a dozen of fans can cool the light. Otherwise PC makers will just screw in a 120mm fan blowing at the CPU instead of using massive heatsink/fan combination. A beefy heatsink is more or less a MUST for high power LED driven at spec or overdriven......unless you incorporate a temperature sensor/thermal fuse to cut off/lower the LED current in the event of overheating. A beefy heatsink also act as a buffer in the event of fan failure. Dosen't really matters in this case since you're using an acrylic casing..........

When the temperature of the heatsink has reached 80deg. The junction temperature had probably exceeded the safe limit, depending on how good is the thermal interface between the junction/metal core PCB and the interface between metal core PCB/aluminium heatsink. Even if it dosen't exceed the max allowed temperature. Why not use a proper heatsink and further reduce the junction temperature? It not only will improve efficiency(LED efficiency increase with decrease in temperature), it will also prolong the lifespan of your LED.

Some people learn from their mistake. Others learn from people's mistake. But there are also those who insist on making mistake before they learn.......

I would agree that having a proper finned heat sink would be the best choice but I would also like to take cost efficiency into consideration. Massive heat sinks costs alot! but if we can compromise performance and costs .. why not?

Several issues we need to consider for heat sink selections.

1) Thickness

The thicker the heat sink, the better spreading of heat. The main criteria is not to allow localised hot spots on the heat sink. The temperature across the heat sink surface must be able to spread evenly. If a heat sink is too thin, the heat sink area directly behind the LED will be very hot while the rest of the heat sink is only warm. In this case, the heat sink is not effective.

Aluminium will expand significantly when heated. This will cause warping and de lamination between LED and heat sink if not rigid enough. Once there is an air gap between the LED and heat sink, the LED will heat up and deteriorate rapidly.

2) Total cooling surface area

The higher the surface area available, the better radiation of heat. Finned heat sinks provides more cooling surface area than flat aluminium plates. To compensate for the lower surface area, you simply use a larger piece of aluminium plate. For me, I use a aluminium plate size of 4'X2'.

3) Color

Black objects radiate(lose) heat(infrared) much better than shiny objects even though they are of the same temperature. That is the reason why I'm saying most infrared thermometers are not designed to work with shiny surface. Objects emit infrared radiation when heated, a shiny object will emit less infrared radiation than a black object even though they are of the same temperature.

The problem comes when most infrared thermometers are calibrated to relate the temperature against the infrared radiation it received of a black object(emissivity=1/0.95). At the same given temperature, a shiny object will emit far less infrared than a black object. The infrared thermometer will then give a much lower false temperature reading. It is okay to use infrared thermometers if accuracy is not much of a issue or if the thermometer is specifically calibrated for that purpose. In our case, even a 10-20C error can make a difference whether the LEDs are operating at the correct temperature or burning itself out.

So black heat sink is always better? Not always the case.

Black color is either a anodized coating or simply black paint. This coating no matter how thin, its a insulator of heat.

So when to use black or bare heat sinks?

As a general rule, use black if there is no active cooling(no fans). The black heat sink would be able to lose heat faster via radiation despite acting as a thermal insulator itself.

Use bare, if there is active cooling. Without the thermal insulator(black coating), the heat sink would be able to lose heat faster via conduction and convection.

In a nutshell,

Using aluminium plate might not necessarily be a bad choice but you have to build it thicker, bigger or introduce more active cooling

[comycus]

good to see all gurus dishing out advice here for anyone to pick up! correct me if I'm wrong:

Keeping all variables constant,

finned heat sink>aluminum bare heat sink w fan>aluminum black heat sink fanless??

'>' = better than

so to compensate when costs are factored in, bigger and thicker is better. Nothing new here

[Jameshong]

I would agree that having a proper finned heat sink would be the best choice but I would also like to take cost efficiency into consideration. Massive heat sinks costs alot! but if we can compromise performance and costs .. why not?

Several issues we need to consider for heat sink selections.

1) Thickness

The thicker the heat sink, the better spreading of heat. The main criteria is not to allow localised hot spots on the heat sink. The temperature across the heat sink surface must be able to spread evenly. If a heat sink is too thin, the heat sink area directly behind the LED will be very hot while the rest of the heat sink is only warm. In this case, the heat sink is not effective.

Aluminium will expand significantly when heated. This will cause warping and de lamination between LED and heat sink if not rigid enough. Once there is an air gap between the LED and heat sink, the LED will heat up and deteriorate rapidly.

2) Total cooling surface area

The higher the surface area available, the better radiation of heat. Finned heat sinks provides more cooling surface area than flat aluminium plates. To compensate for the lower surface area, you simply use a larger piece of aluminium plate. For me, I use a aluminium plate size of 4'X2'.

3) Color

Black objects radiate(lose) heat(infrared) much better than shiny objects even though they are of the same temperature. That is the reason why I'm saying most infrared thermometers are not designed to work with shiny surface. Objects emit infrared radiation when heated, a shiny object will emit less infrared radiation than a black object even though they are of the same temperature.

The problem comes when most infrared thermometers are calibrated to relate the temperature against the infrared radiation it received of a black object(emissivity=1/0.95). At the same given temperature, a shiny object will emit far less infrared than a black object. The infrared thermometer will then give a much lower false temperature reading. It is okay to use infrared thermometers if accuracy is not much of a issue or if the thermometer is specifically calibrated for that purpose. In our case, even a 10-20C error can make a difference whether the LEDs are operating at the correct temperature or burning itself out.

So black heat sink is always better? Not always the case.

Black color is either a anodized coating or simply black paint. This coating no matter how thin, its a insulator of heat.

So when to use black or bare heat sinks?

As a general rule, use black if there is no active cooling(no fans). The black heat sink would be able to lose heat faster via radiation despite acting as a thermal insulator itself.

Use bare, if there is active cooling. Without the thermal insulator(black coating), the heat sink would be able to lose heat faster via conduction and convection.

In a nutshell,

Using aluminium plate might not necessarily be a bad choice but you have to build it thicker, bigger or introduce more active cooling

Well said bro BarraCuda !

I'm ready trying hard to compromise performance, outlook and costs effectively !

[Jameshong]

Update .

[comycus]

Beautiful! extremely neat and clean look

just wondering if the exposed wires should be a cause for concern?