# High Current DC LEDS Strips with Aluminium Plate… is grounding advisable?

I'm in the process of planing to make a rather large LED light panel that uses 24v 130w in total. The LED PSU Driver is rated up to 150w and comes with 3 inputs (earth, neutral and live) and 2 outputs positive and negative.

The LED strips that I'm using have fairly decent PCB copper thickness but to help dissipate the heat I've decided to use a large 5mm thick aluminium plate which hopefully will prolong the life span of the LED brightness.

A simple diagram can be found below:

The question(s):

• Is it advisable to earth the aluminium plate using a wire attached to the plate back to the chassis of the PSU?
• Or is there an alternative option or is it not required?

Extra Information:

Meanwell 150W 24v CLG-15024A datasheet

CLG-15024A Schematic

Picture of the LED STRIPS

Looking at the LED strip it seems to have SMD Resistor with label 121 (120ohms, I think) on the +24v and -24v which then reduces the voltage by half to 12v which is then split between 6 leds in series resulting in 2 volts each.

• What are dimensions of this ESD , antenna (j/k) thermal radiator? – Tony Stewart Sunnyskyguy EE75 May 4 '18 at 16:42
• The plate? 600x400x5mm – Simon Hayter May 4 '18 at 17:13
• Is your LED driver double-insulated? Or is the aluminium plate inaccessible behind an insulator? If neither of those are true, it would be compulsory (in the UK anyway) to earth the plate. – Brian Drummond May 4 '18 at 17:18
• @BrianDrummond seems I got the wattage wrong! but I've updated the question with datasheet of the LED driver and a Schematic. Looking at the Schematic it looks like the input has a ground and the output on the -V rail?, does that mean I can simply link the plate to the end of the negative -V? – Simon Hayter May 4 '18 at 17:33
• It is triple insulated. Transformer, PCB and adhesive tape – Tony Stewart Sunnyskyguy EE75 May 4 '18 at 18:00

Thermal resistance from LED junction to surface is degraded by the adhesive tape but for alum plate heat radiation from SMT soldered to MCPCB I use 5sqcm/W for40’C rise in heat sink which means junction is hotter.

You have 2400 sqcm/ 170W or about 3 times this but a thermal resistance from 2 sided tape that may increase Rca thermal resistance much more than chip to copper so the the alum sheet won’t be as effective but will reduce the junction temp by the rise in sheet alum. Since MTBF is doubled by approx every 10’ drop in junction rise (assuming no process faults) a 30 ‘ rise in metal temp translates to about an 8x improvement in lifetime, so good plan.

Now to the electrical ground or not question.

Will the SMPS radiate Common Mode RF noise? ... probably .. will that interference with high impedance sensors? Maybe if you have any.

Will touching the plate conduct ESD interfere with anything? Maybe but probably not

Will it be a safety issue to ground it or not? Not if the supply is certified but if it has a line filter, leakage current caps may be felt (200 uA approx) if you also touch this with one finger and earth ground. So earthing it is good but does not limit ESD current which might only pose an issue if you had some long wires to some remote drivers controlling it.

So there are tradeoffs and if there are more unknowns in this system then the answer may be more complex (RC shunt to earth).

But if it a simple horticulture lamp in high humidity it may be a simple answer . It doesn’t matter but compatibility issues may arise if there are other unknowns like a ground fault on other equipment or AM radio interference so earthing it is my 1st choice. But if you also have ESD sensitive electronics nearby then a 1M resistor in series is commonly used. Then if RF interference is a problem on AM radios or sensors then a 10nF cap to earth ground with a 100 ohms in series to limit ESD current.

These are my EMC thought processes to any large ground plane with potential compatibility issues or NOT.

• Hi Tony, thanks for your answer. Yea the adhesive and the LED strip PCB is a bad conductor in general and hoping that it helps, even if a little. Lucky the leds I'm using are 3014 which are renowned for keeping temps low but at 240 a meter, that soon becomes issue and hoping with some extra metal and maybe a silent fan somewhere it'll help. MCPCB is an excellent option but can be very costly unless you go COBS but I can't find colour temps that meet my requirements. – Simon Hayter May 4 '18 at 17:53
• The trend now is to use 48V stripLEDs in high power applications to minimize copper losses and can get any CCT in waterproof to avoid moisture ingress degradation. – Tony Stewart Sunnyskyguy EE75 May 4 '18 at 18:03
• Yeh, I just wish you could get LED strips without the adhesive and a copper back, that way you could use 40/60 epoxy/thermal paste. – Simon Hayter May 4 '18 at 19:00
• That’s a great idea but difficult to implement ... the tape can be easily removed but adhesion to the epoxy from its coiled state makes it difficult to hold in place without a large poly sheet To release from the epoxy and careful placement with a cushion to hold it down until cured – Tony Stewart Sunnyskyguy EE75 May 4 '18 at 20:37
• Just had another look at the LED strips and more at the adhesive, it looks like I can pull the adhesive layer off with some ease (fake 3M? maybe?) under the adhesive layer there is 1mm of plastic laminate but think I'll have to keep that otherwise the negative and positive copper rails are exposed and will short out on the metal plate. – Simon Hayter May 4 '18 at 21:03

## UPDATE

No, you do not need to ground the aluminum plate.
The Mean Well is supplying filtered DC power. There is nothing to radiate.

Basically the only possible source of ESD would be the Mean Well CLG and it's enclosed in an aluminum box. Nothing will be radiating from the CLG. Mean Well suppresses the ESD from going back down the power line.

You do not need ESD protection. Even if there was an RF signal on the strips they are mounted to a plate of aluminum that would suppress any ESD.

You are not using lethal voltages so no safety risk.

END OF UPDATE

An aluminum plate will not dissipate much heat. It will absorb heat until it reaches nearly the same PCB temperature as without the plate. Aluminum has good thermal conductivity but not so good convective characteristics without fins. Natural convection relies on buoyancy and gravity (Grashof number) which is minimal on a horizontal flat surface. Unanodized aluminum has poor emissivity and therefore poor heat transfer by radiation.

Rather than use a flat sheet checkout heatsinkusa.com. They have 1" wide heatsinks for 16¢ an inch.

What I do is adjust the current so the PCB temperature is below 55°C (hot to the touch but not burn). There are some strips with an efficacy of about 180 lm/W that do not need a heatsink. I would guess your strips are about half as efficient, maybe 90m/W.

Also if you were to use strips with more efficient LEDs more of the electrical watts would be dissipated as light rather than heat.

The strips I like are Bridgelux EB-Series Gen 2. when in stock the price starts at $4 per 280mm (11"), 1200 lumen and are also sold in lengths of 560 mm and 1120 mm. The 280 mm strips are 19.5V and the others are twice that. Samsung has their F-Series Gen 3 strips at the same lengths and similar cost per lumen and efficacy. The Samsung, with more lumens per strip, can be driven to the point where a heatsink is required but I just decrease the current to the desired temperature. This is a fixture with seven 22" (560mm) using Bridgelux BXEB-L0560Z-30E2000-C-B3 @$7.41 each.
I have two of them with the strips wired in parallel (driver: Mean Well HLG-185H-48B) and series (HLG-240H-C1400B).
The parallel strips did not balance well. Current in each strip ranged from 800mA to 1500mA. The good news was even at the 1500mA (60W) strip did not need a heatsink. Max rated current is 1400 mA.

This is a project to experiment with growing seedlings under 3000K (BXEB-L0280Z-30E1000-C-B3) and 5700K (BXEB-L0280Z-57E1000-C-B3) $4 strips. The LED drive is mounted to the top of a wooden shelf and the strips to the bottom. Tomato plants prefer warm 3000K a lot more than cool 5700K. The PCB I made to set the HLG-40H-54B's current and do power distribution. Two$14.48 each, 4500 lumen, Samsung 560mm strips P/N SI-B8T261560WW.
No heatsink required under 1 Amp (1.12 Amp max), 25 Watts each.

## NOTE

The 24V is not divided into two 12V rails.
The voltage of an LED is determined by current and the LED's forward voltage characteristics, not the applied voltage.
White LEDs, driven at ≈60 mA, have a forward voltage of about 2.8V.
6 LEDs in series will be about 6 x 2.8v = 16.8V.
A 120Ω resistor would mean there is 60 mA flowing through each LED and resistor.
This works out to 1.44 W per sets of 6 LEDs.
If the above is true the LEDs would be spaced at about 6 mm (160/m) which would add up to 24 W per meter.
The maximum electrical efficiency would then be 70%. The resistors dissipating the other 30%.
Why they do this is a mystery. They could power 8 LEDs per 27Ω resistor and get 93% efficiency.

• Sadly can't order items from the US since the shipping will exceed the cost of goods, not good. I thought about using a fan and drilling 2mm holes across the back of the plant which should allow the heat to dis-pate better. Good recommendations through, just need something closer to home (UK). – Simon Hayter May 8 '18 at 1:06
• How hot do these strips get? If you can touch the strip and hold your finger on it for a few seconds, you do not need a heatsink. Touch it after an hour or so. It will take a while for the aluminum plate to heat up. I would highly recommend the Bridgelux EB-Series Gen 2 strip. They would likely pay for themselves in electricity savings. What is your wall current or wattage? How many lumens? What is the CCT and CRI? – Misunderstood May 8 '18 at 2:49
• Posted an update. – Misunderstood May 8 '18 at 3:07
• The aluminum plate has been ordered so can't test the temp until it arrives. In regards lumens, the strip in total will be 6 meters at 240 leds per meter, 144 watts in total, 1440 leds in total. Each LED is rated 12-14 lumens, so if take the more conservative number of 12 lumens, the total output should be around 17,280 lumens. – Simon Hayter May 8 '18 at 12:22
• I purchased these: aliexpress.com/item/… because of the higher usage of thicker copper/pcb looks like Sanan Ingan SMD but more likely Nichia because I can't find a datasheet or mention of Sanan 3014. – Simon Hayter May 8 '18 at 12:45