# DIY Chip on Board LED circuit - help with review of safety.

#### LEDIY

Joined Dec 2, 2016
26
Hi,

I have been putting together a DIY LED circuit for a lighting fixture in my room consisting of 4 high power COB LED's. They will be mounted on heat sinks which have been calculated for the correct thermal dissipation and in an aluminum enclosure.

I am just awaiting the LED bulbs to arrive tomorrow in order to be able to complete the circuit and flip the switch, but I am an absolute laymen at best when it comes to electronics and have simply gone off all of the required ratings for each piece in the circuit and read from various sources how to wire them together.

It is not a complex circuit by any stretch of the imagination, just two dc converters wired in parallel to power the led and fans respectively. Then the fans and LEDs wired in parallel to the dc converters so that they all share the max amperage the converter gives out whilst maintaining a constant voltage. I have used wago clips to split the wires on both the ac and dc side of the circuit.

I have thrown together a rudimentary wiring diagram in paint and was hoping that you fine people could have a look over it and maybe tell me if I am going to electrocute myself with 240 volts when I flip the switch or whether this all seems reasonable enough. I also wanted to ask is there a way to calculate what the AC side of the circuit will be drawing from the mains amps wise, so I can work out whether my rocker switch is sufficiently protected. At the moment it is rated for 6amps @ 240 volts.

I have not used convention in the diagram as I have yet to learn any, so it shows the actual components I will be using.

LEDIY

#### LEDIY

Joined Dec 2, 2016
26
Why is the switch wired as shown ?
Because it has an led in it to indicate on so needs a neutral.

#### AlbertHall

Joined Jun 4, 2014
11,305
Is it a mains rated LED then?

#### Alec_t

Joined Sep 17, 2013
12,060
Is your 36V DC converter a constant voltage type or a constant current type? LEDs need something to control the current through them.

#### LEDIY

Joined Dec 2, 2016
26
Is it a mains rated LED then?
Yeah the switch as a unit is rated for 240vac @ 6amp max.

I wanted to know how I would estimate the current draw for the ac side of the circuit to see if that switch is good enough, because the dc converters dont say on the data sheet or on them what the current draw is in order to convert to 36vdc @ 10amps.

#### LEDIY

Joined Dec 2, 2016
26
Is your 36V DC converter a constant voltage type or a constant current type? LEDs need something to control the current through them.
Hi Alec,

They are constant voltage at 10 amps variable as a max output.

I understand that the LED will draw the maximum current available to it so by wiring them parallel i am assuming that the total amperage would be divided by 4 in this set up and each LED will draw 2.5 amp of dc current each.

I understand there can be issues with current runaway if they all consume slightly different currents, but there is a high margin for error because they are rated at 4amps. And they are all high bin cree leds. I have thought about installing a 3amp inline fuse on the live to each for protection so if one fuse blew the others would be drawing over 4 and all the lights would go out.

Would I need something to regulate current in this scenario?

#### LEDIY

Joined Dec 2, 2016
26
I meant if two of the fuses blew, not one. Not enough posts to edit!

#### Alec_t

Joined Sep 17, 2013
12,060
Would I need something to regulate current in this scenario?
Almost certainly. If each LED Vfwd is rated as 34V (i.e. a typical, not guaranteed, value) and you apply a fixed 36V supply then without current-limiting there is nothing, apart from any cut-out in the power supply, to stop the LED drawing far more than its maximum allowed 4A. It is usual to drive LEDs of this sort from a constant-current source.

#### LEDIY

Joined Dec 2, 2016
26
Almost certainly. If each LED Vfwd is rated as 34V (i.e. a typical, not guaranteed, value) and you apply a fixed 36V supply then without current-limiting there is nothing, apart from any cut-out in the power supply, to stop the LED drawing far more than its maximum allowed 4A. It is usual to drive LEDs of this sort from a constant-current source.
Oh no! Ok so then I need to put a current regulator in there. wire them in parallel to the dc converter and then wire that to the LED.

Would a buck converter be able to do this without loosing more than 2v from the 36v input?

The LED wouldn't fire at less than 34v right?

Do boost converters regulate current as well or just boost the voltage?

This has thrown a spanner in the works. But im glad i didnt flip the switch!

#### Alec_t

Joined Sep 17, 2013
12,060
Would a buck converter be able to do this without loosing more than 2v from the 36v input?
Maybe. You would have to check the specs of specific makes/models.
The LED wouldn't fire at less than 34v right?
34V is not a sharp cut-off. Slightly below that would still give some light. The problem is that the '34V' is only a typical value. Your particular LEDs may have a somewhat different Vfwd.
Do boost converters regulate current as well or just boost the voltage?
Depends on the type. Ideally you want a constant-current regulator which can provide up to 4A per LED at a voltage in excess of , say, 36V (for a single LED) or 70V (for 2 LEDs in series), etc.

#### LEDIY

Joined Dec 2, 2016
26
Maybe. You would have to check the specs of specific makes/models.

34V is not a sharp cut-off. Slightly below that would still give some light. The problem is that the '34V' is only a typical value. Your particular LEDs may have a somewhat different Vfwd.

Depends on the type. Ideally you want a constant-current regulator which can provide up to 4A per LED at a voltage in excess of , say, 36V (for a single LED) or 70V (for 2 LEDs in series), etc.
Thanks a lot for the advice Alec. I have seen an instructable which seems to do what I need with a current loss of only 0.7v and a variable constant current flow. so perfect for me.

It uses an LM350, the thing is the poster advises you would only need a 1, 2 or 3 watt potentiometer but in the comments a guy states that this would need to be able to handle all the current. So for 3amps @ 10volts you would need at least a 30w pot. which is super expensive.

I have attached the wiring diagram he posted. can you tell me if a low watt pot is safe to use on this circuit?

The link to the instructable is:

#### AlbertHall

Joined Jun 4, 2014
11,305
The pot would be taking the full output current but there would be only 1.25V across it (whatever the setting). For 3A the pot would be set to just under 0.42Ω. You would be better off using a fixed resistor - 0.43Ω, 5W would do the job.

#### LEDIY

Joined Dec 2, 2016
26
The pot would be taking the full output current but there would be only 1.25V across it (whatever the setting). For 3A the pot would be set to just under 0.42Ω. You would be better off using a fixed resistor - 0.43Ω, 5W would do the job.
Hi Albert,

Did you mean a resister instead of the potentiometer because i dont need to change the current?

What would the output current be for that resistor? I dont actually want 3 amps i want 2.5amps so I could run 4 rather than 3 LEDs!

Also is it ok to wire these in parallel from the dc converter and then individually to each LED?

#### LEDIY

Joined Dec 2, 2016
26
Hi Albert,

Did you mean a resister instead of the potentiometer because i dont need to change the current?

What would the output current be for that resistor? I dont actually want 3 amps i want 2.5amps so I could run 4 rather than 3 LEDs!

Also is it ok to wire these in parallel from the dc converter and then individually to each LED?[/QUOTE

#### LEDIY

Joined Dec 2, 2016
26
Sorry for the double post my phone messed up.

And thanks for helping guys I know these must sound like stupid questions compared to the other threads on this forum!

#### AlbertHall

Joined Jun 4, 2014
11,305
I dont actually want 3 amps i want 2.5amps so I could run 4 rather than 3 LEDs!

Also is it ok to wire these in parallel from the dc converter and then individually to each LED?
In that case make the resistor 0.5Ω, 5W, and that wiring is good.

#### Alec_t

Joined Sep 17, 2013
12,060
If each LED is drawing 2.5A, the four LEDs will nominally be dissipating 34V x 4 x 2.5A = 340W in total. Some serious heatsinking will be required if you don't want the LEDs to fry. What are your plans for that?

#### AlbertHall

Joined Jun 4, 2014
11,305
If each LED is drawing 2.5A, the four LEDs will nominally be dissipating 34V x 4 x 2.5A = 340W in total. Some serious heatsinking will be required if you don't want the LEDs to fry. What are your plans for that?
I agree that heatsinking will be required but your calculation assumes all the power is converted to heat. What is the efficiency of these LEDs - perhaps 30%. If so then the heat to get rid of is more like 200W.

#### Alec_t

Joined Sep 17, 2013
12,060
Agreed. 340W is the input power; 200W is waste heat (unless you use it for brewing coffee ).

#### LEDIY

Joined Dec 2, 2016
26
Hi guys,

The LEDs will be mounted individually on active CPU heatsinks with the 12v fans attached to the top.

They are rated for 80w of thermal dissipation each and the waste heat for the LEDs is around 50w per light.

Another thing that I have neglected to consider is im assuming that the fans will need a constant current source like the LEDs or they could all be running at different speeds?

Also, the LM350 specs state that the max input voltage should be 34v. But I am putting in 36v.

I have found the LM334Z on another site which, i think, does the same job.

Datasheet:

Can I just substitute this in to the instructions given in the instructable with the 0.5Ω, 5W resister in place of the potentiometer and have something thst will work for regulating my current?