Powering long strings of high powered leds

Okay I have another question arising from my led project which came up in another one of my threads on a different topic so I thought I would post it here in a new thread. Hope that's okay.

If I had 90 3.2V/350mA leds and 48v (300-3000mA) dc drivers, what would be the safest / most space saving way to power them?

The main difficulty here is that the leds have pretty low current draw but high voltage requirements and I understand that for safety I should try to match the voltages as closely as possible and keep each circuit below 50v.

So the max number of leds I can run on each circuit would be 14 (giving me a demand of 45V, but only a draw of around 16W) so I would need to have several circuits. However PSUs that supply that much voltage tend to also supply a lot more watts than I need (usually around 150W) unless I go for those unregulated wall plugs or desktop power supplies.

So my question is, should I just run 7 circuits each with their own 48V 150W ps, or is there a better configuration like using some kind of booster or just a larger psu if I can find them? Ideally the configuration would allow me to use the 48v drivers too since I already bought them and they have some features I want (like analog and digital dimming).

thanks very much!

Okay I have another question arising from my led project which came up in another one of my threads on a different topic so I thought I would post it here in a new thread. Hope that's okay.

That is fine but you should have included a link to your other thread.

Ideally the configuration would allow me to use the 48v drivers too since I already bought them and they have some features I want (like analog and digital dimming).

What drivers are these? The ones you had in the other thread did not allow dimming, have you got a new one, if so please post a link to the data sheet.

If I had 90 3.2V/350mA leds and 48v (300-3000mA) dc drivers, what would be the safest / most space saving way to power them?

With one constant current driver you can only power one series string of LEDs whos total sum of forward voltage drops is less than the maximum output voltage of 48V. So assume the voltage drop is 3V per LED then you can have 48 / 3 = 16 LEDs. I would tend to always use one less than that gives you, so I would say 15 LEDs per driver. So for 90 LEDs you would need 6 drivers each with 15 LEDs in series.

However PSUs that supply that much voltage tend to also supply a lot more watts than I need

That presents no technical difficulty as the PSU will only supply the current the circuit needs. It might be seen as uneconomic but that is another story. Each LED chain will take 350mA so 7 chains will take just under 2.5A, a decent amount of current for a 48V supply. Don't think of the wattage of a supply think of it in terms of voltage and current.

So my question is, should I just run 7 circuits each with their own 48V 150W ps,

No.
You run 7 circuits ( driver and LEDs ) all powered from the same power supply.

Okay link to the other thread here: PWMming high powered leds smoothly - LEDs and Multiplexing - Arduino Forum

and thanks for clarifying my thinking. I got confused by wattage. In terms of voltage, the demand would be 153.2 =48v and 3506=2.1A.

So a single 52v 3A (or more) psu would suffice!

for completeness here is the new driver I am using: http://store.sure-electronics.com/led/led-driver/sure-electronics/ps-sp12153. it is based on the LM3409

So a single 52v 3A (or more) psu would suffice!

Well no, from the link to your driver:-

Input voltage: 10V-45V

So feeding 52V into it would break it. Anyway look around I doubt if you can buy a 52V power supply, they tend to come in standard values.

yes, at 45v I will need to change my total number of leds to 13 per circuit.

I can either have 6 circuits totalling 78 leds drawing 2.1A or 7 circuits totalling 91 leds drawing 2.66A.

Now, I would like to add another circuit for my arduino too so will need a step down transformer/converter but I am not sure whether it's possible without a lot of cooling..

It is not possible without further information about this "other" circuit.

More specifically I am thinking of adding an uno to provide the pwm signals and have it draw power off that same circuit instead of using its own separate 12v source, and what is the safest way to achieve this.

The only way I can think of is to connect an arduino to the same psu in a parallel circuit to the leds with a 48vdc-5vdc regulator but I am not sure such a solutin will need active cooling.

connect an arduino to the same psu in a parallel circuit to the leds with a 48vdc-5vdc regulator

You could but connecting a 48 to 7V regulator would allow you to use the Arduino's internal regulator.

I am not sure such a solutin will need active cooling.

No, a heat sing possibly but that is all unless your Arduino is drawing stupid amounts of current due to what it is driving.

A switch more converter would not get as hot as a linear regulator.

a thought just occurred to me. Since the lm3409 based board I am using can supply up to 3A but is limited to 45v, and my leds are 3.2v and only drawing 350ma, the seller of the lm3409 board is suggesting I don't 'waste' the overhead by using it power several led circuits at once. In theory I suppose the board could power up to 8 parallel circuits, but I do know that load imbalances can destroy the circuit. Perhaps this isn't a problem if all the circuits aren't going to powered on all the time and will only do so occasionally but my lights are going to be on all the time.

so my questions are, would anyone recommend it, and is there anyway to make this circuits a bit safer?

I read one way of doing it here: LED DESIGN FORUM: Avoiding thermal runaway when driving multiple LED strings (MAGAZINE) | LEDs Magazine but... I wonder..

skreech:
so my questions are, would anyone recommend it, and is there anyway to make this circuits a bit safer?

No. Well, Mike certainly will not.

skreech:
I read one way of doing it here but... I wonder..

You do well to wonder. The article is essentially bogus - the last supposedly "safe" circuit fails to fail "safe" as proposed.