what is the most efficient way to generate several voltages from one PS

[mottihoresh]

Hi,

I am new this this forum, have been following the blog for about a year now, I am currently working on a LED driver for my salt water aquarium. Did some research and decided to go with the CAT4101 LED driver.

I am planning to drive about 33 leds, between 2-6 on each array.

My question is regarding the power supply, I have ordered a 24v 10a switching power supply, it would be ideal if I could use just one power supply to drive all arrays. the arrays are:

6 10k - 22.2V
5 6500k - 18.5V
5 Royal Blue - 18V
6 Royal Blue - 21.6V
5 Normal Blue - 18V
2 Pink - 7.2V
2 Violet - 7.4V
2 UV = 7.4V

What would be the best way to generate these voltages from the power supply? would it be better to use resistors as voltage dividers, zener diods or voltage regulators?

Thanks!

[sacherjj]

Are these being driven with just on/off control?

If current requirements are similar in the 7.2, 7.4, 7.4 group, you may be able to drive them in series.  Even if they are not balanced, the proper resistors in parallel could fix that.  The others could just use power rated current limited resistors.

Do you have the current specs for the LED arrays?

[mottihoresh]

the leds are going to be driven using PWM signal, each LED's forward voltage is between 3.6 - 3.7v and draw 700ma. The configuration of the driver is set that the negative (-) end of the leds array will be connected to the driver, and it will sink any remaining voltage.

The datasheet recommends having about .5v as padding voltage which will be dispated in the driver, having more then that could cause the driver to overheat and shut down.

The other 'issue' is that i want to be able to independently turn one array on, and leave the other off, would that cause a problem?
 

[mottihoresh]

The CAT4101 cheap controls the flow of current through the leds.

[sacherjj]

You can do an LM317 at about a buck a pop for the higher voltage lines.  Although, that should give you enough adjustment to use for the LEDs without the other chip.

You would probably want with an +8V power supply or a 24V to 8V DC to DC converter that could handle 3A or more.  Something like this: http://www.ti.com/lit/ds/symlink/tps54331.pdf

[mottihoresh]

Hi, thanks for the info, the main reason for using the other chip is for the constant 700ma current, and the ability to control it with a PWM signal. I am going to order few samples of the chip you have mentioned and try it out!  thanks again! O0

[mottihoresh]

Another quick question regarding that DC-DC convertor, I used https://analogtools.ti.com/SwitcherPro/Main/MainPage.aspx? to help me find the values for the passive components, the Schematics indicates two 'open' capacitors, does it mean 'no capacitor'?

I used the following info when generating the list:
vin min 23
vin max 24
out voltage 8 (might lower that a bit)
out current 3

[mottihoresh]

Since we are going toward the approach of using voltage regulators, why would TPS54331 would be better over linear 8v regulator http://www.ti.com/product/ua7808

[sacherjj]

The advantage of a Buck convertor over a linear 8V regulator is purely efficiency.  If you have 8V linear regulators driving nearly 3 A, you have linear regulators that are dropping 16V x 3 A = 48 W which is a lot of power and a lot of heat sink required.  You can also buy prebuild DC to DC convertor boxes that would just hook up.

[BravoV]

I suspect the op still doesn't understand the current regulation needed for driving led, "not" voltage.

Driving led with voltage regulated is not the best way especially with different so many different colors at multiple series of leds.

[sacherjj]

I suspect the op still doesn't understand the current regulation needed for driving led, "not" voltage.

Driving led with voltage regulated is not the best way especially with different so many different colors at multiple series of leds.

Using a constant current chip that was specified, it should be fine, as long as the voltages for each set of LEDs is brought down within the power dissipation capability of the constant current driver.

[mottihoresh]

Thanks for all the help, i'll post the schematics once i am done.