Using Raw Rectified Voltage

Hi there,

Voltage from the wall (115V/230VAC) goes through a transformer to step it down to something like 36VAC, then that goes through a rectifier to get a max rectified DC voltage of about 36 x sqrt(2) = about 51V. If I want about 50VDC to drive a 12V LED (using a constant-current LED driver), would I still need to use some sort of buck converter or linear regulator to stabilize the DC voltage, assuming I'm using very large smoothing caps after the bridge rectifier?

Thanks!

Probably not necessary to regulate input, but one should check the data sheet for the particular current driver under consideration to see if there are any input ripple concerns.

A constant current driver is typically a buck converter circuit topology with feedback to control the current output rather than the more familiar constant voltage output.

Thanks! That's what I thought as well, the voltage shouldn't matter that much as long as it's higher than the LED voltage. It is floating buck typology.

Thanks!

Now on a separate note, to power a microcontroller running on 3.3V would you use a buck regulator from 50V down to 3.3V or would you first do 50V to 5-6V then use a linear reg down to 3.3V?

Also, how much ripple voltage is considered "bad" for 50V? I'm just trying to calculate the required smoothing cap value.

Are you stuck with that 36volt transformer.
Using 50+ volt for a 12volt LED seems silly.
CC LED drivers for that voltage are rare/expensive/hard to get.
Same for the 50>3.3volt buck converter.
Also note that a 36volt transformer might be ~45volt unloaded.
A better/cheaper choice for a 12volt LED with CC LED drivers is an old laptop supply (19volt regulated).
Leo..

Be careful using very large smoothing caps at such a high voltage. Check rectifier data so You don't blow the rectifier. A start up circuitry might be needed. A radio amateur tought me this.
Wouldn't a voltage like 12+ voltage drop through the current limit device be enough? 2 transistors and a few resistors could do the jobb..
//Railroader

I suppose I could lower the voltage, maybe to 30VAC instead, which would be rectified to about 42VDC. The CC LED driver isn't hard to get, many of them accept up to 65V. I just wanted to drive at a higher voltage so the current would be lower (to stay within overall 18W LED power level) using the CC driver so the LED doesn't get so hot.

Railroader:
Check rectifier data so You don't blow the rectifier. A start up circuitry might be needed. A radio amateur tought me this.
Wouldn't a voltage like 12+ voltage drop through the current limit device be enough? 2 transistors and a few resistors could do the job.

I have looked at the bridge rectifier spec sheet and it can handle the current I need. As far as using large caps, I'm almost leaning towards just using a switching reg so I can avoid the huge caps (like 5-6000uF total) but I'm still weighing my options.

Actually maybe I should choose a much lower voltage, because for constant-current drivers the voltage across the LED will still be the same 12V forward voltage, but the rest of the power will be dissipated through the driver chip. By choosing a lower voltage it wouldn't get as hot, even though the LED is still performing exactly the same. Is that true?

I was reading the answer from "Stuntman" in this post.

You are perfectly right! Check the current regulator for its voltage drop, add 12V and a suitable little margin.
I have built numerous of power supplies in the past just because the switched, and affordable, ones were not invented. Using a ready built supply might be handy, no mounting of components, no wiring, no casing etc.
Good luck!
//railroader.

Thanks!

androidfanboy:
I just wanted to drive at a higher voltage so the current would be lower (to stay within overall 18W LED power level) using the CC driver so the LED doesn't get so hot.

That doesn't make sense.
18watt into a LED from a 15volt supply, or 18watt into a LED from a 50volt supply is the same heat in the LED.
The CC LED driver might get hotter though, since efficiency drops when there is a higher voltage difference.
Efficiency of a transformer is also much lower than e.g. a switching 19volt laptop supply.
Leo..

Yea, that's just what I figured out. I realize what I said makes absolutely no sense now that I think about it :slight_smile:

Hi,
Does your transformer have any other outputs, is it 36V centre tap or just one secondary?

Tom... :slight_smile:

Dual secondary, 18V+18V (I can get many other voltage levels too). I could tap into one side and get a lower voltage before putting it in series but that might be more work than it's worth.

Hi,
Have you thought about this configuration?

Use 18V instead of 36V
18 * 1.414 = 25.4Vpeak, less voltage to drop.

Tom... :slight_smile:

Yes, but I could also just choose a different transformer :slight_smile:

I like the full wave rectifier, saves on capacitor size too.