Author Topic: Good 12V regulator with Enable Capability (and a minimum 200 mA current output)  (Read 2291 times)

0 Members and 1 Guest are viewing this topic.

Offline SnoopKattTopic starter

  • Contributor
  • Posts: 14
Hi,

I'm looking to make an LED array that is controlled by an ATTINY13 microcontroller. Essentially, my goal is to have the microcontroller enable each 12 V regulator one at a time every second (in a loop), and each regulator will be connected to about 60 LEDs (I don't need them terribly bright, so I'm thinking of going for 4-5 mA for each LED instead of 20ish). I saw some 5 V regulators with enables (or more like 4.5 V - TI LP5907 - http://www.ti.com/lit/ds/symlink/lp5907.pdf), but that gets crazy because each LED would need its own resistor instead of being able to pair up two or three LEDs to a resistor. I searched around a bit on a few different IC websites like TI & Linear, and I also used Google to try and find one, but I wasn't really able to find anything.

Maybe I'm using the wrong keywords and I'm sounding like a dolt who can't search, but would anyone happen to know of a good 12V regulator with an enable output that could output at least 150-200 mA (probably more on the 200 mA side to leave a little head room)?

Thanks!
 

Online georges80

  • Frequent Contributor
  • **
  • Posts: 916
  • Country: us
There's plenty of adjustable regulators with enable pins.

But for you application, why not put an nchan FET on the ground feed from the series LED strings and switch that with your uC. It'll perform faster than switching a regulator on and off. It also means you can share the regulator with a bunch of strings since the switching is via the FET on each string.

cheers,
george.
 

Offline uncle_bob

  • Supporter
  • ****
  • Posts: 2441
  • Country: us
Hi

Why regulate the current driving an LED with a linear regulator? Its a very inefficient way to do things. LED's are *not* very picky about supply ripple or tight regulation. A switcher is a very good fit for driving them. Just about every switcher chip ever made has some sort of enable / disable capability.

=====

To at least sort of answer the question:

The LT1764 is a nice low noise adjustable regulator with an enable pin. They can be used for a wide range of things in a whole bunch of projects. They are not a bad thing to stock up on and have running around when you need a regulator in a hurry.

Bob
« Last Edit: March 07, 2016, 02:27:38 am by uncle_bob »
 

Offline SnoopKattTopic starter

  • Contributor
  • Posts: 14
There's plenty of adjustable regulators with enable pins.

But for you application, why not put an nchan FET on the ground feed from the series LED strings and switch that with your uC. It'll perform faster than switching a regulator on and off. It also means you can share the regulator with a bunch of strings since the switching is via the FET on each string.

cheers,
george.
Thank you, that sounds like a good way to do it. Probably part of my issue was that I was searching for fixed regulators instead of adjustable ones. I'll also look into using multi channel FETs!

Hi

Why regulate the current driving an LED with a linear regulator? Its a very inefficient way to do things. LED's are *not* very picky about supply ripple or tight regulation. A switcher is a very good fit for driving them. Just about every switcher chip ever made has some sort of enable / disable capability.

=====

To at least sort of answer the question:

The LT1764 is a nice low noise adjustable regulator with an enable pin. They can be used for a wide range of things in a whole bunch of projects. They are not a bad thing to stock up on and have running around when you need a regulator in a hurry.

Bob
That's a good idea too; I see that chip is an adjustable switching regulator. Probably wasn't including switching regulators in my search either, thanks for the tip and the recommendation! Off to digikey :)
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 13119
Use a low side NPN transistor for each chain with the base driven directly by your MCU pin.   Use an emitter resistor to set the current.   I=(Vdd-0.7)/R, where Vdd is your ATTiny supply rail.  With 5V Vdd, 1K would give you 4.3mA.   String as many LEDs in series as you can for each chain, but make sure that Vsupply-N*Vf still leaves you enough headroom for the transistor to provide a constant current.  i.e. it's collector needs to be above the Vdd supply when on.   As the transistor provides a constant current you can have different numbers of LEDs in each string without changing the resistor value.   If you can run the LEDs at greater than 12V off an unregulated supply, do so, but check there is enough headroom allowing for any ripple with the supply fully loaded.  Any 'jellybean' small signal NPN with an adequate Vce rating should do nicely.   If you can use a high enough voltage, e.g. 48V, you can light 60 LEDs with current control in only five series strings even if LED Vf is 3V. It doesn't get cheaper or easier, and it may even be worth using a boost converter to get a 48V rail and enable that.
« Last Edit: March 08, 2016, 12:23:33 am by Ian.M »
 

Offline botcrusher

  • Regular Contributor
  • *
  • Posts: 192
  • Country: ca
If one regulator per (x) string of lights isn't a needed part of the project, any old ATX supply with work magic.

If you do need do switch several strands on or off, use the supply with a transistor or SSR for each section.
That gives you the master enable on the atx to pull low (use another SSR and ground?), and a sub-enable per segment that you can drive High. If you want dimming, the transistor opton would let you do that.

I'd imagine that this would be more efficient than lots of small supplies ( minimum possible efficieny allowable is 60%, you can get gold rated PSUs at 80% or above) and nearly 100% idiot proof protection circuits.
I've shorted, sent ESD down the lines, drawn more current that I'm supposed to, and even run two in parallel with no aditional circuitry and haven't killed this one yet.
(Please note, by ESD i mean accidently shifted a carpet when a wire fell down and watched the psu suddenly cut out, assumedly from static discharge.)

Edit: terribly sorry, first version said "thyristor" but the bistable nature of some of them makes it completely unsuitable for use as an ON/OFF switch. Mosfet or SSR is a surefire way to do the trick.
« Last Edit: March 08, 2016, 02:27:22 am by botcrusher »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf