2-12VAC to LED?

Thread Starter

sifka4

Joined Nov 9, 2011
8
Hi everybody. I'm a student currently working on an LED project that involves driving LEDs off of 2-12VAC. I have some experience in electronics via my university's lab courses, but we only did a few labs on LEDs and also on voltage regulation/power supplies.

What I am looking for is information that would help me get started on this. While I do know theory behind some of these concepts, I lack the real world experience to tell me exactly what is and what is not possible given the voltage constraints.

Are there prefabricated devices capable of this AC/DC conversion and voltage regulation, or will a bare bones design be necessary? Traditionally, is the AC voltage regulated the rectified, or is it the other way around?

This is a big/long project, so I am not looking for someone to tell me exactly what needs to be done. I just need some guidance on what topics I should begin researching to gain the knowledge necessary to proceed with the project.

Thank you.
 

Adjuster

Joined Dec 26, 2010
2,148
The simplest starting point might be be to rectify the input, and follow this by a switching regulator to create a regulated output current, the voltage being stepped up or down depending on the LED forward voltage compared to the input voltage.

Unfortunately, the efficiency obtained could be very low at the 2V end of the input range, because the rectifier forward drop could be a big percentage of the total, even using Schottky diodes. The conventional four diode bridge may not be the best choice for this. For instance, two opposed (positive and negative) half-wave circuits would develop twice as much voltage for a given diode drop, or a synchronous MOSFET rectifier might be considered.

You might want to ask your lecturer whether it would be permitted to step the voltage up with a transformer first, to get a higher range of voltages.
 

Thread Starter

sifka4

Joined Nov 9, 2011
8
Adjuster, thanks for your reply. I'm not familiar with Schottky diodes or the opposed half-wave rectifiers so I will look into those.

I don't think a transformer would be possible since the entire apparatus must fit into a space roughly the size of a tuna can. I imagine much of the space will be occupied by a heatsink to accommodate the white, HPLEDs.

What are the limitations of skipping the AC/DC conversion and driving the LEDs directly off of the AC supply?
 

SgtWookie

Joined Jul 17, 2007
22,230
LEDs are DC devices; you can only run current through them one way, and the current must be regulated.

You may have to look into "synchronous rectification" to get any kind of output from as low as 2VAC. Google that phrase. It will require some very low threshold power MOSFETs.
 

wayneh

Joined Sep 9, 2010
17,496
I agree that the current must be regulated, but the LED is, after all, a diode. Feeding it AC will light it up on the positive wave. The modern Christmas LED light strings don't use rectifiers.
 

Adjuster

Joined Dec 26, 2010
2,148
LEDs tolerate reverse voltage poorly though, typically 5V is about the limit. With up to 12V coming in some means would have to be provided to keep this at bay. Simple lamps may get round this using AC feeding back-to-back parallel pairs.

Depending on the definition of "2V" (RMS?) and what colour / Vf the OPs LEDs are, s/he might just get one pair of LEDs to work straight off the supply, but the variation of brightness from 2V to 12V would I think be pretty severe with just a series resistor ballast.
 

wayneh

Joined Sep 9, 2010
17,496
LEDs tolerate reverse voltage poorly though, typically 5V is about the limit.
Agreed. The challenge of this project is the voltage range. It's easy to solve at any one voltage, but not so easy for that wide range.

Huh, that makes me wonder. What if you bucked all voltages down to 2V and then just use a joule-thief type boost circuit to power LEDs? Sort of a "lowest common denominator" strategy.

But I can also imagine a bargraph type approach that just switches in more LEDs in series at as the voltage climbs higher.
 

Adjuster

Joined Dec 26, 2010
2,148
My advice to the original poster would be to get the constraints on this design as well clarified as possible:

How many LEDs, what type, what power level (Vf and If)?

Does this thing need to meet any minimum efficiency target?

How much is the brightness allowed to vary?

A fuller statement of the input requirement would help too:

Is it RMS sine wave,? What frequency? Maximum available current?
 

Thread Starter

sifka4

Joined Nov 9, 2011
8
My advice to the original poster would be to get the constraints on this design as well clarified as possible:

How many LEDs, what type, what power level (Vf and If)?

Does this thing need to meet any minimum efficiency target?

How much is the brightness allowed to vary?

A fuller statement of the input requirement would help too:

Is it RMS sine wave,? What frequency? Maximum available current?
The number of LEDs is part of the design, it doesn't matter so long as 600-700 lumens are achieved (in the case of an array, one burnout should not disrupt the other LEDs). The current drawn should be no more than 3A, obviously less is desirable. I'm looking to maintain as constant of an output as possible, but I'm doubting this is not possible with a supply down to 2VAC. The important thing at the lower voltages is that the LEDs stay on. Input is a 60Hz sinusoid.
 

Adjuster

Joined Dec 26, 2010
2,148
This thing is somewhere in the region of 10W then, depending whether you have access to the latest 100 lm/W marvels.

You can probably forget about linear regulators, as the efficiency is going to be rotten with such a wide voltage range. Your little tuna can could get very hot, and LEDs don't like that. This definitely looks like a switching application (but you probably knew that).

By the time you have got through Bill Marsden's notes, you should be well aware that LEDs require current drive for safe and consistent results. I'm going to reiterate it anyway, because it is a source of so much difficulty when people first encounter them.

A typical LED has a voltage/current characteristic where over a small voltage range, the current rises rapidly from a negligible amount through the rated level and up to levels which would cause the device to burn out. The voltage range over which this happens is fairly similar for devices of a given type, but there is generally enough variation with temperature and between devices to scupper any serious attempt at constant-voltage operation. Cheap little torches that may seem to break this rule rely on weedy batteries with significant internal resistance, driving low-power devices.

Good luck with this, it sounds both interesting and challenging.
 

Wendy

Joined Mar 24, 2008
23,415
Over the long haul it is better to acually measure the Vf (forward dropping voltage) of the LEDs you are using. Constant current sources tend to bypass this need, unless you are running it very close to the edge.
 

Adjuster

Joined Dec 26, 2010
2,148
It occurs to me that, especially if this is an advanced project, your trachers may not be content with a simple rectifier-input solution. Such systems cannot easily provide a good power factor, and they draw very non-sinusoidal input currents. Modern lighting systems are being constrained to have better power factors, and perhaps this will be mirrored in your project.

Typically this involves a switching converter at the very input of the system. My initial gut reaction was that this is simply too difficult with the low voltages you are dealing with, but I am no specialist in this field. I think you should perhaps raise this issue with your lecturer or tutor. Maybe I ought to apologize in advance if this turns out to be a complete red herring, but I'm concerned not to give you unreliable advice. In the end, remember that you need to satisfy your examiners, so it is important to understand their requirements as well as possible.
 

Thread Starter

sifka4

Joined Nov 9, 2011
8
Thanks everyone for your input. Unfortunately, I do not have any experience with switching regulators, so it's something I will have to read up on. I am not really concerned about power factor at this point, the most important part is that the solution simply works. I believe the voltage range is so wide in order to account for "real world" transients in such as a large motor starting. The LEDs simply have to stay on although it is assumed there will be a brief, significant drop in output.

Assuming that the times when voltage drops extremely low are infrequent and short in duration, I have an idea. I have some experience in car stereos, and many people think that adding a large electrolytic capacitor (<1F) will reduce voltage dips and supply energy to the amplifier(s) quickly (low ESR) when the alternator can not keep up. While this doesn't necessarily work in automotive electrical systems due to the current limits imposed by the alternator, I am curious if it could help here where the system can provide plenty of current. Could a large enough capacitor help maintain voltage when there is a significant draw on the system for a short period of time?
 

wayneh

Joined Sep 9, 2010
17,496
Could a large enough capacitor help maintain voltage when there is a significant draw on the system for a short period of time?
Just depends what the terms "large enough", "significant draw" and "short period" mean, in terms of numbers. If you can define those terms, the answer is easily calculated. Of course the answer is yes you can use a big capacitor, but the real question is, can it be practical? That answer is often, no.
 

Thread Starter

sifka4

Joined Nov 9, 2011
8
Hey folks. I'm at the final stage of this project, but I'm running into an issue. I've decided to persue an AC only design, so DC power supplies/current sources will not work in this application. I got the idea from a string of christmas lights a few months back. Here is the design thus far:

I am using 2.9V LEDs (CREE XT-E WHT). Because this is an AC application, imagine that I will have a mirror of the circuit I am about to describe that will work on the negative cycle of the AC wave.

I have one LED connected across the source that will turn on at 2VAC. I then have a string of 3 or 4 (havent decided on the final number) connected across the source; this will not turn on at 2VAC which is fine. As I increase the input voltage from 2-12VAC, the lone LED will get brighter, and eventually the string of 3-4 will turn on.

You will notice I haven't mentioned resistors yet, and that is where the problem lies. If I put a resistor on the lone LED, it needs to dissipate quite a bit of power (8W or so) at 12VAC and is extremely inefficient. I am trying to keep the current through the lone LED and the string at 300-600mA each.

Is there any other solution other than having a heavy duty resistor on the lone LED in this application? I just need the lone LED to be on at lower voltages, and the string to be on at higher voltages for greater light output.

Thanks for your insight.
 

wayneh

Joined Sep 9, 2010
17,496
Just thinking out loud here. Could you use a zener in parallel to the LED (or LED plus resistor), to limit the voltage there, AND then a resistor in series with the zener to limit current thru the zener and the LED+resistor. The challenge would be to set the total current to the zener when the input is at 12v, to stay within spec for the zener. At any lower input voltage, it should be able to drain of enough current to hold voltage on the LED.

There are a lot of other ways to do this but I believe you need small and not too complicated?
 

mcgyvr

Joined Oct 15, 2009
5,394
Forget voltage/current,etc... The thermal issue will stop this project dead in the water once its actually working..
If you are still planning on putting this into a sealed "tuna can"/Altoids tin what you will have is a red hot brick and a circuit that will burn up in minutes and GREATLY decreased LED life.
30W/ft^2 in a small can AIN'T gonna happen without some SERIOUS forced air.. And I mean SERIOUS forced air.
 
Top