Connect with us

Reducing 12VDC to 8.2VDC

Discussion in 'General Electronics Discussion' started by Quixote, Mar 13, 2011.

Scroll to continue with content
  1. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Hi guys,

    I am building an automated aquarium lid. For now I am setting up LED lights to be tied into my automated lights. I have LED strips that have 128 LEDs on them, and I will be using white, red and blue lights (for optimal plant growth). I am using a PC power supply. The white and blue LED strips require 12 volts DC, but the red ones need 8.2. Each strip uses about 1 Amp. I need to know how to bring the voltage down in a safe way from 12 to 8.2 volts.
    I found this site that shows how to use Zener Diodes and a resistor:
    http://webcache.googleusercontent.c...=clnk&gl=ca&client=opera&source=www.google.ca

    The problem is that I need to have 4 LED strips connected to the same voltage source, meaning that the amperage jumps up to 4 Amps. Using that page, I can see that there are no Zener Diodes that can handle that amount of wattage that I need (about 33 watts). Can I use more than 1 diode and if so, what arrangement would I use? Would I have them connected in serial or parallel?

    Thanks in advance!
     
  2. davenn

    davenn Moderator

    13,865
    1,956
    Sep 5, 2009
    Hi Quixote,
    welcome to the forum

    easiest way would be to use a voltage regulator ... say a Buck DC-DC converter or say a regulator like a LM338T adjustable 5A capable regulator.
    google search a datasheet for a LM338 to see how to use it :)
    it would need to be very well heatsunk drawing that sort of current

    Dave
     
  3. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    When you say a strip requires 12V it tells me it might incorporate current stabilizing resistors, but when it says 8.2V I start to suspect that it might not.
    For LED's it's the current that's important. The voltage drop across them is approximate and just for reference for calculations. How about a link to the strip data?
    To avoid the high power loss mentioned I'd use dedicated switchmode LED drivers instead of regulators or resistors. They can be found quite cheaply these days.
    Another thing to observe is that PC PSU's are/were usually not made to be loaded on the +12V only, they need a significant load on the 5V (&/or 3.3V) to regulate properly.
     
  4. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Hi,
    Thanks for the pointers!

    I totally forgot about having to put something on the 5v. Here are links to two of the LED strips that I'll be using:

    http://www.dealextreme.com/p/8w-128...num-alloy-light-strip-white-12-12-5v-dc-13087

    http://www.dealextreme.com/p/8w-lsh-15ni-120-lumen-red-128-led-light-strip-aluminum-shell-8-2v-13547

    I'll look into switchmode LED drivers as well as the LM338T, since I want to be certain it is power efficient and done the best way possible. I don't mind spending a few extra bucks since I've already spent a fair amount on this project already. Can I use that with my PC power supply? I have already done most of the wiring, so I want to keep it in the system. Can you give me a link to something that might help me? I would prefer to buy a ready-made solution than to try to build my own since I am pressed for time and it seems like this is getting a little more complicated than anticipated.

    Thanks again.
     
    Last edited: Mar 13, 2011
  5. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    DX is very poor at presenting information about their products, and the reviewers are not much better at knowing what they're doing - or what needs to be known.
    However (regarding the white strip) one of them happened to mention two LED's burned out quickly, and another mentioned he could run it down to 6V.
    That tells me there are 64 (parallell) strings of 2 LED's (in series) each, probably with one resistor per string. Running it at around 12V seems to be acceptable then.
    So far no-one seems to have bothered to check out the red strips though, so it's anybody's guess as to its real data. A real good closeup pic might reveal something.
    But, trusting 8W @ 8.2V = 975mA is correct, you could try using a 3.9V 5W zener for each red strip, or a 3.9 Ohm 4W resistor each.
    What's best depends. If it has internal resistors then using a zener is best. If it has no internal resistors then the resistor might be better (unless the LED's start to pop).
     
  6. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Thanks, Resqueline.
    Currently I have one white LED strip set up in the lid of a small aquarium and it works great. I have attached it directly to a 12vdc adapter without adding any resistors. I haven't been running it for too long, however, so I can't really comment on its lifespan.
    It is impossible to see if there are resistors just by looking at them. I have inspected them thoroughly and there is no indication. I would need to strip one down, and at $25 a pop, I'm not going to do that.
    The way I have set up my power supply, I have one terminal (with a positive and negative connection) that all four red strips will be hooked up to. The reason for this is because I have limited relays on the automation hardware that I will be using and lots of other LED strips to control as well. I will be turning on a certain number of strips with delays between each new set turned on so that I can simulate a sunrise (a blue, red, purple and white sunrise ;))
    Is there a way that I could modify the voltage for all four red strips at once without having to attach components to each individual strip inside the lid? I'd like to stick to the way I have planned since I've put a ton of work into the power supply box already.
    Thanks for all of your help.
     
  7. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    My guess is that neither have series resistors. As such I would be extremely reluctant to recommend connecting them to a voltage source.

    Here is something you could build for the lower voltage (red) emitter. Obviously it would need some modification to operate at a lower voltage.

    If you had access to more than 12V, then a similar circuit could also be used for the 12V white LED array.

    Alternatively, you could look at constant current LED drivers. I believe DX has some of these that are mains operated. Naturally, they would require mains wiring, and mains wiring near water requires knowledge of what you're doing.
     
  8. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    I can't find any constant current LED drivers on DX that would supply me with 8.2vdc at 4 amps with an input voltage of 12vdc.
    Thanks for the schematic, but that is definitely outside of my scope of ability and looks like it would be way too much trouble for me for what I am trying to do.
    Can you explain to me why it seems to be the consensus that I should use resistors? I am using a very high-end PC power supply and the voltage appears to be extremely stable at 12vdc.
    Isn't there a simple way to step down the voltage with DC just as you can with AC and a transformer? I can't believe this simple project is starting to look like I should just abandon the idea and use DC adaptor(s) plugged into the wall.
     
  9. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    There is a sticky about driving LEDs and it telly you why not to use a voltage source. The very quick summary is that as the device heats up, it can draw more power and eventually destroy itself. Series resistors help to reduce this tendency.

    Constant current sources are the best (for lots of reasons I won't go into here) however they are not trivial to come across. The ones from DX are mains powered and it can be tricky to figure out which one is appropriate.

    I would insert a series resistor that would drop about 1V at the operating current for the white LEDs. This shouldn't reduce the light output too much, but will help prevent thermal runaway. From what I read, they are 350mA @12V nominally, so a series resistance of around 3 ohms rated at about 1W (anything between 2.7 and 3.3 ohms would be OK).

    For the red LEDs. find out the voltage (8.4V?) and the power (8W?) and use this to determine current (8/8.4 = 0.95A) Use this to calculate the resistance required (12-8.4)/0.95 = 3.8 ohms. 3.9 ohms is easily available and you'd want a 5W resistor.

    If my figures don't exactly correspond, plug in the correct ones -- the results should be similar. I haven't given you calculations to determine the power, but unless the current and/or resistance varies a lot, there is sufficient margin (but let us know anyway)
     
  10. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Thank you. I'll give it a shot and test the circuits before hooking any LEDs up. I'll let you know how it goes once I receive the red LED strips (Still in transit). I have the white ones though, so I can work with those for now.
     
  11. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Can you guys proof this schematic for me, please?

    I plan on using 1.6 KOhm 10 Watt resistors for the first set of 4 and then a couple of 390 Ohm 2 Watt resistors for the second parallel set.

    If my calculations are correct, this should give me about 8V. I'm going to glue the resistors to a heatsink with thermal epoxy and use a fan to make sure nothing gets fried.

    Thanks!

    [​IMG]
     
  12. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    Get rid of R2.

    Place a single 4 ohm 5W resistor in series with each load. At 1A the 4 ohm resistor will drop 4V (and dissipate 4W) leaving 8V for the load. The voltage will vary a little if the 8V 1A is nit the exact load, but it should be close.

    Note that the 4 resistors will be dissipating 16W of heat (which is a lot) and the LED strips will be dissipating 32W (which is a hell of a lot) The LEDs will almost certainly require heatsinking and you may be better off with 10W resistors. if there is not some airflow in the area where they are placed.
     
  13. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Hi Steve,
    Thanks for your reply.

    Placing a resistor in series is pretty much not an option. I have an enclosure that houses the power supply, along with fans and heatsinks and terminals for wires leading to the lights in a remote area (please see my second reply). The LEDs do have heatsinks on their backs and I may put a fan where I am situating them as well.
    The terminal is on the enclosure that the four red LED strips will be attached to and I don't want to have resistors on the exterior. Not only would I not like the way it looks, but the heat factor would make me uncomfortable and a bit nervous. If I set up everything in the enclosure, I have a large fan that is pumping air through it and multiple heat sinks to attach resistors to.
    If I did, hypothetically, place a single 4 ohm 5W resistor in series with each load and remove the two R2 resistors, would the bulk of the heat not be dissipated by the 4 R1 resistors?

    I have placed a 3.3 Ohm 1W resistor in series with each of the other loads since my enclosure and relay board could afford the real estate to provide them each with their own terminals. I only need to resolve the 4 red LED strip situation. :mad:

    Thanks for your patience.
     
  14. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    The resistors labelled R2 in your circuit will only serve to increase overall dissipation.

    Run 4 separate wires from the LED bars back to the resistors if you want. There's no reason why they need to be close to the bars.

    You *could* place all the resistors in parallel, and the loads in parallel, but you can't guarantee that the current will be shared equally and you may accelerate failure.
     
  15. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Well, I guess I'll have to tear out the terminal and see what I can come up with to connect each strip separately. Major PItA, but if that's what I have to do, so be it.

    So just to confirm, I leave the R1 arrangement with the 4 parallel 1.6 KOhm 10 Watt resistors, remove the two R2 resistors and it's associated circuit, and then place a single 4 ohm 5W resistor in series with each of the LED strips anywhere in their individual circuits? Maybe I'll put the 4 Ohm resistors somewhere on the positive wire leading to each strip.

    Thanks.
     
  16. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    No, remove all the resistors, and place a 4 ohm resistor in series with each LED strip.

    The 1.6k resistors will limit the current to a handful of mA rather than the 1000mA that you apparently require.

    Note that I am relying on your rating of 8V 1A being correct.

    I would tend to err on the side of slightly higher value resistors as this will reduce the LED current and prolong their lives.
     
  17. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    Ok, thanks. I'll try that.
    What would happen if I use 10 Watts resistors? Would that just give them more resilience to heat, or would that affect the current in any way?
     
  18. Resqueline

    Resqueline

    2,848
    2
    Jul 31, 2009
    Using higher wattage resistors will only result in them having a lower surface temperature, nothing else. Higher resistance = lower current = lower dissipation = lower temperature.
    Here's the preferred wiring diagram, just to make sure.
     

    Attached Files:

  19. Quixote

    Quixote

    21
    0
    Mar 13, 2011
    OK, thanks guys!
    I'll try it out and let you know how it goes once I receive the red strips.
     
  20. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,497
    2,839
    Jan 21, 2010
    My advice to you is to connect up a single strip first, and to do so with a higher series resistance -- maybe 2 of the resistors.

    Measure the voltage drop across the resistors and ensure that the LED strips are operating well within their limits.

    Then reduce the resistor to just the value suggested and do the measurements again.

    It is always best to minimise the risk of error, even if you're sure you're doing the right thing.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-