Jump to content
ricecube

2.5 gallon LED nano build walk thru

Recommended Posts

Here's a follow up to a suggested build found in an article here called "Full spectrum LED layouts." This is my first LED build and first aquarium in 30 years. Let's get started with the pictures and I'll add all the caveats and pardons along the way.

 

The LED line-up is True Violet, Royal Blue, OCW, and Neutral White. LEDGroupBuy.com makes it easy to figure out what to buy.

DSC02516_zps97de0408.jpg

 

The ABS plastic hood with guts taken out and small fan installed.

DSC02525_zps78fdba30.jpg

 

DSC02522_zpsbfa43627.jpg

 

I wanted a way to make the LEDs adjustable yet not spend a fortune on the fancy PWM modules. The only way I could figure out how to do it was to use linear LM317 regulators and adjust the voltages of the LEDs so that they don't exceed their voltage/current specs. If you'd like a small tutorial on how I did that, leave a comment and I'll follow up. The printed circuit board was leftovers and I used a Dremel tool to slice traces. I don't recommend anyone to follow my hacking suggestions since they are the worse methods... but I just needed to get this project done within the couple of spare hours I had. You could also use perf-board from your local electronics store.

DSC02527_zpsaa980f69.jpg

 

Finished product. Power supply board is connected to the LED heatsink via a DB-9 connector. Notice we have four LED colors, four pots, but five LM317 voltage regulators. The fifth LM317 is used to drop the 12v supply down to 6.8 volts so the fan runs at a happy and silent speed. No need for full 12v fan speed.

DSC02647_zps6b218a86.jpg

 

DSC02649_zps6908403d.jpg

 

Heatsink was cut from leftover to fit the hood and LEDs. Sticky thermal tape was used instead of screws to make it quick and easy. I'm not 100% sure this is a good way to mount the LEDs but they seem to be stuck quite hard even after months of running hot with no fan.

DSC02659_zps8a5197b0.jpg

 

Drilled the holes for potentiometers and 5mm 12 volt power plug. I had tiny 300mA 120v drivers mounted inside this hood before and hated the idea that 120v switching power supplies were hanging in the hood. I also couldn't vary the LED brightness and found algae bloom to be a huge problem. That's what started this whole project. More on that topic later but let's stay focused on LEDs and lighting for now.

DSC02656_zpse8cf6279.jpg

 

Debbie QA'd this hood. It works.

DSC02667_zps08b0d2b3.jpg

 

Finished product from the back side. As you may have noticed, the hood is missing 1/3 length of lighting. This is from the built-in filter taking up 1/3 the tank surface. (no, I'm not hanging an external filter because I'm favoring all-in-one size and simplicity versus maximizing tank space) More pics to come when the tank settles down from the algae bloom.

DSC02665_zps961ebe82.jpg

 

To conclude, the beauty of this build is that everything is compact. It also lends itself to be modified for sunrise and sunset capability... maybe that's next. A 4 or 8 channel digital potentiometer chip can be connected to an Arduino, Pic, or whatever for simple control.

 

Update:

The devastating algae bloom is under control so here's what happened. Using simple 330mA constant current drivers on the LEDs with no dimming resulted in the anacharis plants to form a nasty brown fuzzy algae which ultimately killed the plants. Plus, the color wasn't natural. With the new dimming, the lights can be tuned to look more like real and natural lighting or even that nice un-natural actinic blue found in fluorescents depending on my mood and desire to twist the controls. The anacharis is growing at about an inch per day with no brown fuzzies... no exaggerations on the growth rate. My new problem is having to clip the plants before they choke the lighting, but that's a good problem.

 

DSC02742_zps28d22fb1.jpg

 

The camera never does color and low lighting any justice, but here's a moonlighting shot. You'll have to use some imagination.

 

DSC02747_zps8f91b059.jpg

 

Some side notes... if you're wondering about the rock path up the backside of the tank is for, it's Mr. Krab's ramp to a small perch so he can get out of the water. Apparently fiddler crabs like to be on the beach as much as in the water. But LFS's sell these crabs as if they are fine in regular tanks which isn't right. Also, I have the LEDs on a timer that breaks the lighting into 5 hours for morning, off at noon to 3:00pm, then on another 5 hours. I see the tank at breakfast and dinner. Somewhere on the net I read that algae photosynthesis doesn't really start growing until it sees more than 5 hours. But plants start photosynthesis rather quickly. All I know is that my algae problem is gone, learned a lot about LED lighting, light spectrums for freshwater, photosynthesis, and patience. I can only imagine the sensitivity of reef life to different spectrums and durations.

 

Thanks for sticking thru this post for the 2.5 gallon LED build.

  • Like 1

Share this post


Link to post

Looks good. Did you make a constant current LM317 driver, or did you just regulate voltage?

Share this post


Link to post

In short, it is five simple voltage regulator circuits without the filter capacitors C1 or C2. The adjustable constant current circuit is wicked complicated for this project so I had to chose the easy way.

 

The long and thorough explanation... I took a standard LM317 voltage regulator circuit and chose a resistor value to put in parallel with the 5k pot that will limit the maximum voltage going to the LEDs. I had to measure each led series (the two blues, the two violets, the two whites) to see what voltage would make 500mA flow. Of course it was around 7.2v (for two in series) like the graphs said it would be. The logic is that at maximum light, we want 500mA - no more than that. If we want to start to dim our LEDs to about 275mA, we just need to lower the voltage to about 3.3v each or 6.6v for two in series by simply turning the R2 pot on the outside of the hood.

 

Now that we know what our max output voltage limit should be, we need to know what resistor to put in parallel with the 5k pot (R2). There are lots of equations on the net to calculate Vout against R1 and R2. Well, bottom line, those equations are nice but with a 12v input and a bagful of resistors, I had to simply plug and play several different resistors until the max Vout was 7.2v.

(LM317 schematic)

LM317.gif

 

For example, let's say we use two LEDGroupBuy True Violet LEDs in series at 500mA. This will mean they want to see ~7.2 volts total. (disclaimer, I am not affiliated with LGB, I just credit and reference them because I'm using their parts and graphs) From the graph you can see 3.6 volts (for one LED) will draw about 500mA. So we can either limit current or voltage to achieve the same goal*. However, that asterisk means for single brightness applications, it is more meaningful and safe to use constant current and not worry how many LEDs you put in series if the goal is to pump 500mA thru 1 or 10 LEDs. You won't have to worry about voltage because the circuit will adjust voltage as necessary. In the constant current 500mA scenario, you simply have to adjust the pulse width modulation to adjust brightness... but that takes us down a different discussion path. (let me know if PWM needs more explanation)

TVFV__40112.1333602048.1280.1280.JPG

As you can see, instead of using constant current PWM techniques, we can simply vary the voltage and brightness from Max voltage = 3.6v down to about Min voltage = 2.9v. The LM317 circuit could probably drive 3 LEDs in series for the larger LED builds.

 

Any larger than 3 LEDs in series and now we're talking more than a 12v power supply. (Let's go down this thread a tiny bit) Let's say you want 6 LEDs at 500mA. That means max voltage across the LEDs would be 3.6v x 6 = 21.6 volts. And let's say we have a 28volt 1A power supply power pack from that online auction place or leftover old wireless router. This means the voltage regulator is dropping 6.4v (28 - 21.6 = 6.4) at 500mA. Power = I x V. Watts = 500mA x 6.4v = 3.2 watts of heat. Or, ~2.1 watts at 1/3 brightness at 250mA. That's not tons of heat, but it's a lot for a small LM317 package with 25cent heatsink. And also, there are 4 circuits you need to cool down for a total of 12 watts in a tiny hood that shares LED heat. Bottom line, you'll have to experiment and be cognizant of heat if you want to go to higher voltages and higher strings of LEDs in series using the LM317 method.

Share this post


Link to post

Smart. I'll wait for Evil and Jedi's take on this though; I'm not sure what they'll think of your method.

In any case, it seems very efficient for a simple driver.

 

Also, for larger projects, there are the lm350 and lm338 can take higher currents. 3 amps and 5 amps respectively.

Share this post


Link to post

The big issue with regulating voltage for a dimmable driver is that very small changes in voltage make huge changes in current, and as a result, brightness. With the resistor values you have now, you are limiting your current based on a max voltage, which is fine. There will be some variation based on resistor tolerance, but it should still be fine, as long as you are monitoring what the actual voltage is for each channel (really only needed the first time you power everything up). The single turn pots will make adjustment very twitchy. If it becomes difficult to adjust, then you can switch to multi-turn pots that will give you finer control.

 

It's not the best way to do it, but it's certainly one way to do it. Really though, using an LM317 as a constant current source isn't any more complicated than setting it up as an adjustable voltage regulator. That can be made adjustable with a high wattage resistor to set the max current, and a small low resistance trim pot in parallel to add the adjustment. I've made these before, and they work fairly well. Just not efficiently :)

 

The reason why we don't (normally) drive with voltage is that LEDs are current driven devices. When an LED is tested, the results are very consistent and predictable at a given current. The voltage of an LED varies as it's temperature changes, so you can start seeing why regulating voltage isn't the best idea. The forward voltage of an LED at a certain current can change quite a bit too from LED to LED, so going by the voltage curves in the datasheet isn't the most reliable way to predict where to set the voltage to get a certain current. Driving with current, and letting the driver deal with setting the voltage to the appropriate amount is really the best choice for driving LEDs.

  • Like 1

Share this post


Link to post

As for the 3 and 5 amp voltage regulators, yes, those are very good options given the right heatsinks, mounting positions, and removal of heat. If one wanted to be frugal and mount them on a real LED heatsink sled, I bet it could work very well.

 

Regarding the slight variation on resistor tolerances, I totally agree that the 2% or 5% standard tolerances influence a lot. That's why I manually chose each resistor to parallel the R2 pot and measured the max voltage of each circuit to ensure no chance of burning out LEDs. If the R2 pot becomes noisy, the worse that will happen is that the LEDs will go full brightness or full off for a tiny microsecond while the knob is jiggled. For those who can't envision this, it's like a noisy volume knob on your stereo. I thought of this problem while building the circuit but no noise has been an issue so far. But pragmatically thinking, this setup works extremely well for adjusting the brightness from zero to full... works linearly and nicely as expected and mathematically works in terms of proper currents and voltages since they are proportionally related. No need for multi-turn pots. This acts just like your standard dimmer for house lights.

 

And moreover, I agree this was not the best way to do it. But I was desperate to drop the blues from a massive and destructive algae bloom that literally devastated my tank. I had hours to solve this problem with very little electronics resources to do it. That's what made me think that it's worthy of being shared with those who want an LED solution but don't want to do a super hi-tech build and have limited electronics experience. The beauty of the current situation is that I get to learn what levels of each color produce the best results for this tiny tank by simply turning a knob. 100% blues are not the answer... I can tell you right now.

 

With disclaimers and caveats that might exist at the top of this thread, I would never suggest this build for larger applications unless it's the only method left on the table. The amount of linear heat that needs to be dissipated on top of the need for tighter current regulation really begs to have a more efficient constant current solution with PWM for brightness... so I see your comments as very valid. But I am a proponent for this quick and frugal solution which solved my immediate problem.

 

What I'm really excited about is the idea that this circuit can be easily controlled by a 4 channel digital potentiometer chip... it's begging to have a sunrise sunset function attached to it. I really like the idea of gentle light ramping so fish don't have to experience bright lights slammed onto them from darkness... we would hate the same done to us at midnight. But with the little time I have to hack on things, building a microcontroller and digital pot circuit is just a dream right now. I manually turn the brightness down at night before the timer clicks off so that in the morning it turns back on really dim. I'll get tired of this soon enough.

Share this post


Link to post

A "digital pot" doesn't work with the LM317 circuit unfortunately. Digital pots change the voltage, whereas this works off some kind of reference resistance or something; I haven't looked it up in much detail.

 

However, this is a fantastic solution for driving leds IMO, and finer control can be achieved simply by using trimmers. Although even though this is simple, the DIY-savvy people on this site are few, and by "DIY savvy" I mean they know what a soldering iron is. :D

Share this post


Link to post

UGH... I hear ya, I forgot about the max Vadj being too high for it. As for trimmers, yeah, I wish my closet of parts had some trimmers, but in total disbelief none were to be found. There are a couple of sample circuits out there showing digital pots with the LM317 but it requires a tad bit more circuitry than I'm willing to toss at this (a simple transistor and op-amp per circuit, plus power supply just to run the op-amps... etc. etc.). So it's really unfortunate that my dream has turned into a nightmare for sunrise control. :-( *sigh*

 

Maybe buying PWMs is the only way out of this mess, but now we're talking $$ for the whole setup.

 

There's gotta be a poor-man's version for auto sunrise/sunset... and maybe it's making someone else manually turn up the knobs at 6:30am everyday during breakfast. ;-)

Share this post


Link to post

You can make your own Typhon controller. I'm pretty sure there's a wiring diagram somewhere.

If you used an Arduino Nano from ebay, an lcd from ebay, and buttons from ebay or ripped out of something, you could make a Typhon led controller for about $20.

Share this post


Link to post

Dive in and share your thoughts!

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recommended Discussions

×