DIY LED driver for reef lighting

Great Job

Great Job

I have been following along, but had to take a break in October:sad2:. I have read the original thread and this one (finally caught up). Great work with lots of informative information:bigeyes:. I was toying with this, but really wanted fewer parts and easier configuration for different currents. I know stugray mentioned a while back the STCS1 - which looks promising except for the thermal pad on the bottom.

Is anyone interested in starting a new design from scratch (probably a new thread). If not I may just try the one here since all the work is done. Once again great job!!


Please PM so we don't cloud what is going on here. Thanks
 
TheFishMan, I might try that chip eventually just to see what it's like. Anecdotally, from two or three people that have tried it (SpacedCowboy, active in this thread, and a user who's name I forget that was active 3 or 4 months ago in one of the other threads) it's REALLY sensitive to keeping the voltage drop small enough such that the chip doesn't overheat. That and the obvious difficulty of that ground pad if you don't have a reflow oven. Might want to check in with SpacedCowboy on that chip. If you guys do get something going in another thread, please at least link it here!

And if you guys want to put your material online, you can use the google code project I started. I intentionally gave it a generic name (HP LED) so we could use it beyond the NCP3066 driver I did, if we wanted to.
 
TheFishMan, I might try that chip eventually just to see what it's like. Anecdotally, from two or three people that have tried it (SpacedCowboy, active in this thread, and a user who's name I forget that was active 3 or 4 months ago in one of the other threads) it's REALLY sensitive to keeping the voltage drop small enough such that the chip doesn't overheat. That and the obvious difficulty of that ground pad if you don't have a reflow oven. Might want to check in with SpacedCowboy on that chip. If you guys do get something going in another thread, please at least link it here!

And if you guys want to put your material online, you can use the google code project I started. I intentionally gave it a generic name (HP LED) so we could use it beyond the NCP3066 driver I did, if we wanted to.

FYI on the STCS1. If I keep the current down to ~500mA, I can happily run the STCS1 for a week or so at 24/7. I've no reason to suspect it wouldn't run longer, that was just how long I left it on for. It was cool to the touch at the end of that week.

When I switched sense-resistors to get 700mA, it ran fine for a day, but the next day I went into the garage, it had shut down. I'm not yet sure what went wrong, whether a LED burnt out, or if the chip melted itself - I've been busy, my wife's birthday is more important to her than LEDs :)

As DWZM mentioned, the STCS1 is *very* sensitive to input voltage - to the point that having a potential difference of even 2v with respect to the input supply and output voltage will cause meltdown in minutes, as well as probably burning out/melting the input bypass resistor. I've only successfully run it on tunable power-supplies although you could place a voltage-regulator in front of the chip to manage that. I think I read about someone doing just that, to keep the chip temperature in the manageable range.

I have an SMT oven, and all this is with the central underside pad solder-pasted to large copper-fill areas to dissipate the heat. It's only 0.5oz Cu, but it's still a lot of heatsink.

At this point in time, I'm going to build a version of DWZM's circuit in a config that suits my needs, and see how easy it is to work with that one. If it's significantly easier, I'll probably ditch the STCS1 (there ain't no NIH syndrome here, baby :) ). There may be more components on DWZM's circuit, but because of the STCS1 price, it's not significantly more expensive IMHO to go with DWZM's design.

Simon.
 
Great work,

Does anyone know what the minimum runs are to contract the building of this unit out to a shop in China? Your project seems almost ready to go as far as required documentation. I have read in the back on electronics mags of orders as small as 500 units. Maybe we could coordinate and get a group going and benefit from economies of scale on component prices and get the drivers AT cost. AND have someone else build them for us? If we have excess I am sure the rest could be sold at cost here on the club to recoop any gap. Just a super high level thought for now but I am curious to hear if the key players would be interested in pursuing something like this.
 
barderer, so far my interest has been limited to doing a (mostly) through-hole board that novices could easily make at home. I'm going to get a large run of the PCBs made to get my own costs down, and distribute the others to hobbyists on here at cost.

If I were to try to do this design and have someone else assemble it, I'd probably change it to all SMT (I'd guess that would be cheaper), make it more compact, and change some other things. That said, if the min quantities are really in the hundreds, I'm guessing it would be a challenge. Right now I've probably got solid commitments for 50 - 100 of the PCBs and I really don't want to try to run a wide-scope group buy or other such venture.

That said, if someone else wants to bankroll it and enough parties magically appear to make it viable, I'd be more than interested in participating. :D
 
Does anyone know how many leds would be needed for an 80 gollon tank? I'm not sure what kind of corals I want yet, but I would like to have enough light to keep anything I want. Everything I check out doesn't give a basic rule of thumb to how many leds are required for applications. I have a friend that said he thinks his 12 leds are equal to a 150w MH. I'm not sure this is true, but his corals seem to think so. This is the kit he bought http://www.rapidled.com/servlet/the-35/12-Premium-LED-Do-dsh-It-dsh-Yourself/Detail

Is this a good kit? I have NO experience wiring up lights, so this easy kit seems like a good purchase for me. I've seen that these are the same type of leds that everyone seems to reccomend, but I just don't know how many I need. My main concern is power comsumption. My electric bill is high enough and I don't want to watch my meter spin every time I turn on my lights. Thanks for the help guys.
 
Power supply

Power supply

Anyone know how many buckpucks/leds you could run off of a 24v 8.3a power supply? I know I have seen the math somewhere but this post has become too large to search through. Thanks for the help.
 
8 strings of 6 running at 1000mA.
They take about 3.5V each, so 3.5 X 6 = 21V. Then you can put 8 of them in parallel and be pulling 8 amps. That's a lot though. You might want to limit it to about 6 amps or your power supply will get really hot.

Also, if you run them at 700mA, you should be able to get 11 strings of 6, again at upper the limit of the power supply

** edit**
Actually, I think you saw the math in a different post. This one isn't very large at all...
 
You guys might want to check out the mothership LED thread:

http://reefcentral.com/forums/showthread.php?t=1678127&page=43

It covers a wider range of topics, while this thread is mostly concentrating on the DIY driver project.

Does anyone know how many leds would be needed for an 80 gollon tank? I'm not sure what kind of corals I want yet, but I would like to have enough light to keep anything I want. Everything I check out doesn't give a basic rule of thumb to how many leds are required for applications. I have a friend that said he thinks his 12 leds are equal to a 150w MH. I'm not sure this is true, but his corals seem to think so. This is the kit he bought http://www.rapidled.com/servlet/the-35/12-Premium-LED-Do-dsh-It-dsh-Yourself/Detail

Is this a good kit? I have NO experience wiring up lights, so this easy kit seems like a good purchase for me. I've seen that these are the same type of leds that everyone seems to reccomend, but I just don't know how many I need. My main concern is power comsumption. My electric bill is high enough and I don't want to watch my meter spin every time I turn on my lights. Thanks for the help guys.

That's an OK kit. Some of rapidled's kits use the Q4 bin cool white XR-E, IMHO you want to avoid those. But this kit uses the better Q5 bin. However, it uses a less-versatile driver. Meanwell also makes a driver called the ELN-60-48 that allows for dimming, while the one in that kit does not.

For your 80 gallon tank, you might want 50 - 75 LEDs. It'll depend on the intensity you want, drive current, optics, etc. But it'll be a vastly smaller impact on your power bill than any comparable lighting, no matter how many you use (within reason of course).

Regarding your friend's comment about 12 LEDs comparing to an 150w MH, it's hard to make generalizations like that, but if you drove them high and used moderate optics, and were comparing to a poor or average MH lamp, that would probably be a fair comparison. In general, you can assume a LED rig will be ~40% the power of a comparable MH rig.

Anyone know how many buckpucks/leds you could run off of a 24v 8.3a power supply? I know I have seen the math somewhere but this post has become too large to search through. Thanks for the help.

You're probably thinking of the mother post I linked above. For reference, a buckpuck running at 24v with 6 LEDs can be assumed to draw about what the output current will be. This is a rough estimate, because - being a buck driver - it'll actually need LESS current than the output current, the difference depending on the drop in voltage it has to create. However there will be a ~10 - 15% efficiency loss, which is usually about what the drop in current is, so it ends up being about an even tradeoff. So, if you're driving at 700mA, you can assume each buckpuck will draw about 700mA.

Now, just multiply that number by the number of buckpucks you want to use, and add 20 - 30% for some headroom (power supplies lead short, hot lives when pushed above 70 - 80%). Or do the math the other way - cut 30% off your power supply's rating and divide by drive current. So, for your 8.3A supply, you can happily load it to 6 - 6.5A. That's 8 or 9 buckpucks at 700mA. In the end, the power supply ends up being one of the cheapest major components in a build like this, so IMHO it's safe to leave a good head margin like this vs. loading them to the hilt.

EDIT - jma beat me to the punch by 7 hours! That's what I get for opening a thread at night, leaving it open, and replying in the morning without refreshing. :D
 
Regarding your friend's comment about 12 LEDs comparing to an 150w MH, it's hard to make generalizations like that, but if you drove them high and used moderate optics, and were comparing to a poor or average MH lamp, that would probably be a fair comparison. In general, you can assume a LED rig will be ~40% the power of a comparable MH rig.


Thanks! Any suggestions on kits then? I don't want to have to wire up a driver, but I don't want to spend a fortune either.
 
Thanks! Any suggestions on kits then? I don't want to have to wire up a driver, but I don't want to spend a fortune either.

rapidled's kits are fine, just make sure you know what you're buying in terms of LED bin and driver capability. Also might want to check out the other usual sources: ledsupply.com, nanotuners.com, etgtech.com, dealextreme.com and so on.

Also, please check out the big LED thread I linked back a page - it focuses on LED builds in general, while this thread is focusing specifically on DIY drivers. :)
 
Presenting, the dual-driver prototype:

IMG00099-20100114-1411.jpg


*some assembly required.

Thank you again to Simon for prototyping the pcb. I will report back when it's assembled. Probably this weekend.
 
Anyone else not getting orders from DealExtreme? I ordered my LED's well over a month ago and they are still sitting as temporarily out of stock and they won't respond to email requests of when they may have them in stock. In their catalogue they show that they are in stock. And, of course, my credit card has been billed.
 
I've gathered some of the material I've been promising for weeks (months?) in the various LED threads and figured if I make a thread about it, I'll be motivated to get the rest of the material together if there's enough interest.

If you're new to HP LEDs for reef lighting, here's some reading:

The Mother Thread:

http://reefcentral.com/forums/showthread.php?t=1678127

Some others:

http://reefcentral.com/forums/showthread.php?t=1662682

http://reefcentral.com/forums/showthread.php?t=1751598

http://reefcentral.com/forums/showthread.php?t=1672708

I'm starting this thread to detail the DIY driver I've mentioned a few times in those various threads. I decided to DIY drivers for a few reasons:

1) It seemed like an interesting project for someone like me with moderate DIY electronics skills to learn some more and expand my skillset
2) It gave me control over the end product I was getting, vs. living with the limitations of a COTS product.
3) It'll end up being a few bucks per driver cheaper than any of the commercial products that compare, though that's NOT accounting for time involved.

If you want a fun DIY project involving LEDs, read on. I'll be doing this thread in sections as I get time to document this project. Also, for all you "real" EE's out there, if you'd like to rip apart my design, please do so. It'll help me learn. :D

dont wanna straight hijack your thread, but heres another link for more reading..

http://blog.makezine.com/archive/2010/01/photon-blasting_60w_rgb_led_array.html
 
A note to emphasize something for anyone trying this themselves. THIS IS A BOOST DRIVER! NOT A BUCK DRIVER!

I explained earlier, but a I think a quick recap for emphasis is due.

With a buck driver, i.e. a buckpuck, you string together a bunch of LEDs in series, add up their forward voltages, then try to get a power supply thats ~2v higher than the Vf of the LEDs. This is because a buck driver cuts the voltage going to the LEDs.

With a boost driver, like this design, THAT WILL FRY THE LEDs. You MUST use a power supply such that the Vf of the LED string at your target current is HIGHER than what the PS is putting out. A boost driver CANNOT cut the voltage to something lower than what the power supply is giving - it can only boost to a higher voltage.

Some examples. With a buckpuck, if you want to run 6 XR-E, you've got a 22V drop. So, you want a power supply that's around 24v.

With this driver, if you wanted to power that same string of 6 LEDs, you MUST use a power supply LESS than 22v. Really, anything within a few volts is fine. You might aim for an 18v supply since they're common. If the difference gets too great, efficiency drops (this is true with either style.)

The example I keep referring to in this thread - using a 24v power supply to drive 8 LEDs - is probably the best for most people, as it maxes efficiency, and maxes the number of LEDs the driver will handle. But, if, for some reason, you want to use this driver with fewer than 8 LEDs, that's OK, but you MUST carefully choose the power supply, else you will cook the LEDs!

So, just remember, the math is backwards when using this driver vs. buckpucks. When in doubt, please ask. Or just stick with the proven formula of a 24v power supply and 8 XR-E. It'll also work for 8 Luxeon Rebels, though they have a slightly lower Vf at typical currents. The XP-G has even lower Vf, and a 24v supply would be marginal for 8 of them. If you wanted to use this design for 8 XP-G, make sure you get a DC power supply with a trimpot, then just turn it down. The power supplies I'm using, from mpja.com, will turn down to ~20v, so they give you tons of adjustability.
 
Hi, i am newbie here. currently working on something similar to this. i am using NCP3066 from on semi too. the only problem i have is the PWM dimming.. I need some advice on how to do to provide controllable PWM pulses to adjust the brightness of the LEDs. I read from this thread using arduino. really never use it before.. i also read about the 555timers.. Not really confused what to choose.. thanks..
 
Hey kimimaro, I sent you a PM, but for the benefit of the wider audience, here are some comments on the options you mentioned:

1) Using an arduino: You can read about arduino here:

www.arduino.cc

There are lots of versions, but this is an extremely simple task for that platform, so it doesn't matter which one you get. The easiest to use is probably the standard reference design, which is currently called the Duemilanove. To use an Arduino for dimming, you basically connect one of the PWM pins on the driver to the PWM pin on the Arduino, then connect the Arduino's GND to any GND on the driver circuit (for instance, you could connect it to the GND pin on the board, or the "-" terminal on the output side of your power supply.) Then, upload code to the Arduino to dim as you want. There are examples of PWM dimming in the sample sketches that come with the development environment, or you can ask on these forums, as several other people are working on this. This sounds complicated but really it's very easy - if you can solder up one of these drivers, using an Arduino is going to be simple.

If you go with an Arduino, the only limit is your imagination. You can set the lights to fade on and off in the morning and evening, you can wire a pot into an analog input and use that to dim the lights, anything you want.

2) A 555 timer IC. If you go this route, you can use a pot to adjust the duty cycle fo the PWM signal generated by the 555 and dim the lights manually. This requires soldering up a circuit for the 555, but that's no more difficult than this driver project. I started a thread about using a 555 IC to make a PWM fan speed controller, but I guess it didn't interest anyone because there were zero replies. :D Here is that thread:

http://reefcentral.com/forums/showthread.php?t=1676242

You could easily use that same circuit to give a PWM signal to an LED driver. You'd want to eliminate the transistor and cap on the output side, though.

I know I keep promising this, but once I'm done testing the prototype of the driver, I will start a thread about arduino use for LED dimming. I've soldered the prototype up, and while it works, I'm going to make a few small tweaks (some of the hole sizes aren't QUITE right, for instance, and I'm going to turn up the isolation for the ground planes, since it's too easy for sloppy solderers like me to bridge to it). If anyone else out there is already making one of these, or thinking about making one, you might want to wait a day or two till I get the revised eagle files up - or, proceed at your own risk and double-check hole sizes and package outlines against the parts you'll be using.
 
Thanks for the quick reply.. Looks like i may go for 555timer.. other than these two ways, do you have some other advise? it's just for some comparison which is most suitable for the project i am working on.. thanks a lot for the info..:D
 
Back
Top