DIY LED driver for reef lighting

It's a "Non Recoverable Engineering" fee.

Each design requires a setup charge for their engineering costs to produce the films, stencils, pin board checker programming, etc. It's a standard charge and frankly, I'm surprised that it isn't more with them since it's a proprietary process.

The "cheap" board houses obviously waive that charge, but you don't get the same quality, the boards may not be electrically checked. Our last run of boards from Gold Phoenix had a few bad ones, but they always overshoot the quantity to compensate.
 
Boards came in this weekend but I'm still waiting on some parts from China to come in. I did have enough on-hand to build out one of the drivers on the board and things are working well so far. I did find one bug with the address pads for the temp chip where the ground pad didn't connect to ground. I'll have that fixed shortly on github but it was easy enough to fix by adding a jumper to the other ground pad..

I'll be able to finish testing all six drivers once the rest of my parts come in and I've made some good progress on the libraries.

All my parts finally came in. Seems things take longer coming from Hong Kong than mainland China. Go figure. I've got 4 boards built out with 650ma parts that I pulled from the DX drivers. I'm also building out 6 1.2A boards that I have all the parts for now except the ss24 diodes I just ordered. They're coming from China so I expect I'll have them in a couple of weeks.

I have half of a Ecoray 60 conversion done now. Had to drill the PCB to split the strings into 5 leds per string and then solder wires to the PCB. Easy enough.

I still need to calibrate the second board for 350ma output then I can wire up the other half and finish the assembly. I'll take some pics and post them up of the entire build process. I'm still waiting on some red led's I'm going to swap with some of the whites so it might be next week before I finish if they don't show tomorrow.

I've also completed the arduino libraries for I2C control of the boards (temp sensor and the pwm chip). Not all features of the pwm chip are implemented yet but I can address each channel individually and set the 8bit or 12bit PWM value.

I'll post up full details about everything once I get this one fixture completed.
 
Just in case anyone is looking to purchase any CATs I thought I would let you know that digikey is willing to match future electronics prices. (future is currently out of stock and won't ship there "reserve stock" out to US customers) I did however purchase 50 so it might be different if you purchase less, but its worth a try. All of the other companies that I tried weren't willing to match the price.
 
Hello All,

Just wondering if anyone is ordering any LM3409 boards soon. I was hoping to add some boards for me to your order.

Phillip
 
It dims the output if the temperature gets too hot.

I can't watch the video on this PC but honestly, I don't know if that is a big deal in our environment. Temperature and other environmental conditions should be stable/predictable over time. An unusually high temperature should probably be a "full stop" alarm condition - just shut things off. You can accomplish that via a thermal fuse in series with the DC power to the driver.
 
I also can't watch youtube here, but if it is what I think, it can be done with a thermistor on one of the voltage dividers. Temp goes up, divider brings voltage down - leds are dimmed (I love analog dimming :D) . Just need to select the right part.

As for a thru-hole lm3409 board, it should be straight forward. However I'll probably leave to someone else to spend the hours sorting through mouser and digikey to find suitable replacement parts since I prefer SMD components over thru-hole.
 
If I were to do a "through hole" design of the LM3409, I'd leave the passives and other components as SMT and just change the IC itself. They're all easy to solder compared to the LM3409 itself.
 
Good luck with a thru-hole version. Hope you need a lot of them since it's VERY likely you would have to purchase one thousand of them if you want one. Nobody will ever carry them so you would have to place a "factory order" and that would likely be a thousand or more pieces.
 
Good luck with a thru-hole version. Hope you need a lot of them since it's VERY likely you would have to purchase one thousand of them if you want one. Nobody will ever carry them so you would have to place a "factory order" and that would likely be a thousand or more pieces.

Actually, the MOQ is an extremely manageable 275. I haven't heard back from our account rep yet in regards to our corporate price, but it should be less than the $1.59 I was quoted.
 
That might be easier to solder, but I haven't tried soldering my boards yet so I have to admit SMT might not be as hard as I think it is. Might do it this weekend.

Lots of good advice in this thread for how to do it, I think I need a solder braid, I have most of the rest of the stuff. (fine tip temp controlled iron, not air though).

== John ==​
 
That might be easier to solder, but I haven't tried soldering my boards yet so I have to admit SMT might not be as hard as I think it is. Might do it this weekend.

Lots of good advice in this thread for how to do it, I think I need a solder braid, I have most of the rest of the stuff. (fine tip temp controlled iron, not air though).

== John ==​

You'll definitely want some solder braid. I finally finished mine up tonight. The braid really helped. I asked how other solder the LM3409 earlier today.

http://www.reefcentral.com/forums/showthread.php?t=2156300
 
Well it depends. You could, actually, use the non-HV version with 12 XPGs (12*3.3=39.6Vfw), however you'll have to find a 42V PSU. And from my short research it doesn't look like an easy task. 48V is much more common. Thats why I'm using the HV version.


In other news, here is a time-lapse of me populating a board. I found that for this chip it is easiest (for me) to just put some solder on the 10 pads, position the chip, and just touch the pins of the chip with the iron until I see the solder melting and covering the pin. Also use LOTS of flux!

<object width="651" height="366"><param name="allowfullscreen" value="true" /><param name="allowscriptaccess" value="always" /><param name="movie" value="http://vimeo.com/moogaloop.swf?clip_id=27958445&server=vimeo.com&show_title=1&show_byline=0&show_portrait=0&color=00adef&fullscreen=1&autoplay=0&loop=0" /><embed src="http://vimeo.com/moogaloop.swf?clip_id=27958445&server=vimeo.com&show_title=1&show_byline=0&show_portrait=0&color=00adef&fullscreen=1&autoplay=0&loop=0" type="application/x-shockwave-flash" allowfullscreen="true" allowscriptaccess="always" width="651" height="366"></embed></object>
Is this video still available? I only see a padlock and "private video"...

Got the pcbs today and am looking for some hints to get this tiny LM3409 soldered.
 
hi...
I have always problem with this driver.
I have built three board for 48v 700mA with:

Vin = 50.5V
Vo = 48v
Roff = 16000 ohms
L1 = 18uH
Rsns = 0.25 ohms
Frequency = 98Khz

___________________________________

Vin = 50.5V
Vo = 48v
Roff = 7870 ohms
L1 = 47uH
Rsns = 0.33 ohms
Frequency = 199Khz

____________________________________

Vin = 50.5V
Vo = 48v
Roff = 2200 ohms
L1 = 12uH
Rsns = 0.33 ohms
Frequency = 713Khz


all three boards whistle, some more and some less
if dim to 0% or 100% it does not whistle!
if dimming from 1% to 99% it whistles!
what is the correct frequency not to do it whistle?
 
@Baghyyy: look around #1712 in this thread. I had the same problem with the CAT and decreased PWM frequency the way it is described in the link above.

@terrahz: thanks for fixing the video.
 
Terahz,

JC286006 hooked me up with some of the latest version boards. They are the V0.7. with the heatsink. Great board BTW.

When the whites come out, I plan on running mostly the newer, lower Vf XTE LEDs. I should be able to fit about 4 per string with a 12V PS or 5 with a 15V PS.

Do you think you could make a BOM for a 12V and 15V driver? I have these two sizes at home already.

Thanks
Phillip
 
Back
Top