DIY LED driver for reef lighting

[Spaced Cowboy]

Is there a better method for those who don't have enough experience to solder the pins individually?

The way I usually do small parts (when I'm not using the SMT oven [grin]) is to put lots of flux all around the part with one of the flux-pens, then drop the part onto the pads and maneuver it into place with tweezers. I add a small amount of solder to the soldering iron tip so I get a small ball of solder on the end of the tip; then just touch the small ball of solder to the end of the pad (not the pin coming from the chip, the pad it rests on, as far away from the pin as possible). Keep it there for a second or so, and the solder will wick its way onto the pad, joining the pin and the pad together.

I can do really fine-pitch parts like this (0.5mm), although it does help to wear these, and of course to not let anyone actually see you wearing them. That could be a symptom of my advancing years, but I really do find them invaluable for fine-pitch parts.

Simon.

 

[Gorgok]

All i do is put a small dab of solder onto one pad of the component, then with tweezers bring the part in and re-melt that pad and get it to join the component. Using 0.015 wire i can easily solder the other pads individually, but the smallest i have worked with is 0608 components. For larger components, like a CAT4101 i don't bother with the dab of solder, instead i use a small spring clamp to place the component and just solder all the legs individually. Plenty of flux is always applied to make sure everything happens quick and smooth.

There is actually a few videos by EEVblog that demonstrate this method very well in my opinion. His videos also have good ways to not individually solder pins of small parts, rather drag the solder across and such. Very good advice i think.

 

[firechild]

Thanks for the advice, I'll try a different method with the next board. I have been using a flux pen to get plenty of flux around the pads but I think it's going to take a lot more practice. I guess 0603 components and 0.5mm pitch ICs isn't really the ideal way to learn how to solder surface mount components.

 

[O2Surplus]

Has anyone tried using the hot air method? I've managed to learn how to do it, and now find it's actually much easier to use than conventional soldering methods. I preheat my boards to around 200 degrees F on a hot plate with the solder paste and components already placed. I then rapidly heat the top of the board with a hot air gun set to about 600 degrees. The neat part is that all the components will "self align" to their respective solder pads when the solder melts. The only real hard part of using this soldering method is learning to judge the proper amount of solder paste to be used. The paste contains flux, which burns off during the soldering process, so placing just the right amount of paste- takes a bit of practice. Fine pitched parts that end up with solder bridges between pins can be reworked with some additional flux and a solder wick. I now only drag out the conventional soldering iron for the few through hole parts that I still use.

 

[der_wille_zur_macht]

Are you using just a plain old heat gun, or an actual hot air rework station meant for soldering?

 

[O2Surplus]

Are you using just a plain old heat gun, or an actual hot air rework station meant for soldering?

I bought a new heat gun with adjustable temp from 0 - 1000 degrees F and adjustable air flow. I have to adjust the flow down to just a whisper otherwise the components will blow off the board ,lol. I've soldered enough, using this method, that I can produce an SMD version of the typhon in about 10 minutes. I'll eventually by a hot air rework station, if the quality and price are too good to pass up.

 

[TheFishMan65]

DWZM, was the 604k too dim for moon lights. Still thinking about a second CAT for moon lights, but I have not figured out the resistance value yet. Did you?

 

[der_wille_zur_macht]

Even at 100% it was reeeeeeeallly dim. I didn't measure current but it was only dropping ~15v across 6 XP-E RBs. A string of 6 LEDs at whatever current that was would probably be much too dim, at least for my tank.

I have a CAT4101 as the moonlight driver on the Typhon plugin I'm working on for driving moonlights. I haven't built it yet, as the PCBs are still enroute from China. But, I am theoretically running it with a 1k sense resistor in series with a 50k digital I2C pot for dimming (without tying up a PWM channel). I am planning on only two or three LEDs for my tank.

 

[TheFishMan65]

I will be interested to hear how it turns out.

 

[der_wille_zur_macht]

Yeah, I figure it's easy enough to swap the static sense resistor so it's really down to selecting a pot with enough range to cover the variation I wanted, so I made a spreadsheet that shows mA per step of the digital pot and picked the 50k value based on how it looked. Higher values don't seem to add any real resolution as the current just tapers off faster. So given that this pot is correct to give good resolution, I can just swap the static sense resistor to move the range brighter or less bright.

 

[shark boy]

Cat4104 PCBs

Cat4104 PCBs

If anyone is thinking of ordering some pcbs for the triple cat 4104 drivers I would be willing to pay up front for 3 or 4. I need the boards, I'm just not good at ordering them.

Shark boy

 

[der_wille_zur_macht]

Some anecdotal evidence. When dimmed with an Arduino, the LM3409 aliases under 50kHz or so at certain duty cycles. If anyone is using this combo, aim higher. I am targeting 100kHz for my production version.

 

[terahz]

How did you get to 50kHz? I think Arduino's default frequency is about 500Hz. And the switching frequency for the IC should be about 400kHz.

 

[O2Surplus]

I thought there was a range of switching frequencies that the LM3409 could operate at? I aimed for 500khz with my component selection, since that was between the recommended frequencies of 100khz - 1000khz. What PWM frequency would be best to use from the arduino? I've had good luck with the Cat4101's low end dimming performance using PWM frequencies down to 150hz. I hope the LM3409 will react favorably to similar PWM signals.

 

[der_wille_zur_macht]

50kHz switching frequency on the LM3409. It's frequency is indeed dependent on component choice. The datasheet suggests operating frequency at least an order of magnitude or more above your pwm frequency to avoid aliasing. Looks like it is more like two orders of magnitude. I am getting aliasing at duty cycles between about 90 and 97 percent.

It's worth pointing out that you kinda have to try really hard to get the frequency of the LM this low. I don't think this will be an issue for most people but I figured I would point it out anyways. Even if you found yourself in this situation you can effectively adjust frequency without changing components by playing with the difference in Vin and Vout.

 

[Krakonosh]

Feature request

Feature request

Hi guys,
I just wanted to say thanks for all the hard work you guys put into this.
I'm planing to use TeraHz's v0.4 design in my 12 driver design with Arduino Mega.
I would love to also implement a current monitoring on each of the drivers.
What do you guys think of the hi side current monitors like this one.
Any chance we could include it on the board (I know we have some room left before we hit the 5 x 5 cm limit) ?
I could more less figure out the circuit, but I'm hopeless in Eagle.
Thanks.

 

[der_wille_zur_macht]

Interesting concept, Denis. Can I ask what your desire is for monitoring current? The whole idea of the constant current driver is that current doesn't change despite changes in the environment, so in theory it would always show the same number.

 

[Krakonosh]

Well, it's just an idea.
My overall concept is like this:
I'd like to have manual pots to fine tune the color and use the Arduino PWM output (converted to analog) for main sunrise - sunset dimming (full 255 steps).
The current monitoring would help me to tune the string currents between the 3 modules I'll have and also allow me to set a bench mark between adjustments.
Also later on I could use it to alarm any open circuit strings.
I did give a PLEDs a thought too, but it seems like a lot of work to implement.
I'm planning to put a driver on every string, no parallel strings here.
It's all in my head right now, open to any suggestions.
Thanks,

 

[terahz]

Sorry to anyone waiting for me to say the final word on the analog dimming for the LM3409 board...

I finally sat down to test the v0.4 board today and unfortunately it still isn't ready. The board behaves as expected until I plug in the LED psu. Here are the two situations:

Situation 1:
5V supplied for reference
LED psu (15V in my case) is off
0-5V dimming range (0.1V cutoff voltage)
Things work as expected.

Situation 2:
5V supplied for reference
LED psu (I'm testing with a 15V switching PSU) is [strong]on[/strong]
0-5V dimming range([strong]1.2V cutoff voltage, one way[strong])
After the LEDs are on (Adj voltage goes over 1.2V), they never shut down even if adj voltage goes to 0.

So, obviously, the second PSU messes up the comparator, but I haven't got a clue why and how.

Time to start googling. If anyone has ideas, don't be shy .

PWM is still working fine.

 

[terahz]

Sorry to anyone waiting for me to say the final word on the analog dimming for the LM3409 board...

I finally sat down to test the v0.4 board today and unfortunately it still isn't ready. The board behaves as expected until I plug in the LED psu. Here are the two situations:

Situation 1:
5V supplied for reference
LED psu (15V in my case) is off
0-5V dimming range (0.1V cutoff voltage)
Things work as expected.

Situation 2:
5V supplied for reference
LED psu (I'm testing with a 15V switching PSU) is [strong]on[/strong]
0-5V dimming range([strong]1.2V cutoff voltage, one way[strong])
After the LEDs are on (Adj voltage goes over 1.2V), they never shut down even if adj voltage goes to 0.

So, obviously, the second PSU messes up the comparator, but I haven't got a clue why and how.

Time to start googling. If anyone has ideas, don't be shy .

PWM is still working fine.