TheFishMan65
New member
Look for two connector that stack together. That is what I used in my design.
DIY LED driver for reef lighting
- Thread starter der_wille_zur_macht
- Start date
http://www.mouser.com/Search/ProductDetail.aspx?R=1984659virtualkey65100000virtualkey651-1984659
http://www.mouser.com/Search/ProductDetail.aspx?R=1984785virtualkey65100000virtualkey651-1984785
Just found them on page 30, my bad but this is very long threat!!!!!! I have been jumping all over it several times and keep getting new details....
http://www.mouser.com/Search/ProductDetail.aspx?R=1984785virtualkey65100000virtualkey651-1984785
Just found them on page 30, my bad but this is very long threat!!!!!! I have been jumping all over it several times and keep getting new details....
TheFishMan65
New member
You realize that one of those connector has the wires coming out the top and the other out the side? Maybe that is what the design calls for or maybe what you want. I will also point out these 2 pin connectors:
Terminal Blocks PTSA 1.5/ 2-3.5-F
They are very close to the same spec, but are only $0.26 each (even less if you buy 10). From the specification you cn mount 2 next to each other for a 4 pin and 3 next to each other for a 6 pin version.
Terminal Blocks PTSA 1.5/ 2-3.5-F
They are very close to the same spec, but are only $0.26 each (even less if you buy 10). From the specification you cn mount 2 next to each other for a 4 pin and 3 next to each other for a 6 pin version.
der_wille_zur_macht
Team RC
Difference in those caps is indeed only the reel size.
I have tried to not spec a single header because I felt that people would want to pick something that worked with their application. Top exit, side exit, pin headers, screw terminals, and so on - as long as it has 3.5mm spacing and isn't huge it should work. I designed the board to take standard-issue screw terminals which are about the biggest thing you might end up with, so other formats should fit fine.
I have tried to not spec a single header because I felt that people would want to pick something that worked with their application. Top exit, side exit, pin headers, screw terminals, and so on - as long as it has 3.5mm spacing and isn't huge it should work. I designed the board to take standard-issue screw terminals which are about the biggest thing you might end up with, so other formats should fit fine.
mprygh
New member
Fantastic thread, but I was really confused about half way through with the changeover to the CAT4101.
The big problem I had was the concept of tweaking the power supply
to make the voltage differential tiny, and thus getting the efficiency
back that you lose using by a simple linear regulator.
That is great for 1 power supply per string, but I don't get how
you can do that with several strings on one supply.
With the LED Vf variation, the strings will simply not match.
An XP-G can vary from something like 2.8-3.7 at a low 350mA.
It only gets worse at 1A.
That is upwards of 30%. Sure, real life will be less, and they average out.
But even assuming half that, at 15%, that is 3.5V.
Add in the 0.5V dropout, you are losing 17% through that regulator.
Too much.
Am I missing something??
On the plus side, you can actually run any voltage you want, as long
as you keep the voltage on the device at < 25V. Easy enough
with a zener, resistor, and some planning. So you could run
a 200V supply, and put everything on one string. :hmm4:
The big problem I had was the concept of tweaking the power supply
to make the voltage differential tiny, and thus getting the efficiency
back that you lose using by a simple linear regulator.
That is great for 1 power supply per string, but I don't get how
you can do that with several strings on one supply.
With the LED Vf variation, the strings will simply not match.
An XP-G can vary from something like 2.8-3.7 at a low 350mA.
It only gets worse at 1A.
That is upwards of 30%. Sure, real life will be less, and they average out.
But even assuming half that, at 15%, that is 3.5V.
Add in the 0.5V dropout, you are losing 17% through that regulator.
Too much.
Am I missing something??
On the plus side, you can actually run any voltage you want, as long
as you keep the voltage on the device at < 25V. Easy enough
with a zener, resistor, and some planning. So you could run
a 200V supply, and put everything on one string. :hmm4:
der_wille_zur_macht
Team RC
You're not missing anything, except that in practice I don't think it would be as bad as you're predicting. The Vf variations from LED to LED that I've seen in my own testing have been significant, but not nearly THAT significant - more like a tenth of a volt, max. And as you add LEDs into longer strings, the potential range for delta Vf from string to string grows, but the distribution also gets much skinnier - that is, your chances of ending up with "like" strings (i.e. small delta Vf) get much better as the strings get longer. With 6 LEDs per string, I think it's safe to assume you'll be within reasonable ranges.
And, while it's nice to tune the supply to get the desirable minimum drop of .5v, it's by no means a strict requirement. Depending on cooling on the CAT4101 chip, they can take a much higher drop and still operate happily.
All that said, IMHO it's still probably a good plan to group strings containing different LEDs on different power supplies; i.e. if you're mixing royal blue XR-E with cool white XP-G, you probably want a "blue" supply and a "white" supply, since the strings of XR-E and XP-G will likely have different Vfs.
And, while it's nice to tune the supply to get the desirable minimum drop of .5v, it's by no means a strict requirement. Depending on cooling on the CAT4101 chip, they can take a much higher drop and still operate happily.
All that said, IMHO it's still probably a good plan to group strings containing different LEDs on different power supplies; i.e. if you're mixing royal blue XR-E with cool white XP-G, you probably want a "blue" supply and a "white" supply, since the strings of XR-E and XP-G will likely have different Vfs.
theatrus
100-mile-commuter
The efficiency reason and allowing multiple types of LEDs per supply was part of my motivation for the LM3404HV buck converter version.
And an updated on the "FlexDriver": Boards are on order, parts are sitting on my desk. I'm going to give it a spin on my lighting rig for my new desktop ADA30C nano cube, running off of a Lambda LS100-48 (48V) power supply.
And an updated on the "FlexDriver": Boards are on order, parts are sitting on my desk. I'm going to give it a spin on my lighting rig for my new desktop ADA30C nano cube, running off of a Lambda LS100-48 (48V) power supply.
mprygh
New member
Interesting data on how the Vf matches.
I guess that makes sense, given the extensive part binning.
Good info.
My original leaning before reading this thread was one large 36V power
supply and possibly a LM3406 buck type.
I actually consider 48V to be a bit unsafe, so did not want to go there.
And I wanted to run my XP-G at 1200 mA, series with XP-E B + RB in parallel,
so 600 mA each, hence the larger current.
With the main power pins next to each other, I was just going to dead-bug it.
(thermal epoxy it upside down to a heat sink, and wire to pins)
Ground planes are over-rated.
I have some more thinking to do though.
I guess that makes sense, given the extensive part binning.
Good info.
My original leaning before reading this thread was one large 36V power
supply and possibly a LM3406 buck type.
I actually consider 48V to be a bit unsafe, so did not want to go there.
And I wanted to run my XP-G at 1200 mA, series with XP-E B + RB in parallel,
so 600 mA each, hence the larger current.
With the main power pins next to each other, I was just going to dead-bug it.
(thermal epoxy it upside down to a heat sink, and wire to pins)
Ground planes are over-rated.
I have some more thinking to do though.
First off you guys rock
I think I'm going to build the DWZM designed CAT4101 3x driver for my little project.
I am going to start with an 18-21 led setup for my fuge. Then go bigger into led's with my main reef.
can any one spare a board or two? (please pm me)
It was probably covered but, if I don't want to use the PWM feature on my first build do I just provide 5v to the en/pwm leg?
Thanks, Steve
I think I'm going to build the DWZM designed CAT4101 3x driver for my little project.
I am going to start with an 18-21 led setup for my fuge. Then go bigger into led's with my main reef.
can any one spare a board or two? (please pm me)
It was probably covered but, if I don't want to use the PWM feature on my first build do I just provide 5v to the en/pwm leg?
Thanks, Steve
redhanded77
New member
Im thinking of building a led fixture for a 75g tank, and i was wondering if 96 leds (48 xpg white and 48 xre royal blue) setup at every 3 inches front to back and side to side would be suffecient enough to keep sps. Im not sure on what height it will be hung at, I guess it will depend on coverage and PAR numbers. What would you guys recommend?
imagex
New member
I've also lit up a 75 with 48 leds and got another friend that did 48 on his 120 and pretty good coverage and spread... 2 lines of 24 but he has his fixture up 16" from water level..
also another thing this thread isnt for led questions.. it's for diy driver questions.. should probably start a new thread on other subjects..
redhanded77
New member
thanks and sorry bout that. I meant to post it on the other DIY led thread.
TheFishMan65
New member
So Red do you often get got RedHanded posting in the wrong thread
Image,
You can use one or two supplies. For 72 LEDs that is 12 string so you would need at least 2 of the supplies you point. If you run them at the full 1 amp you might need three. IIRC kkcress recommended not running these above 80% or 4-5 amps.
The reason for 2 is if the white and blue have different voltage drops. Let' say you have a 24 (non adjustable supply) and are running at 1 amp. Further suppose the blues have a drop of 22 volts and the white 20. This means that that the CATs for the blue need to dissipage 2 volts * 1 amp = 2 watts. The white need to dissipate 4. This can be a lot of heat and wasted energy. Now if the supply is adjustable and IIRC the recommended overhead is 1/2 volt so set the supply at 22.5 then you only have .5 volts * 1 amp = .5 watts, but still have 2.5 watts for white. With two supplies you can set the second for 20.5 and also only have .5 watts for white.
Hope this make sense.
Image,
You can use one or two supplies. For 72 LEDs that is 12 string so you would need at least 2 of the supplies you point. If you run them at the full 1 amp you might need three. IIRC kkcress recommended not running these above 80% or 4-5 amps.
The reason for 2 is if the white and blue have different voltage drops. Let' say you have a 24 (non adjustable supply) and are running at 1 amp. Further suppose the blues have a drop of 22 volts and the white 20. This means that that the CATs for the blue need to dissipage 2 volts * 1 amp = 2 watts. The white need to dissipate 4. This can be a lot of heat and wasted energy. Now if the supply is adjustable and IIRC the recommended overhead is 1/2 volt so set the supply at 22.5 then you only have .5 volts * 1 amp = .5 watts, but still have 2.5 watts for white. With two supplies you can set the second for 20.5 and also only have .5 watts for white.
Hope this make sense.
der_wille_zur_macht
Team RC
Yep, if you just want it to turn full-on supply 5v to the PWM pin.
So has anyone done anything more with the PT4115E chip? I found some drivers based on this chip on DealExtreme for around $2 each. They should be very easy to mod to drive whatever the chip is capable of driving. I tried to post about this a couple days ago too but I think the post is lost in moderation limbo.
mprygh
New member
An interesting chip.
It appears it can handle a completely unregulated supply.
So you could get a simple 120V-12V AC transformer (10:1), a couple
of rectifier diodes, and you are done. Big savings on power supply.
With a 100uF cap, the inrush should not be that bad, and you could
add a resistor to help.
Power factor correction might be an issue though.
