Aquastyleonline.com LED questions.

[00Warpig00]

Dread... you got my mind turning... I've been crunching some numbers for the last hour or so. I have an idea... not sure if it will work yet (I'm leaning toward yes it will at this point). I need to crunch more numbers but it's time for bed. I need to look at your code in your sketch to try to deduce some things too. I haven't done programming in a LONG time am just a programming hack and don't know how robust the arduino sketch language is... but if we can do basic math in a sketch (I assume we can) and compare some variables and make some decisions based on those comparisons and possibly add a time wasting loop/delay this may be workable. I cant think right now I'm fried, but I will look more into it tomorrow.

Nick

 

[00Warpig00]

Dread... just looked at your sketch. I'm not 100% familiar with the syntax of the sketch language yet but from what I do understand about it from looking at it for about 15 minutes I think what I'm thinking is workable in the software dept. A little more thinking tomorrow but I need sleep. I will come up with something as soon as I can to try to communicate my idea so everyone can pick at it... I think i will have to come up with a rough outline that resembles a program/sketch with my limited programming experience and see if we can hammer things out. Are there multi channel I2C digital pot chips that contain either 2 or 4 channels per chip and are 1K ohm? I didnt search hi and low cuz it's late but from your writings above you may have found some 1K I2C's

Nick

 

[want2reef]

Just placed my order... Will keep you guys updated on how it goes.

Oh ya, 24 Royal blues 6 10000K 6 6500K.
If I end up liking them, I will be ordering more.

 

[jwoyshnar]

Just placed my order... Will keep you guys updated on how it goes.

Oh ya, 24 Royal blues 6 10000K 6 6500K.
If I end up liking them, I will be ordering more.

you will like them. I said the same thing. I am ordering more soon.

 

[dread240]

I need to call mouser sometime next week (or tomorrow if they're open). They have lots of chips in the ranges I'd really like for this project, but when pulling up datasheets it all says they're different resistances and like 10k ohm end to end on the coil.

The other idea I'm toying with right now too is resistor arrays, but I've never played with one of those before andthe main company I saw 'bourns' website was down

 

[want2reef]

Anybody running 20 LED's per driver on the maxwellens?

Just wondering if they can handle it.

 

[dread240]

I am, I run 20 royal blues per driver

 

[want2reef]

Another dumb question.

Does the blue LED's put less par than the whites do, kinda like 20K MH's put out less par than 10K MH's?

 

[dread240]

I never tested individual par between the blues/whites, just my overall fixture, I can try that tomorrow for you though

 

[dread240]

so... another update.

I did what I should have actually done beforehand, and thoroughly tested the dimming circuit and outputs of the maxwellen drivers before I got all willy nilly and started throwing digital pots all over the place. Good thing about it though... is that this may turn out even simpler then I thought by just using a multi-tap resistor array.

Resistance - current
22 ohms - 36ma
21 ohms - 37ma
20 ohms - 38ma
19 ohms - 40ma
18 ohms - 42ma
17 ohms - 44ma
16 ohms - 46ma
15 ohms - 49ma
14 ohms - 51ma
13 ohms - 54ma
12 ohms - 58ma
11 ohms - 62ma
10 ohms - 67ma
9 ohms - 74ma
8 ohms - 81ma
7 ohms - 91ma
6 ohms - 104ma
5 ohms - 121ma
4 ohms - 140ma
3 ohms - 174ma
2 ohms - 230ma
1 ohm - 346ma
.9 ohm - 376ma
.8 ohm - 394ma
.7 ohm - 416ma
.6 ohm - 463ma
.5 ohm - 490ma
.4 ohm - 535ma
.3 ohm - 570ma
.2 ohm - 620ma

(this also goes to show why people were saying the pots aren't very smooth... it's because the output isn't linear and the pot is linear)

I'm going to use a 16 channel multiplex switch from mouser. Actually gives us 17 positions with the dimmer circuit open totally (which I didn't measure but it's probably like 10ma)

channel 1 - full power
channel 2 - .2 ohm 620ma
channel 3 - .3 ohm 570ma
channel 4 - .4 ohm 535ma
channel 5 - .5 ohm 490ma
channel 6 - .6 ohm 463ma
channel 7 - .7 ohm 416ma
channel 8 - .8 ohm 394ma
channel 9 - .9 ohm 376ma
channel 10 - 1 ohm 346 ma
channel 11 - 1.3 ohm 300+/-
channel 12 - 2 ohm 230 ma
channel 13 - 2.5 ohm 200+/-
channel 14 - 3.7 ohm 150 +/-
channel 15 - 6 ohm 104ma
channel 16 - 14 ohm 51ma
Switch open - infinite resistance 10ma

They're not the smoothest, but I looked at the intensity by eye between those changes, and it's barely noticeable. I also stacked more channels on the top end as well too for finer control of the driver.

As for how it works, you link all your resistors in series. do (2) .1 ohm resistors to get to .2 ohm's. You just build a line of them to get to the values you want for each step. In between each resistor you tie one channel of the switch. Whatever switch is closed to the common (which will be one of the dimmer wires, the other dimmer wire goes to the front of the resistor array) determines how far down the resistor array you travel. If you close channel 5 for instance you will have passed through 5 .1 ohm resistors in series, for the .5 ohm resistance needed for that output.

Near the end of the array you'll get a bit more tricky with resistor selection, but it won't be bad. I may modify the values just for ease of assembly at the end of it. Should actually make programming it a bit easier too as it's just bit bang for the switch

 

[dread240]

http://microcontrollershop.com/product_info.php?products_id=4579

That's a link to the breakout board we'll be using for the resistor array for those that are curious

They also make this in dip style to where a breakout isn't needed, but pricing wise this looks better on a breakout board (and can probably build the array right on it as well too)

http://bildr.org/2011/02/cd74hc4067-arduino/

And a bit of info on how it works and how it interfaces to an arduino

 

[00Warpig00]

NICE! this looks very promising. Your info above about the actual current vs resistances makes the idea I was thinking through a non starter as it looks like much of the magic happens in the less than 10 ohm resistance realm which my idea cannot produce. i took a quick look at your links. looks like a great idea. Also looks like if someone wanted to build a longer string of resistors and use two of the breakout boards at a cost of 4 more digital pins on the arduino they could have 32 steps of resistance and an even more fine control. I also love the idea that you can leave the dimmer wires open (the ~10ma) this would allow me myself to use an entire string of my blues in 10ma mode as a moonlight.

Alright alright... I'm almost convinced to order an arduino mega.

Thanks for all your work on this Dread.

Nick

 

[00Warpig00]

I have to admit, I am a bit concerned about more than 100+ ma flowing through the matrix chip when in the brighter positions. Sounds like a fried IC to me. I guess maybe the addition of 16 relay's or transistors to the project and just have the base of the transistor driven by each matrix pin switch each resistor on/off. but then you need to account for any emitter-collector resistance.

Nick

 

[dread240]

It's something that I'll be testing before I give anybody the go ahead. We use these types of chips in a lot of our plc systems which run through 12v or 24v dry contacts at 480ma. The spec sheet on resistance and current is for the control circuitry

Also remember that only the .7v dimming circuit is running through it.. not the driver outputs.. I will verify dimmer current tonight though before i order 2 chipsets. The ma values listed previously were the driver output compared to resistance in the dimmer circuit

 

[Neurovet]

I'm enjoying spying on this thread..especially with Dread's controller contributions. I have a 80 DIY setup on my 90 gal bowfront....46RB, 16 4500, 8 6500, 10 10K. I know the white termp mix is unconventional, but I like it. I clustered the lights as described by evil on nr.com. They are on an apex lite controller. I am thrilled with the overall color temperature and growth of SPS. I got a number of frags from Mr. Coral and all are polyped out 24/7 and encrusting their plugs. My friend has the same frag pack under 2 x 250MH and is not getting as much growth. Only other difference is in the birds nest growth. His grows like a more transitional bird nest, while mine is more ball like with thicker branches. I am able to keep croceas on the sandbed. Like everyone else, I have decided to add another 20 for fuller coverage. I'll get some coral growth pics up soon but here is my setup.

 

[Neurovet]

One other thing....purple corals appear white under this color temp...Anyone have experience with alternate colors that will make the pink shine through the actinics?

 

[dread240]

I'm waiting to hear back from sparkfun. I firstly wanted to verify they did include the mux chip for that price, which it is, so no need to solder soic components (cheaper this way too)

I've also messaged them about the possibility of 200ma through the IC contacts. I sent their technical department a quick sketch to see if they recommend isolating it or just going directly through the chip. I honestly think the chip can take it, and I haven't had anything stick out in the datasheets to me as a 'max current through contact's' type deal. Either way, I'll see what they say first.

If it's all good, parts will be on order tomorrow, and hopefully assembled by this weekend, working for real this times

 

[00Warpig00]

Also remember that only the .7v dimming circuit is running through it.. not the driver outputs.. I will verify dimmer current tonight though before i order 2 chipsets. The ma values listed previously were the driver output compared to resistance in the dimmer circuit

I've also messaged them about the possibility of 200ma through the IC contacts. I sent their technical department a quick sketch to see if they recommend isolating it or just going directly through the chip. I honestly think the chip can take it, and I haven't had anything stick out in the datasheets to me as a 'max current through contact's' type deal. Either way, I'll see what they say first.

If it's all good, parts will be on order tomorrow, and hopefully assembled by this weekend, working for real this times

Awesome! Brain fart on my part. I thought you were measuring through the dimmer circuit. You are probably correct the dimmer circuit I would imagine isn't flowing 100-650ma or so. I cant imagine the full output of the driver being put through the dimmer. These drivers cant be that cheap. lol

Nick

 

[dread240]

lol, no not at all. I did see 190ma at full power through the dimmer circuit, but again that's at .7v. The traces I believe really only care about current though, not voltage, so it will still be the limiting factor. I'm thinking we're still ok though as we use multiplex systems all the time in our controls, and all of our outputs are 400-500ma. Granted I've never taken one apart completely to see if it's isolated or not, but alot of the transistors I was looking at require anywhere from a 50ma-500ma signal anyways. This is newer stuff for me so I won't be surprised if I'm wrong and we do need to isolate though.

 

[symon_say]

Everything you say is over my head, lol, but great job.