New tank setup with LEDs and microprocessor

[der_wille_zur_macht]

Check out this thread:

http://www.reefcentral.com/forums/showthread.php?t=1783424

for a successful pH amplifier implementation by user terahz.

 

[cyrusthevirus]

Yeah this is the circuit I started my design with. I made a couple of mods to it.

Stu-
I am using the LF356 sorry for the typo.

Interesting circuit. I am using a PIC24F chip by Microchip. It provides 9 channels of hardware PWM. In addition I decided that instead or using a high current PS, I would go with a 1A PS. That way if something goes wrong, the circuit is still in the high, but acceptable range for the LEDs.

 

[der_wille_zur_macht]

Stu will shoot me down, but another option for the LEDs is to just build a driver around a purpose-designed IC. The CAT4101 chip is what most of us in this thread are using:

http://www.reefcentral.com/forums/showthread.php?t=1759758

It's cheap and easy to use, only 3 external components per circuit. You get a PWM-dimmable constant current driver with reasonable efficiency and built-in protection.

I wouldn't rely on your power supply's max current rating to limit current in the event of a disaster. IME, many power supplies will happily supply more than their max rating (before some internal thermal protection shuts them off or they just break) for plenty long enough to cook LEDs.

 

[stugray]

"Stu will shoot me down"

No I wont shoot you down, I am even trying to layout my own CAT4101 boards in eagle this week as the family is on spring break ;-)

However, another simple solution is to set the max current per string using kcress's "simple" method, then achieve the dimming using the MOSFET & PWM signal.

It sounds like cyrusthevirus's design will work fine.

Stu

 

[cyrusthevirus]

All I was looking for when I started this project was a chip that provided PMP, at least 4 PWM channels and 2-4 ADC capable pins. My first chip was a 28 pin PIC from Microchip, but the sharing of pins became a pain to program. So I upgraded to a 100 pin chip. I was worried about how I was going to prototype with such a high pin chip until I found the schmartboard. It made prototyping much easier.

As far as the LED circuit, I originally had a fuse but learned that by the time the fuse blew, the LEDs would already be blown. I also had a LM317 to limit the current, but determined that would not be much better.

I guess when it all comes down to it, the circuit is built in such a way that I am comfortble I did all I could to not blow all of the LEDs either the instant the sun rises for the first time or at noon.

 

[cyrusthevirus]

Yeah my LEDs finally arrived from DX. Now it is time to start building.

 

[tahiriqbal]

Tagging along..... very nice build, can't wait to see the rest.

Good luck with your entire project.

 

[JohnsonSBK]

ok for someone who doesnt understand electrical lingo how are you making the leds dim automatically? i know how to solder would it be easy to do if you gave a schematic? is there anyone who can explain step by step how to build this controller and make it work?

 

[cyrusthevirus]

The first post has a schematic of the circuit although I am still tweaking it. The heart of the circuit is a microprocessor, basically a Pentium III chip. The chip is programmed to produce a PWM signal which increases or decrease the intensity of the light. The PWM signal works not by decreasing the amount of voltage to the led but by turning it on and off very fast. So if it is on 50% of the time, the birghtness is 50%. If it is on 10% of the time, then the brightness is 10%. My code is far from perfect and not for for general consumption, but PM me and I can zip it up and send it over to anyone who want to look at it.

If this does not make sense, what parts are confusing?

 

[cyrusthevirus]

Although I have the ability to control the temp with the microprocessor, I decided to use a Ranco 111000. I wired it using the instructions below.

http://www.rcreptiles.com/blog/index.php/2007/05/11/how_to_wire_a_ranco_etc_111000_thermosta

I made the temp probe waterproff using Plasti Dip as you can see in the pics below. I did not thin it out and only dipped once.

 

[cyrusthevirus]

On to the LEDs. My for the LEDs is to put them into the old housings I have for standard aquarium lights. I removed all of the guts and plan to install two fans in each light. Then cut out a vent hole and attach the leds to "C" channel aluminum and then to the polished surface. This should provide a light weight standard looking housing.

Here is the concept:

Here are the old guts, TRASH!!:

The empty housing:

Fan placement:

 

[cyrusthevirus]

In doing research for setting up the LED banks I came across melev's setup in the reef discussions. Here is his lighting plan. Now time to create enough banks to be able to do the same, but with LEDs.

 

[der_wille_zur_macht]

Cyrus,

I had pretty much the same setup on my last large MH/VHO tank (a 125g). It made me feel good, but the more I thought about it, the more I realized it was woefully unrealistic. I started experimenting with LEDs trying to approximate that effect melev and I had with MH/VHO when I realized how unrealistic it is. I had LEDs in different banks from left to right over a tank and faded them on and off separately, trying to convince myself that I was approximating the sun moving over the reef.

But, when the sun travels over a real reef, you'd pretty much NEVER see that pattern of colors and intensities! Instead, you'd see an overall color and intensity shift that would be the same everywhere (as opposed to one end of the tank being intense and white and the other end being dim and blue), and you'd see a change in the angle of the light through the day.

For us with LEDs, the real intensity and color shifts are easy. Just program the whites and blues to fade on a bell curve or some approximation. The only part that's still eluded me is approximating the changing angle of the light through the day. One way to do this would be to have the LEDs in narrow banks on axles so you could turn them to adjust the angle. But I can't think of a way to implement that with any degree of reliability over a 10+ year lifetime. Another way would be to have smaller fixed banks of LEDs at steep angles at each end of the tank to use during dawn/dusk to simulate a steep angle for the sunlight. I'm pretty sure that's the direction I'll take with my big tank - it's easy to implement, reliable, and "good enough" to achieve the effect I want.

 

[cyrusthevirus]

Interesting. I just ordered the last peices I need for this build so now I am just waiting on FedEx and UPS to deliver them. I ordered many more circuit boards than I needed because there was a minimum order. I also ordered a couple of extra probes. The pH probes only cost me $6 each. Shipping was a little high, so that is why I ordered extras. I figured if they worked, others might be interested in them.

 

[ke6guj]

cool thread

 

[cyrusthevirus]

pH/Salinity and ORP circuit boards have been soldered. They are below. Now I am just waiting on the probes, They should ship next week. I received all of the LEDs, the U chennel and the fans, now I just have to find the time to build.

pH/ORP before soldering:

pH/ORP after soldering:

Salinity before soldering:

Salinity after soldering:

 

[Reef_Matrix29]

Sweet!

 

[cyrusthevirus]

Okay time for another update. I have been working on this for the past several nights. The LEDs are mounted to the 1 3/4" "U" channel using Arctic Silver Thermal Adhesive.

The SilenX iXtrema cas fans have been installed. These fan are 12VDC and only 14dBA so they should be ultra silent.

The mounted LEDs have been soldered together into 8 different banks and are about 1 3/4 to 2" apart.

I am still working out the details of my constant current. After feedback from Stu and DWZM I decided to review my overall power circuit. Besides them many people advised me to avoid trying to power LEDs with a constant voltage and instead they require a constant current. If you power based on a constant voltage, as the LEDs heat up and cool down, their voltage usage changes. These changes could have a negative impact on the life of the LED.

My original plan was to use an LM317 to control the circuit, but after the forward voltage of the LED I only have between 1-2V for regulating the current and the LM317 needs 3.25V to work correctly. Instead of removing an LED I did some searching and came across the following: http://www.candlepowerforums.com/vb/showthread.php?t=236260

This only requires < 1V and could eliminate several components from my original design and increase the simplicity.

So once I get these components I can finish wiring the LEDs and provide power to them.

 

[der_wille_zur_macht]

Maybe my perception is skewed, but why not just go with a simple purpose-built LED driver IC like the CAT4101? Smaller parts count, a bit more efficiency. . .

Not trying to change your mind, just wondering if you have thoughts. It interests me to see someone building a driver "from scratch" in this manner, too.

 

[cyrusthevirus]

The last thing I started working on was an ATO for the tank. After alot of research I decided on using a solenoid, timer and mechanical float to ensure I could manage the water level in my sump.

Here is how I am proposing to put it together. (I am waiting for all of the parts before I can build it.) My plan is to have the water flow directly into the sump from the RODI.

I have two 12VDC 100% duty cycle solenoids. They will be connected together in series. If one fails open, I am hoping the other will still function and remain closed.

The solenoind will also be connected to two float switches in the tank. One will open the solenoid and the other will then close them. The power for the solenoids will be controlled by a time which will run at least once a day. That way I can ensure the RODI is able to produce at least 1-2 gallons to avoid TDS creep.

Finally after passing through the RODI, It will pass through a mechanical float. The mechanical float will be placed several inches above the electrical floats. That way if everything fails, the mechanical float will prevent total disaster.

All of the parts for the ATO come to less than $130 including shipping. The solenoids and miscellaneous plumbing came to $60. The mechnical float is $30. The electical float and relay came to $33.