New tank setup with LEDs and microprocessor

[cyrusthevirus]

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.

I have not found one that meets my specs. My lights are comprised of 16 LEDs and powered by a 3.3V PIC which requires a driver which has a very low RdsOn. I have not found anthing that can do this without building it myself.

 

[cyrusthevirus]

If the above was not enough I have two tunze 6105's that I plan to control using my two remaining PWM channels. I was able to find two old keyboards and cut the cables off for connecting into the power supply.

 

[cyrusthevirus]

Okay so I have been very busy. Here are some pictures of the LED setup for my build:

Tracing the fan for the hole:

Hole cut rubber gromets in place

The fan, only 14 dBA at 12VDC

The fan mounted on reflector:

All four fans mounted and placed back in light hood.

Now on to the LEDs.

 

[cyrusthevirus]

It's LED glue time:

Mix up the epoxy:

Begin laying them out and attaching them. Remember to work in small batches, you only have 5 minutes before the epoxy sets.

And with the miracle of the Internet, they are all attached and some wired. (2) 3' strips and (4) 4' strips.

Mostly wired up:

I ordered the PCBs for the constant current circuit today so I will get 37 of them in a week. It was cheaper to get 37 than the 8 I really need. This circuit will take a 58V 1A power supply and drive 16 LEDs either with or without a PWM signal. The overall cost for the DIY driver falls to less than $15.

 

[cyrusthevirus]

Sorry for all of the updates, I have not had the time to post.

While I wait for the PCB to drive the LEDs, I decided to move on to my ATO.

The heart of the ATO are two 12 VDC solenoids which run at 1.7A each. Two are used for redundancy:

These are wired to a relay which controls an upper and lower float switch. The low switch opens the solenoids and the upper switch closes them. That way I can run enough gallons to prevent TDS creep.

The relay:

The DIY hanger:

The DIY clips for the floats:

Here it is all put together. For snail guards, I used new pill bottles. They are free just ask your local CVS for two.

On the right of the picture you will notice a mechanical float. The is the final failsafe. One thing I found out is the model I bought from the filterguys already had the switch built in. All I need to purchase was the float valve. I which I realized this before I made my purchase of a mechanical float kit.

Here is a shot from above:

 

[taklein65]

Looks good tagging along.

 

[stugray]

Sorry, but...You MUST NOT use brass valves on the output of your RO/DI.

Hopefully those are on the INPUT to the RO/DI????

Stu

 

[cyrusthevirus]

They are on the input. The water flows through the solenoid and then into the RODI.

I am curious. Why no brass after the RODI?

 

[stugray]

Brass is an alloy of copper & zinc.
RO/DI is partially corrosive until it gets some of its Ions back.

Running RO/DI through brass would definitely put copper back into your system ( one of the things the RO/DI filter was used to remove ).

Stu

 

[Titeo]

any update of your ph/salinity and orp working???

 

[cyrusthevirus]

I got the probes, connected them and I am getting some readings, so they appear to work, I just have to spend some time calibrating them. I was trying to avoid purchasing calibration solutions and using something DIY, but it looks like there are no good DIY options available.

I am currently working on the Molex connections between the lights and they are a PITA. Once I get them figured out, I will move back to the PIC and the ORP, EC and pH.

 

[cyrusthevirus]

I received the constant current boards (all 37 of them).

Here are the eight I need for my build:

Here is a close-up of one:

Now I have to determine where I am going to put them. Close to the power supply or near the LED strings.

 

[cyrusthevirus]

any update of your ph/salinity and orp working???

I just ordered pH 7 and pH 10 buffer solutions. Once they come in I can begin testing. I hope by next week.

 

[cyrusthevirus]

I got my test solutions and found out that I musted have missed a wire when I setup the circuit in the Eagle software. No problem, I just ran a jumper wire.

The pH circuit seems to work fine. I just started with it and do not have it connected to my pic yet. I am using a multimeter to measure the voltage. 1.709 = ph 7 and 2.441 = pH 10. I am all calibrated so let's see what happens.

 

[dainiusiva]

very cool project! I'm impressed by your knowledge

 

[cyrusthevirus]

I wish I could take credit for the knowledge, I have been more of an aggregator. Taking pieces of other people's projects and reengineering them. No need to reinvent the wheel.

 

[dainiusiva]

No need to reinvent the wheel.

Words of wisdom

 

[cyrusthevirus]

Can you show me how your LCD and your controller board to control the Led strings??? Like pwm, timer and the sunrise/sunset. Is your board get hot when operate?? Because 58v is a lot.

I was asked the following questions in a PM so I wanted to share it with everyone incase they had the same question.

If you look at the post on 5/3 you will see a new circuit I am using for the LEDs. I am no longer using the ltc1157 to control the LEDs (I am however using them to control my Tunze 6105 pumps.)

The LCD is only for display, it does not control the LEDs. The PIC24 controls the LEDs using one of the PWM outputs. I then use the RTC with alarms to run a sub routine which slowly increases or decreases the lights.

On the LED controller board there is only one element that can get hot and that has a heat sink in the picture. If it is running without the PWM it gets warmer than running with the PWM.

Although 58v may seem like a lot, it is only running at 1A. Plus the Fv of the LEDs reduces it down to 1-2v.

 

[cyrusthevirus]

Started building the power center. It consists of six power supplies, six outlets, six switches and six constant amp circuits.

Here is the build:

 

[cyrusthevirus]

Here is the enclosure for the leds. It is a converted fluorescent hood.