Need help from a Math wizzard/LED Geek

Paul B · Nov 14, 2015

twilliard said:
That is calculated at 6 volts.
Radio shack may not carry metal oxide I think they are all carbons, well where I live.

What is calculated at 6 Volts? Don't I need to supply each LED with 3.3 volts? They are in parallel so the voltage will remain the same, no?

twilliard · Nov 14, 2015

Paul B said:
What is calculated at 6 Volts? Don't I need to supply each LED with 3.3 volts? They are in parallel so the voltage will remain the same, no?

No that is the forward voltage.
Hard to explain..
At 3.3v it will light up now current will drive it.

Paul B · Nov 14, 2015

OK, so I need to supply each circuit with 6 volts? Doesn't sound right but I am a little dim

twilliard · Nov 14, 2015

No Paul your not dim.
Its all about current. The resistor will drive that chip at 330ma at 6 volts.
I can do the calculation at a lower voltage but but puts the resistor at a range you will have a hard time finding.

twilliard · Nov 14, 2015

Here are all the resistance values you can find

twilliard · Nov 14, 2015

Also Paul this is assuming they are 3.3v chips.
Best way to tell is to set up a single chip and slowly raise the dc voltage till it lights up. That will be your forward voltage of the chip.

Paul B · Nov 14, 2015

Well I know if I put a voltmeter across a lit up LED it reads 3.3volts. So 3.3 sounds good. So if I got this correct, I can feed the copper tube with 6 volts, put an 8.2 ohm 1 watt resistor in between the LED and the tube and that will work? OK, that is easy. I just need to go out and get 60 1 watt 8.2 ohm resistors. Radio Shack used to be great but now they basically sell cell phone cases so I should probably get these resistors on line. If I have time tomorrow I will build the frame and radiator. Thanks, I will love to try this out.

Like Duh, I forgot about the resistor when figuring the voltage. See what happens when you get old. I just thought about that. :mad:

Jimmyb819 · Nov 14, 2015

The resistor will be taking care of the extra voltage above the 3.3 volts that the LED needs.

Paul B · Nov 14, 2015

Yeah I got that after I wrote it. So now I am wondering if I should buy and solder all of those resistors, which is a lot of work because I think it is less work to just run the stupid things in series without the resistors. I wanted to use the pipe as a conductor to eliminate wires and some work, but it seems it will be a lot more work to use 60 resistors as that would still add a lot of wires to the tubing and make it ugly.
I am thinking.
OK, I have another question for you guys on the same circuit. Can I run 20 LEDs on each tube in series and use the tubing for the return to the existing driver that I have? The problem is that for the 3 circuits I will be using that means I will have to have one wire from each driver connected together to the tube. Can the return wire from each of the 3 drivers be connected together? This would still allow me to only have only one wire going along the tube which will be the wire connecting the LEDs together and I can still use the tube as the return allowing me to use the drivers I already have. I still think I would need a resistor in there because those drivers are running about 30 LEDs now so 20 may not be enough. Or am I thinking wrong?

twilliard · Nov 14, 2015

Digi key is a great company!
I order custom wire wound resistors and they have never failed me :-)

Paul B · Nov 14, 2015

Like this. 20 LEDs on each run. No resistors but the drivers are still connected to the tube. If I did this, I would eliminate the resistors, still have only one wire to deal with as the other wire will still be the copper. Is this doable and would I need a resistor because there are now about 15 less LEDs for each driver from what was running on the original fixture. I am not sure if I can connect the drivers together like that.
I mean as long as you guys said you like designing this stuff.

TheEngineer · Nov 14, 2015

Paul B said:
I want to dump the heat into the cooling system and not dissipate it through resistors. Each of these LEDs only has a tiny bit of heat sink built in but that is what they were designed to use. They were bolted to an aluminum sheet and we already determined that copper is much better heat transfer so at least that is good news. Of course there is some loss because the tubing is round and the LEDs are flat. This I can help if I want to solder flat disks to the pipe about as large as the heat sink on the LED. That would be very easy and I think I will do it that way. That way 100% of the base of the LED will be in contact with the metal that would be soldered to the tubing so it will be a very efficient heat transfer. Much more efficient than this LED fixture was designed for. I know you guys can figure this out. I have great faith in you. :D

Sorry, I'm clearly not explaining myself very well....

In parallel you will be driving a constant voltage (~3.3VDC ideally). You then need to add resistors to control the current since you can't control the LEDs with voltage wired in parallel. The resistors do their job by resisting the flow of current. This does two things: 1) limits the amount of current running through that node and 2) creates heat. There is no way to avoid generating heat across the resistor, that is the physical effect that causes the current to be limited. The energy changes from a flow of electrons into a flow of heat.

TheEngineer · Nov 14, 2015

Paul B said:
Like this. 20 LEDs on each run. No resistors but the drivers are still connected to the tube. If I did this, I would eliminate the resistors, still have only one wire to deal with as the other wire will still be the copper. Is this doable and would I need a resistor because there are now about 15 less LEDs for each driver from what was running on the original fixture. I am not sure if I can connect the drivers together like that.
I mean as long as you guys said you like designing this stuff.

You can do this, but you'd be better off using a couple extra feet of wire instead of wiring to the copper at all. It doesn't add anything for you here. It isn't as much effort as it seems to do this soldering, you get into a rhythm.

TheEngineer · Nov 14, 2015

Paul B said:
Icecool, you have the perfect name for heat transfer.

Ha! My 10 year old self was very prescient. :)

twilliard · Nov 14, 2015

Running with a 3.3v line doesn't allow you to control the ma of the chips. That is forward voltage for the diode to function.
Running a 6 volt line will allow you to run the chips at any given ma rating depending on the resistor value installed.
This is how we drive led chips. Current

Paul B · Nov 14, 2015

Icecool. I know about power heating up resistors. I just didn't want to use to much power that I had to later reduce with resistors. And I am very good at soldering. I have been an electrician for 40 years and went to quite a bit of schooling in electronics, but that was before LEDs. I also invented and patented the Majano Wand which has quite a bit of soldering. But I am quite stupid when it comes to figuring the fv, and resistance needed and all that when it comes to LEDs. As an old school kind of guy my mind works in things like wood, copper and did I say wood? I know I can just run another wire from the LEDs to the drivers, but "can" I use the tube?
There are a lot of things I should do but even more things that I would want to do just because it is more interesting. My tank runs a reverse undergravel filter and I built most of the rocks just because it is more interesting.

Paul B · Nov 14, 2015

I want to thank all of you guys for all the help. Now I have some different ideas on how I am going to wire this. The plumbing will be easy as soon as I find a small pump that I can plumb into the system. I need to look around to see if I have a pump that will work. Then I need to build the safety's like the thermal switch that will turn off the system if the water is not flowing and the tubing gets to hot. I will post pictures as I build the thing.
Thanks again, Paul :cool:

TheEngineer · Nov 14, 2015

twilliard said:
Running with a 3.3v line doesn't allow you to control the ma of the chips. That is forward voltage for the diode to function.
Running a 6 volt line will allow you to run the chips at any given ma rating depending on the resistor value installed.
This is how we drive led chips. Current

Running the LEDs in series is a voltage driven control circuit, running them in parallel is a current driven control circuit. For the parallel setup, the voltage needs to remain constant. To dim the lights he would need to change the resistance of the current limiters. No longer an issue since Paul said he is going to do this in series now.

Paul B said:
Icecool. I know about power heating up resistors. I just didn't want to use to much power that I had to later reduce with resistors. And I am very good at soldering. I have been an electrician for 40 years and went to quite a bit of schooling in electronics, but that was before LEDs. I also invented and patented the Majano Wand which has quite a bit of soldering. But I am quite stupid when it comes to figuring the fv, and resistance needed and all that when it comes to LEDs. As an old school kind of guy my mind works in things like wood, copper and did I say wood? I know I can just run another wire from the LEDs to the drivers, but "can" I use the tube?
There are a lot of things I should do but even more things that I would want to do just because it is more interesting. My tank runs a reverse undergravel filter and I built most of the rocks just because it is more interesting.

OK, we were talking past each other. :) You made the majano wand, eh? That's pretty cool!

The only way to avoid getting too much heat into the resistors is to get as close to the forward voltage and reverse current of the LED as possible. Since you decided to go with the series route, it isn't an issue anymore.

You certainly can use the tube, but not for each LED in a series configuration. You could use it for the ground for the last LED in the chain. Make sure you make all of them common though.