Come in and join me....

[evilc66]

Sure. You need to add the pot in series to the resistors that set the current. That way you always have a maximum current value set by the fixed resistors and you can't overdrive the LEDs. You just need to make sure that the pot is rated for the same power as the fixed resistors. Make the pot as small as possible too. No more than 50 ohm. Larger pots will make it very hard to adjust. This would be perfect. At 25 ohms it would give you about 20mA of current. The LED probably wont light under 100mA though. Should give you fairly fine control.

[KDris]

Sure. You need to add the pot in series to the resistors that set the current. That way you always have a maximum current value set by the fixed resistors and you can't overdrive the LEDs. You just need to make sure that the pot is rated for the same power as the fixed resistors. Make the pot as small as possible too. No more than 50 ohm. Larger pots will make it very hard to adjust. This would be perfect. At 25 ohms it would give you about 20mA of current. The LED probably wont light under 100mA though. Should give you fairly fine control.

 

Thanks you very much for your help. Based on my understanding, a 1.2 ohm resister and the 25 ohm pot will give precise tuning and not allow the current above 1amp, correct?

[evilc66]

With a 1.2 ohm resistor, you will have a max current of about 0.5A. The equation that was used on the web page stated that output current was calculated by 0.5/R3=CurrentOut. To get 1A, you need 0.5 ohms of resistance. This current should be set based on the LEDs you are driving. Unless you have excelent cooling, I wouldn't drive most LEDs beyond 700mA (about 0.7 ohms resistance).

 

This resistor will set your maximum current, and the pot will tune it down. A 25 ohm pot should give you pretty good control.

[KDris]

With a 1.2 ohm resistor, you will have a max current of about 0.5A. The equation that was used on the web page stated that output current was calculated by 0.5/R3=CurrentOut. To get 1A, you need 0.5 ohms of resistance. This current should be set based on the LEDs you are driving. Unless you have excelent cooling, I wouldn't drive most LEDs beyond 700mA (about 0.7 ohms resistance).

 

This resistor will set your maximum current, and the pot will tune it down. A 25 ohm pot should give you pretty good control.

 

Ok I just looked back at your original equations. So to calculate the resistance you divide the 1.25 source voltage by the desired current output. In my case that would be 1.25V/1A=1.25 ohms, correct?

 

"Using ohms law, dividing the voltage from the output (1.25V) by the current you want to supply the LEDs, which in this case will be 350mA for now. 1.25/.35=3.57 ohms. Digikey has pretty much every resistor value you could ever want."

 

Also I plan on driving them between 350 to 700ma probably closer to 700,I just figured that 1A would be a safe maximum.

[evilc66]

Just saw what you are doing! You are refering to the original regulator that I started using way back when. The equations that I stated in the last post were for the new regulator that runs considerably cooler than the original. Your math does work out though. I would recomend looking backa page or two and find the link I posted for the newer design. Few more parts, but still simple. Make sure you size the parts for the wattage at the highest current.

[abe]

i sure can show you mine. i'm still running the 48" but the 72" should be coming in 2 weeks. so, what would you like evil? i can't rip it open, sorry.

 

i can mabe open up the power supply though

 

EDIT: how do i subscribe to thread?

[evilc66]

Look back to post #302. You see that some of the blue LEDs are a slightly different color? Do you have the sma problem with yours, or are yours all a uniform color? A good look in the power supply unit would be cool if you can.

[abe]

Look back to post #302. You see that some of the blue LEDs are a slightly different color? Do you have the sma problem with yours, or are yours all a uniform color? A good look in the power supply unit would be cool if you can.

i will look into it. all i know is that when i was programing the fixture, there are two different sets of blue lights, daytime blue lights, and moonlights, as there are two sets of white lights, which are daytime and moonlights as well, but they are just different colors, there is not difference in the power, its wierd, because i never see white at night. i will try and get pictures of the power supply and also of the programing menus. will also take closeups of a section of lighting.

[KDris]

Just saw what you are doing! You are refering to the original regulator that I started using way back when. The equations that I stated in the last post were for the new regulator that runs considerably cooler than the original. Your math does work out though. I would recomend looking backa page or two and find the link I posted for the newer design. Few more parts, but still simple. Make sure you size the parts for the wattage at the highest current.

 

Doh! My mistake, thanks again Evil. In your opinion would it be worth it to build the new design? I have a half dozen mosfet heatsinks lying around if that makes a difference. Can you connect 3 leds with 12V or only 2 like your first regulator?

[evilc66]

The current regulator based on the LM317 runs too hot to be practical at anything above 0.5A, even with heatsinks. The new one runs so much cooler. Like I mentioned before, I have 9 of these driver circuits in a small Radioshack project box with a small fan blowing across them, using the lid as a heatsink and it gets warm to the touch at ~700mA. There are a few more parts, but nothing out of the ordinary. All parts can be obtained through Digikey or Mouser.

 

Digikey parts list as follows for what I used:

 

FQP50N06L-ND N-Channel MOSFET x 1

2N5088BU-ND NPN Transistor x 1

PPC1.5W-1CT-ND 1.5 Ohm Resistor @ 1W x 2

100KXBK-ND 100K Ohm Resistor @ 1/4W x 1

 

I used 2 of the 1.5 ohm resistors in parallel to acheive about 0.75 ohms to get the current I wanted, or close to it.

 

Runs less than $2 per driver if you order in quantity (10 pieces, not 1000's).

[KDris]

The current regulator based on the LM317 runs too hot to be practical at anything above 0.5A, even with heatsinks. The new one runs so much cooler. Like I mentioned before, I have 9 of these driver circuits in a small Radioshack project box with a small fan blowing across them, using the lid as a heatsink and it gets warm to the touch at ~700mA. There are a few more parts, but nothing out of the ordinary. All parts can be obtained through Digikey or Mouser.

 

Digikey parts list as follows for what I used:

 

FQP50N06L-ND N-Channel MOSFET x 1

2N5088BU-ND NPN Transistor x 1

PPC1.5W-1CT-ND 1.5 Ohm Resistor @ 1W x 2

100KXBK-ND 100K Ohm Resistor @ 1/4W x 1

 

I used 2 of the 1.5 ohm resistors in parallel to acheive about 0.75 ohms to get the current I wanted, or close to it.

 

Runs less than $2 per driver if you order in quantity (10 pieces, not 1000's).

 

Thanks for clarifying

[evilc66]

No trouble. Sometimes simple information like this can get lost pretty quickly in a 17 page thread

[abe]

i could not open the powersupply... couldn't find the screwdriver...

 

 

light display:

 

menu:

 

that's not the whole menu, but it might explain why there are different shades of colors for the lighting... i still don't get the green, but it makes it look nice.

 

the lights look really wierd btw, because its on auto, with cloud cover on, so all the lights are controlled separately. if it was manual, it would be more evenly shaded

[coolwaters]

here i found this just now

http://www.dealextreme.com/details.dx/sku.13985

you can run 10-17 3w LEDs.

i would run 11 or more just incase..

[evilc66]

Thanks for the pictures Abe. At least now it looks like the inconsistencies with the blues may be limited to the unit I was testing against. Yours looks fairly uniform. The two whites in the second row from the left look very yellow. Maybe thats just from driving the LEDs at a very low current from the cloud cover mode. I don't think they are a warm white. PFO has never made any mention of using them before.

[Gomer]

Extremely tempting.

http://www.dealextreme.com/details.dx/sku.13985

[evilc66]

would be nice if it was dimmable.

[Gomer]

would be nice if it was dimmable.

even if it isn't.

One of these is less then a single xitanium, and does the work of 3 of them.

two of these, and 6x6 array. Lovely!

 

I wish there was a cree strip MCPCB for high density arrays.

[Gomer]

would be nice if it was dimmable.

 

 

Perhaps someone is really good at figuring out the circuit topology. If we are lucky, we can swap the regulation resistor for a pot.

[evilc66]

Thats a possibility. Would be interesting to see what the driving chip is on the bottom. Chances are it follows the reference design of the manufacturer and would be easy to hack into. Might be worth buying to find out. I'd like to get a scope on it too to see how clean they filter the ac noise after the bridge.

[Zombo]

Thats a possibility. Would be interesting to see what the driving chip is on the bottom. Chances are it follows the reference design of the manufacturer and would be easy to hack into. Might be worth buying to find out. I'd like to get a scope on it too to see how clean they filter the ac noise after the bridge.

 

Filter caps FTW!

[evilc66]

It looks like it has some pretty good size electrolytics on it, but that doesn't nessesarily mean it working well. Maybe I'll order one and take a peak at it. $18 and free shipping isn't bad for an experiment.

[Zombo]

Let us know what you find out, since I'm not too hyped on dimming for a 4g Finnex, it would be perfect for driving 2 or 3 Q5's and half a dozen blues

[KDris]

Evil where did you get the pcb's for your leds? I can't find bare stars anywhere.

[Zombo]

Here you go!