newbie Lamp-dropper

[Viitalahde]

Anyone tried to make white light out of a red, green & blue leds?

 

[Guest]

[quote author="Svart"]before you make the mistake of using bulbs, which you will have to change fairly often (as i have learned the hard way..), I bought a bag full of warm white leds. these are so close to incandescent color temp it's uncanny, .[/quote]
Espeically if you use these PCBs
http://www.groupdiy.com/index.php?topic=178369&highlight=epe#178369

 

[Svart]

www.ledtronics.com

they have most every kind you can think of.

ebay has them from hongkong and such fairly often.

Hey Jaakko, I tried that once, you can get fairly close but the intensity of the LED is key. you have to get the intensity AND the color band right or it will look strange.

 

[kruz]

thank you svart!

 

[enthalpystudios]

[quote author="pstamler"]The smoking gun here, if you'll pardon the expression, is when you said "LM317 48V reg." The LM317 isn't really designed to handle 48V; you can trick it by floating the adjust terminal, but it still fries when you ask it to drop that much voltage between its input and its output. So while the LM317 has current limiting in it and therefore ought to be immune to short circuits, it's not immune when you're floating it for 48V.

Since you've already fried it, try substituting a TL783 high-voltage regulator, once again available from TI in a TO220 package. It's designed for this sort of thing.

Meanwhile, get a good magnifying glass and figure out what voltage that bulb wants to see, then measure its resistance with your handy Fluke. Ohm's Law tells you Vbulb/Rbulb = Ibulb; figure out how much voltage you need to drop from your power supply (Vdrop = Vsupply - Vbulb), and calculate Rdrop = Vdrop/Ibulb. Ibulb(squared) x Rdrop = power dissipated in the resistor; make sure the resistor's power rating is at least double that.

Peace,
Paul[/quote]

You know, I have like 4 tl783 that I ordered from a fellow in the black market for a g9... i figured i'd order more since they're a) cheap and b) hard to find.

I can try this. I'll investigate the datasheet myself, but perhaps someone knows off hand if it uses the same formula to determine the resistance between Input and Adjust?

Well, I'll check, maybe i'll learn something. Because I've wondered why, when my hands go a-fumbling, no matter what shorts, its the 48v reg that blows, the +/-15v regs (also 317/317) dont seem to crap out. And thats why: floating the input voltage negates the current limiting protection... right?

and also, thanks for the note about the lamp, i'll try that, for the learning and all. I may, however, take svart's note and go for putting a white led behind the jewel. same look with all the plus's of an led.

thanks again you guys!

billy

 

[enthalpystudios]

well, looks to me like they both calculate Voutput based on the same formula, and both have the same internal reference voltage, and are both to-220, so I should be able to drop it in place.

I'll go give it a go, it'll be nice to not worry (in use) that my voltage reg might go out and pump 70 volts into a mic.

billy

 

[The Kid]

Only if you hang the resistors on a constant-current source. Connect them to a voltage source, and...well, let's say it's 10V. Hang 100k on it and you get 10/100000 = 0.1mA. Hang another in parallel and it will draw another 0.1mA for a total of 0.2mA. Add two more and you're drawing 0.4mA through your total of 25K.

Right, but that's not what he did.


Pstamler & Al:

I'm trying to use a series resistor to drop the voltage going to a "power on" lamp. It's a 6V lamp, and I've got 15vdc and 15vac to feed it.

If I parallel 4 100k 1/4 watt resistors, do I have a 25k 1 watt?

The problem at hand is he wants a series resistor for the lamp. He only has certain values though, so he wants to sub 4 in parallel for the series. (If that's not what enthalpystudios is saying, then please correct me!) That would give you a series-parallel situation, and what I said applies. You are correct in your parallel rule usage, but it doesn't apply here, unless he actually did hook them in parallel across a voltage source & the lamp, which would make no sense at all! :thumb:

 

[Svart]

yikes. this got confusing even for me. I am not even sure what he wants anymore... :?

Since I usually use LEDs I'll start with the specified I and then solve for the R based on the V. if it will dissipate too much heat I try a regulator before it. Usually you can find a low enough rail to hang an LED on and not even put a dent in the current draw.

 

[alk509]

I say tomatoe, you say tomato... I say 4X = Y, you say Y/4 = X.

Same ****, different smell.

Peace,
Al.

 

[The Kid]

This is confusing enough for newbs without bringing conductance into it! :grin:

 

[alk509]

The thing with lamps is that their resistance changes a lot from when they're cold to when they're hot, so without the lamps resistance, you can't really apply Ohm's law right away...

Do the pot thing someone mentioned above to get your resistor value in the ballpark, build the circuit (and disconnect everything else, ferchrissake! :wink, then measure voltages across the resistor and the lamp and figure out what's up.

Peace,
Al.

 

[enthalpystudios]

yeah, i think i've actually got a hold on the series/parallel things, despite the back and forth. But since I think these lamps are cheap, and I'm not even sure if they're 3 or 6 volts anymore, i'm going to but bright whites in and use the cool housing. Should do the trick well enough.

So instead, lets say I have a white led with a forward voltage of 4.5 volts. With a 15v rail, I use a 350R (maybe 390?) resistor for an led with a 30ma forward current.

My question is, is the voltage drop over the led constant? That is, if I increase my resistor size, am I just limiting current?

If so, than I guess this is all about deciding how much current you want to run through it. I didn't think that was an option, I though the current was the constant.

Billy

 

[mwkeene]

[quote author="enthalpystudios"]yeah, i think i've actually got a hold on the series/parallel things, despite the back and forth. But since I think these lamps are cheap, and I'm not even sure if they're 3 or 6 volts anymore, i'm going to but bright whites in and use the cool housing. Should do the trick well enough.[/quote]

Yeah, I hope the back and forth didnt confuse you too much, I'm almost sorry I said anything at this point...

[quote author="enthalpystudios"]So instead, lets say I have a white led with a forward voltage of 4.5 volts. With a 15v rail, I use a 350R (maybe 390?) resistor for an led with a 30ma forward current.[/quote]

15v - 4.5 v = 10.5v

10.5 V / .03 = 350 ohm.

yes, sounds good.

[quote author="enthalpystudios"]
My question is, is the voltage drop over the led constant? That is, if I increase my resistor size, am I just limiting current?[/quote]

Yes, picture your LED as a voltage source that subtracts from the 15V source. you now have a fixed 10.5 volts across your resistor. If R goes up, then I will come down cause V is the same. If R goes down, I will come up for the same reason. The voltage across the LED will change a little with current since the VI curve aint exactly straight up and down, but not enough to really worry about in this case. VI curve for diode Look at the upper right side quadrant. Notice how the voltage changes very little for a big change in current? Thats what makes this device non-linear and therefore will subtract about the same voltage for a big change in current.

[quote author="enthalpystudios"]If so, than I guess this is all about deciding how much current you want to run through it. I didn't think that was an option, I though the current was the constant.[/quote]

Not on an incandescent. Like Al said, with DC and stable T it acts more like a resistor.

-Mike

 

[PRR]

> is the voltage drop over the led constant?

A good red LED will be 1.6V to 1.7V at any useful current. Bad ones may rise a bit at high current, but maybe 2V. The exact value of current is usually not at all critical: 10mA and 12mA look the same to the eye. So even if the source is only 5V, and you can't know if the resistor will drop 3.4V or 3.0V, it really does not matter.

Green and white LEDs have higher drops. IIRC, white has a much higher drop than classic LEDs. Use the Typical spec, calculate a resistor. They try to find a Min/Max spec on voltage drop, and see if that resistor with that supply voltage is in any danger of putting current too low to see or so high the LED fries.

> say I have a white led with a forward voltage of 4.5 volts. With a 15v rail, I use a 350R (maybe 390?) resistor for an led with a 30ma forward current.

Well, OK. Although I might throw in a 1K resistor, ~10mA, and see if it is bright enough. We often do not have to run LEDs at maximum current.

 

[pstamler]

[quote author="enthalpystudios"]Well, I'll check, maybe i'll learn something. Because I've wondered why, when my hands go a-fumbling, no matter what shorts, its the 48v reg that blows, the +/-15v regs (also 317/317) dont seem to crap out. And thats why: floating the input voltage negates the current limiting protection... right?[/quote]

Yes, and here's why. Let's say you're using the 317 in its normal state, with say 20V at the input and 15V at the output. Short the output, it goes to 0V, so you have 20V between input and output, which it doesn't mind in the slightest, and the current limiting limits and prevents the thing from frying.

But let's say you're floating it, with 55V on the input, 47V on the adjust terminal, and therefore 48.2V on the output. You're cool; the voltage between input and output is 6.8V. Now short the output; it's at 0V. Your input-output voltage is 55V. Boom. It's not the current that fries the chip, but the excess voltage.

Peace,
Paul

 

[bcarso]

Most white LEDs are blue LEDs with a semitranslucent phosphor overcoat. So they have the same forward voltage vs. current characteristic as blues.

 

[enthalpystudios]

awesome, thanks you guys, both things, the leds and the lm317's, are making quite a bit of sense now.

I'm going to give the led a shot at 1k. Also seems I could use a 1/4 watt at that current, which is nice.

I have tons of 1/4 watt values, only a few in 1w and higher.

billy

 

[Larrchild]

And it only took 3 pages!

Billy,
don't be a hero
come back again.

 

[Svart]

Most white LEDs are blue LEDs with a semitranslucent phosphor overcoat. So they have the same forward voltage vs. current characteristic as blues.

Actually they are more UV than blue still with the phosphor coating though. same principle as a flourescent light.

I'm doing work with 5w white LEDs.. talk about damn bright .

 

[enthalpystudios]

yeah svart, i'm going to try something like that I think.... there are lots out of hong kong on ebay for decent prices.

and blues, for another project

billy