Sciencemadness Discussion Board - Photochlorination apparatus, setup of multiple UV LED lights, help please...

Photochlorination apparatus, setup of multiple UV LED lights, help please...

BromicAcid - 5-7-2005 at 17:11

Recently I bought a bundle of 100 UV LED lights from eBay. They work well but I know nothing about setting them up to function as more then one at a time. Acutally, when I do get one to light it's overkill and the light starts out as that UV blue color then turns orange suddenly then goes out forever. So I don't want to make a setup of 50 or so lights and run power through it and suddenly burn them all out at once.

So, here is the specifications for the LED lights that I got:

Quote:

Specifications
# Size (mm) : 5mm
# Emitted Color : ULTRA VIOLET
# Forward Voltage (V) : 3.2 ~ 3.8
# Reverse Current (uA) : <=30
# Life Rating : 100,000 Hours
# Viewing Angle : 20 ~ 25 Degree
# Absolute Maximum Ratings (Ta=25 Degree)
# Max Power Dissipation : 80mw
# Max Continuous Forward Current : 30mA
# Max Peak Forward Current : 75mA
# Reverse Voltage : 5 ~ 6V
# Lead Soldering Temperature : 240 Degree ( < 5Sec)
# Operating Temperature Range : -25 Degree ~ +85 Degree
# Preservative Temperature Range : -30 Degree ~ +100 Degree

So, what I want to do is insert 50 of these lights into a setup and have them all function at once. And here are my questions:

Can I connect them in series? And if I do so what should be the specifications of my power supply, volts, amps, how can I figure it out so I don't blow all of these bulbs or get no illumination.

I did some work in my physics class with light bulbs and series and parrallel setups and such but the resistances were much larger and we only used two bulbs at once, not fifty so any other pointers would be great.

IrC - 5-7-2005 at 17:43

They are turning color as they scream in pain and then die forever. I would give you a value but you did not specify what voltage you are going to use. You should wire them all in parallel, not series! The drop will vary too much in each one and they will not all turn on the same. Some will be too dim, some will probably burn out. Each LED needs it's own resistor, typically 470 ohms 1/4 watt if you are using say 12 to 14 volts to power them. So you will have 50 LED's and 50 resistors, all 50 +'s (long LED lead) together, and all 50 minuses together, with a resistor on one side or the other of each one. Since R=E/I, if E is say 12, and I is say .03 or 30 Ma, then R= 400. Best to use the standard value of 470 ohms, this will lower the current to about 25 milliamps in each led, giving very long life. Total draw would be .0255 X 50 = 1.275 amperes. So a nice little 2 amp 12 volt supply would be good for this setup. Do not go any other way here or you will lose LED's and have poor light output.

Remember after you decide upon voltage to just figure ohms = volts/amps to get the resistor value, and then pick the closest standard value. Personally, I would have used a mercury bulb which was designed to go in the little eprom erasers (little bulb about like a flash bulb in size with threaded base the standard nite lite size), much higher output and more in the high UV end of the spectrum. Also real easy to hook up as they went straight to the 117 Vac line through a 10 uF 250 volt nonpolar cap, along with a switch. Very simple! You will find the LED's are real low in the UV range in frequency. I know, I have a box, likely from the same seller you got yours from. In any case pick up a bundle of 1/4 watt 470 ohm resistors before you start building your setup. I know it's a lot of work wiring but make a little circuit board and it will be much nicer. At least you don't have to do 98 like I did when I built my night vision illuminator from hell using IR LED's.

BromicAcid - 5-7-2005 at 17:51

Why 12 V if it necessitates the need for a resistor for each LED? I will purchase whichever power supply I need so wouldn't it just be easier for me to get a power supply with a lower voltage to eliminate that need? So parrallel is the way to go, no problem, any other suggestions?

IrC - 5-7-2005 at 18:14

You could eliminate the resistor if you used a constant current supply, but then there would still be variances between LED's. It may work well enough for your project, just multiply the total draw of each by the number of LED's and set the constant current supply to that value. You would need a steady 1.25 amps to run all 50 at 25 mA. The best way to go is the way I described, each LED is ballasted on it's own, and you have a lot of leeway as to supply voltage. What I was trying to give you was a very reliable way to go with all LED's of similar brightness and good protection from harm. If you use less than 12 volts do the math for a smaller resistor value. You could also wire them all in parallel with no resistors and step your way up until there was a 1.25 amp draw for the 50 LED's. There are many ways you can go here. Burning them up or having poor output is hopefully not one of them.

I forgot to mention but think of a supply more in terms of current, rather than voltage (keeping in mind the voltage across each LED should not exceed the specifications). Also, you could do a series-parallel string, but try to stay at a max of 3 in each series string, all of them in parallel. Much more in series than about 3 and you will have a lot of variance in brightness among invididual LED's.

[Edited on 6-7-2005 by IrC]

Twospoons - 5-7-2005 at 20:17

You do want them in series groups of three, with a 68 ohm resistor for each group. Then you put these groups in parallel with each other and run the whole thing from 12V. Doing one resistor per LED is a big waste of power.
This is how I built my UV glue curing lamp, as in the picture below. Thats 105 LEDs on that sucker, all mounted on small PCBs (made by me) and aimed at a point about 60mm from the middle.

The resistors are absolutly necessary, as the LEDs have bugger all resistance themselves, so the current will rapidly get out of hand if the voltage is out by more than 0.1V. The resistor provides current control (of a sort).

[Edited on 6-7-2005 by Twospoons]

Damn. Forgot to add that you can expect those LEDs to get fairly warm, so when you mount them leave about 2cm of wire hanging out the back of each one to act as a heatsink. I left about 1cm on mine, and I'd say the LEDs get to about 40C in one minute. Add a fan if you're concerned about overheating during long runs.

[Edited on 6-7-2005 by Twospoons]

UVLAMP.JPG - 17kB

Twospoons - 5-7-2005 at 20:37

Thought I'd add a pic of it running. You can see the heatsinking wires sticking out the back.

What did you pay for your LEDs? We got 500 for US$ 60, direct from the manufacturer.

[Edited on 6-7-2005 by Twospoons]

UVLAMPrunning.JPG - 10kB

IrC - 5-7-2005 at 20:51

If he used 52 ohm resistors he would have almost 29 milliamps in the LED's. At 68 there would only be 22 mA, and these LEDS fall off rapidly below 25 mA. Since they have little short wave UV output, with most of the energy density at the very bottom of the low UV range, I would want to try to keep at least 25 to 27 mA in them. As to one resistor for each yes it's wasted power, but in the watttage range here it's not much, compared to having very nearly equal brightness in each LED, as compared to dividing the ballasting by three. But as we have both pointed out three per ballast will be adequate for decent performance. Like I said I would up the current above the 22 mA though, just to get more energy in the higher frequency UV range, as well as greater power output. If say the total output is 4000 MCD and only 10 percent of that is at the high end of the range, then 400 MCD of higher energy UV would be present. Say the output was reduced to 2,500 MCD, then only 250 MCD of shorter UV would be present.

I have not seen specs which rate the energy density per UV frequency VS LED current, probably it is not a large variance, so the main thing would be just getting the total output as high as safely possible to gain the greatest amount short wave UV possible. He has not stated a voltage so the values for resistors here are only ideas, if a number can be chosen then the exact value to get the right current would be easy to calculate. Just keep in mind that it is important to very carefully stay within the range of 25 to 29 mA in each LED, no more, no less.

PS: that LED assembly is very nicely made. If I ever decide to build another one of my super IR illuminators I wouldn't mind buying the boards from you. I especially like the work on creating a good focus.

[Edited on 6-7-2005 by IrC]

Twospoons - 5-7-2005 at 21:30

Glad you like it!

Yeah, we spent a bit of time figuring out how to NOT fork out $5000 for a proper curing lamp.

The focus isn't that tight, I get a spot size roughly the same as one of those little triangle boards.

The resistors: I actually used 62ohms, since we have lots of them here, but its an oddball value, so I suggested 68. And 12V is a convenient supply to run off. Also remember that the forward voltage will drop as the LEDs heat up, so it pays to have some margin on the current.

IrC - 5-7-2005 at 22:08

62 is good. I thought 12 volts was a good number also as it is a common supply voltage. What he needs is a little more than a wall wart can do, but those 13.8 volt 3 or 4 amp radio shack supplies are common, in just about every garage sale and flea market on the planet. For mine (though IR and not as much heat in the UV LEDS) I drilled holes in an aluminum reflector, where the LEDS just protrude through it. The metal carries the heat away quite well, but drilling holes in Al is drillbit hell for sure. As to runaway, that was why I mentioned a constant current supply. Years ago I built many laser diode boards for Meredith Instruments, and it was very important, especially in the early laser diodes which gave out extremely easy. In any case sure a lot of work to do 50 LEDS no matter how he goes.
I wish you guys had some of those small mercury bulbs for eprom erasers, at the top end of the UV spectrum with very high output. One bulb would do the whole job with much more UV of shorter wavelength than with a box full of the UV LED's.

Twospoons - 5-7-2005 at 22:31

Eraser bulbs are no good for me. Only a few watts, and its a large diffuse light source, so I can't get the flux density needed. Might work for Bromic though ...

IrC - 5-7-2005 at 22:54

It was just a thought. One of the germicidal tubes would also be cool, I keep thinking for Photochlorination he would need greater shortwave output and I was not real impressed with the UV LED's I bought as far as short wave UV goes. But if it would work there is no doubt it would be a cool project, though time consuming to wire all the LED's. I hope thought is given to making a PCB. My first array was handwired and after that I took the time to make boards, no more of that handwiring a large number of small objects again for me.

Tacho - 6-7-2005 at 03:22

Bromic,

You need to use those resistors, sorry. The led needs a constant current to work. Typically 10mA for common leds. Check the datasheet for your led, but I believe 10mA keeps you on the safe side.
The led internal resistance is close to zero, so you are probably burning them.

So you should put them all in parallel, each with it's own resistor.

Resistor=Voltage/current (R=V/I)

Any voltage should work if you do the math right, but for a 5 volt supply:

R=5/0,01 or 500 ohms

The commercial value you will find is 470 ohms, it should work.

Never mind the resistance of the led. However, if you put a few of them in series, you may need a higher voltage supply, since each led puts a voltage drop in the circuit. It's not due to electric resistance, but to the semiconductor junction (N-P or Arsenide, or gallium or god-knows-what).

Good idea, by the way. I use a large germicide lamp as UV source, but a bunch of leds around a test tube sounds much cleaner and microscale.

NeutralIon - 6-7-2005 at 16:15

Most LEDs work at best efficiency at higher currents -- usually above the max continuous current. This is done by pulsing the current thru the LED.

You might consider a design using a 70 mA peak current [< 75 mA max] and a 25% duty cycle which would give an average current of 17.5 mA. Even at the max. 3.8 V drop the dissipation is <70 mW.

The pulsing could be done using a divide by 4 counter and a 2:4 mux to select one of 4 strings of LEDs that are turned on at any one time.

Twospoons suggestion of 3 LEDs in series with a current limiting resistor is also a good idea. There should be enough drop across the resistor so that the current stays within safe limits regardless of the actual voltage drop of the LEDs.

Twospoons - 6-7-2005 at 20:44

"... higher efficiency..." Well, thats not really true. There is always a maximum pulse current, regardless of the pulse duration, because of the limits imposed by the thermal transient. LED efficiency actually drops as you get into higher currents, though if you're simply after the brightest pulse then its worth doing. There's no real benefit in terms of average optical power.

Besides, running continous at 25mA with a resistor is waaaaaay simpler.

[Edited on 7-7-2005 by Twospoons]

NeutralIon - 7-7-2005 at 16:04

Most LEDs do have a higher efficiency at higher currents. That is [in general] you could expect to get more light out of a 25% duty cycle, 70 mA peak current than with an equivalent continuous current of 17.5 mA. To really know the best operating point one would need to see a curve of lumens vs. current.

Agreed that a pulsed drive is more complicated but might be worth it if you can get more UV photons.

BTW, what is the peak wavelength for these diodes?

IrC - 7-7-2005 at 16:15

The problem with pulsing is fracturing of the substrate. You need to carefully tailor the waveshape and peak amplitudes. Plus it is a lot of circuitry to mess with when you are lighting up 50 at once!

Twospoons, do you have an image of a photoetch pattern for your LED PCB's? I'll gladly pay tuesday for a hamburger today.

12AX7 - 8-7-2005 at 00:35

Pffbt.. just a transistor and 555. Heck, I could do a full-blown adjustable PWM, F and I<SUB>pk</SUB> circuit on a single solderless breadboard. It's pitiful what ICs today, well from the 70s and 80s, can do!

Im

Twospoons - 11-7-2005 at 14:44

IrC : Here's a pdf of my files - schematic, overlay, and pcb copper pattern. I set the pattern up for 0603 resistors, and the borders form the power distribution. You can (of course) use any size resistor you wish, just solder where convenient. Enjoy!

Attachment: UvLampFiles.pdf (47kB)
This file has been downloaded 873 times

IrC - 11-7-2005 at 14:49

Thanks for putting it up. Acrobat wouldn't open the file, saying it was either not a pdf or was broken or something.

chemoleo - 11-7-2005 at 16:12

It works fine here. May be save it first, then open.

Bromic, what do you intend to use the UV light for... I can sense some chemical purpose in this!
Do you know the wavelength range it emits at?

Twospoons - 11-7-2005 at 16:46

The pdf was built with Acrobat 6.0, so you'll probably need Reader 6.0

Incidently, Nichia have some truely stunning UV Leds (Nichia) such as the one that puts out 100mW at 360nm. Probably cost an arm and a leg though.

[Edited on 12-7-2005 by Twospoons]

IrC - 11-7-2005 at 16:57

I saved it 2 or 3 times and each time it would not open. I have noticed for weeks now that any att I d/l here has a delay before the file comes in. So I did what I hate to do, fired up IE and used the plug in, which it opened fine, and then I saved the file there. Now clicking on the file opens it in my reader (6) like it should. Going back and using hex edit on the bad file, there are some missing bytes in the file. At least the save from the plug in works fine.

The wavelength for his diodes should be 400 to 410 nanometers.

12AX7 - 11-7-2005 at 17:19

Quote:

Originally posted by IrC
I have noticed for weeks now that any att I d/l here has a delay before the file comes in.

This happens to me 99% of the time too with attachments, it comes in like everything minus 4 kilobytes then just STOPS dead. Loading images I can hit stop and it displays what it's downloaded, which is basically everything but the last row of pixels, but complete files like .PDFs don't like to come in very well. It usually takes 5-10 minutes to finally finish loading?!

Hmm this post should probably be moved to forum matters...

Tim

chemoleo - 12-7-2005 at 11:32

It' s a shame that LEDs don't do less than 360 nm.
Say, at 200-260 nm things should become more interesting, as the most simple bonds (such as carbonyl) start to absorb, making this invaluable for a variety of organic reactions!

The narrow bandwidth of LED's

Lambda - 12-7-2005 at 11:40

A very nice Spectrophotometer can be built with UV-Leds including other LED's due to there mono spectrosisity. They all show quit a narrow bandwidth , which could be guided through a quarz cuvet containing the sample. A stable constant current source would be required here. No expensive filters or special filterd gas discharge light sources will then be needed. For long term stability the LED's may be artificially made older. To do this, just warm them up to maybe ~80 degrees celcius or more and cool them down in you freezer. Do this a couple of times, and you will have antique LED's. Check this out by looking at the max. and min. storage temperature of your LED's, for the above mentioned temp. may be drastically higher. 150 degrees celcius may be used for transistors, chips and resistors etc., but the bottlenek for LED's wil be the lenz material.

Due to the narrow mono spectric emmision of UV-LED's, you may find yourself slightly pissing next to the pot in it's use in chemical synthesis. The mercury filled UV gasdischarge source of a browning lamp may be better to use, for the intensity is very high (100-150 watt lamps) and the spectum broader. If you use lenses with UV , make shoure you get quarz for this purpose. Normal glass absorbes a lot of UV radiation.

Twospoons - 12-7-2005 at 13:44

Quote:

Originally posted by chemoleo
It' s a shame that LEDs don't do less than 360 nm.

You can bet on Nichia working towards ever shorter wavelengths. They seem to be the leaders in this field - they built the first blue diode laser IIRC.

Maybe their violet laser could be frequency doubled to 200nm? Though at this point mercury discharge starts to look simpler!

BromicAcid - 12-7-2005 at 19:38

Well, I just bought some resistors, 470 ohm, 1/4 watt, 5% tolerance. They also had 1/2 watt but they didn't have enough of them, will these 1/4 watt resistors work?

Anyway, I realize regular glass absorbs a lot of UV but I'm not too worried about it. I plan to use my setup to make some carbon tet from tetrachloroethane in a closed system, no HCl made so I can keep everything nice and closed up. The only difficulty is in breaking that C-C bond that is left, it happens readily with strong enough UV radiation but this might not be good enough now that I've had these opinions on the lack of intensity of these UV lights. Anyway, going to go for three LED's in series in parralel with one another with these 470 ohm resistors on each set of three.

IrC - 12-7-2005 at 21:03

Well you got me to playing so I ordered a hundred from the Hong Kong guy. 470 won't work for three in series, just one, and with a 12 volt supply they will not really be running quite bright enough. For three you need about 68 ohms. Just figure that 3 in series is 9 or 10 volts, say 10, minus 12 = 2 volts. I=E/R so I= 2/68 = 29.4 mA, about right. P=E*I, so in the resistor P=2*0.0294 = .0588 watt. Since the resistor is rated 0.25 watt, you are safe. You should always derate 50% if you don't like failures. For your 470 ohm on a single LED you have 12-3=9, so you need to drop 9 volts in the resistor. I=E/R = 9/470= 19.1 mA, not quite enough. This also depends upon your supply voltage which you still have not stated, but 15 volts would be better for the 470 with a single LED. You need to calculate these things out before you buy resistors, and before you can do this you must choose a supply voltage.

Compromising let's say you are going to go with what you have, 470, and guess that 3.3 volts at 29 mA is closer to the proper turned on LED voltage and current. To get 29 mA through 470 ohms: E=IR = .029*470= 13.63 volts, plus 3.3 volts of the LED = 16.93 volts. Rounding to 17 volts gets you there. The power in the resistor is E squared/R = .6 watts so you would need 1 watt resistors here, kinda wasteful. I would go back to the three in series, guess 10 volts, and pick the resistor for a supply of 12 volts, meaning the resistor needs to drop 2 volts at 29 mA, giving a 68.9 ohm resistor with a power dissipation of .058 watt. A quarter watt resistor with a standard value of 68 ohms is fine. If you would decide upon a supply voltage we can give you exact calculations. 470 ohms usually works great for the visible 4,000 MCD LED's but to get the UV you want you really need to get them close to 25 to 29 milliamps.

Building a large array like you want is much work, so do not begin until you have calculated everything out and are firmly set in all parameters.

PS: U2U me a mailing address and I will mail you 40 of the half watt 68 ohm resistors for free. I have more than I will ever use in my life. I do not think it would cost me more than a buck to get them out to you and they likely cost me 2 cents each to buy, so call it my donation to making your UV stuff work. 40 would do 120 LEDS in strings of 3 and your address will not get out to anyone.

[Edited on 13-7-2005 by IrC]

BromicAcid - 16-7-2005 at 14:06

Thanks for the offer but I have quite a bit of store credit at radio shack that I have to burn up in a somewhat short period of time.

My problem so far with this thread is that I've been telling everyone what kind of LED lights I have and how I want to set them up and I want everyone to tell me what kind of power source and resistors I need. The problem with that (that I'm just realizing is) is that I am giving an equation with two variables and it's hard to solve for both of them with only one equation. So, let me revise this, what would be the minimum requirements for a power supply to connect these bulbs in sets of three, fifty of them.

I will try to get something close to the supply specified here and once obtined I will post what I got and it should be a simple matter of myself or someone else using the formulas so graciously supplied in order to determine the resistance necessary in the system.

My problem as of now is that I know next to nothing as to how to determine the specifications for the powersupply and such. I mean, do I figure out the number of amps that I need by multiplying the amps required for each bulb by 50 to get the total amps? Does this chance since I want to do series of three in parrallel with one another? The voltage should be what? These are 3.4 - 3.6 V bulbs so should it just fall in this range, could it be more, less, does the voltage I get affect the amps I should have? I know none of this, I can try to learn it but so far whenever I learn an aspect of it what I know in that one aspect throws everything else off....

12AX7 - 16-7-2005 at 15:25

Quote:

Originally posted by BromicAcidI mean, do I figure out the number of amps that I need by multiplying the amps required for each bulb by 50 to get the total amps? Does this chance since I want to do series of three in parrallel with one another? The voltage should be what? These are 3.4 - 3.6 V bulbs so should it just fall in this range, could it be more, less, does the voltage I get affect the amps I should have? I know none of this, I can try to learn it but so far whenever I learn an aspect of it what I know in that one aspect throws everything else off....

Bromic:

LEDs are constant-voltage devices. Apply a current to them and a certain voltage (within a relatively small range, hence "constant voltage"

will be developed on the terminals. As voltage rises, current rises exponentially (on the current vs. voltage curve) until it explodes, hence a constant voltage power supply is a big no-no.

The reason is stabiity. Say voltage rises by one volt, from 3V to 4V. Current rises from say, 20mA, to 20mA * 10^(4/3) = 430mA! (That sounds possibly a bit high, I don't know what the actual base is. e gives too low a result. Oh well, I'm no semiconductor physicist.) This raises the junction dissipation from a comfortable 3 * 0.02 = 60mW to a whopping 4 * 0.43 = 1.72W, easily leading to smoke buildup inside (then bursting and going outside) the little epoxy case. Remember, the magic smoke has to stay inside for it to work.

Since by far the easiest power supply to build is constant-voltage, we have to use something to interface them safely. A constant current source (or sink) is perfect, but a simple resistor works as well.

Since an LED is essentially a CV element, and a power supply is also CV, a resistor placed between the two passes a current proportional to the difference in voltages, no matter the current level: i.e., I = (Vcc - Vf)/R (derived from I = V/R, ohm's law). If you connect LEDs in series, you can subtract more of the PS voltage. Too many in series and no current will flow at all because Vf(tot) > Vcc! (Yes, in reality some will, but not much because you're on the short side of the exponential curve.)

So let's say you have a 12V supply and a bunch of 3.6V LEDs. Three can be connected in series for a total 10.8V, but four cannot (for a total 14.4V). Two would be even more comfortable, as with three, it takes a change of merely +10%/-5% to change the current by a factor of two! The AC line is not controlled quite this tightly, though YMMV.

Anyways, the resistor is R = V/I = [Vcc - Vf(tot)] / If, where If is the rated forward current of the LEDs.

Do NOT parallel LEDs, as CV sources do not parallel well. You want series so they all have the same current (the same current going into a wire has to come out, right?).
Else what will happen is, inevitably, one LED will have a lower forward voltage, thus pulling down the voltage on the other LEDs paralleled to it. If the resistor is sized so that all LEDs in parallel get the rated current, most of it will go to just one or two, putting those well over limits and making a micro-cheronobyl. The other reason is, LEDs, indeed most - if not all (I don't know) semiconductors - have a negative temperature coefficient. Forward voltage DROPS as temperature increases. The lower voltage causes one to hog current, causing it to dissipate more power, which raises its temperature, which ultimately explodes it!

So to connect a whole whack of LEDs, you have to use high voltage (something greater than Vf, that is) and many resistors (one resistor per series string). If you want to run 100 x 3.6V LEDs in series, you'll need more than 360V, however twenty strings of five each will run something more than 18V (25VDC or so would be good, with one 220 ohm resistor per string). Current adds in parallel, so if each string takes 30mA, you'll consume 0.03 * 20 = 0.6 ampere.

I don't know how much of this is patronizing but I'm too lazy to read the entire thread and see what I'm repeating. It's summarized in one at any rate.

Tim

IrC - 16-7-2005 at 17:44

I think he just wants to know what to buy rather than the theory behind it all. So here goes. Yes, multiply the number of strings by 29 mA: (50/3) (and round to 17) which gives 51 LEDS, with 17 strings. 17 times 29 mA is about a half amp, so derate this and buy a one amp 12 volt wall wart. Wire 3 leds in series with a 68 ohm, half watt resistor. Do this 17 times. Put all 17 strings in parallel with the wall warts 12 volts. Mount on a nifty twospoons board and sink with lead length as he mentioned. Plug it in and glow away. You need 17 resistors, each 68 ohm half watt, 51 of your UV LEDS, a one amp 12 volt wall wart, and whatever circuit board or mounting of some kind. I emailed the MFG for a graph of MCD versus nanometers and MCD versus current but have not heard back yet.

Archimede - 16-7-2005 at 23:10

The best way to do it should be with one resistor per led. If you group them in mixed serie/paralled setups, and one burn the others get a higher current.
The leds dont have much internal resistance , thats why they burn if you dont limit the current with a resistor.
In theory you could put em all in parallel with one big resistor but I would not recommend that. And if you put em all in series and one burn, the all thing stop working.

[Edited on 17-7-2005 by Archimede]

IrC - 16-7-2005 at 23:48

Ya think? Like the very first answer to the question back on page 1 of this thread. Does anyone actually read these threads?

Twospoons - 17-7-2005 at 16:14

One thing about the 12V 1 Amp wallwart - don't buy a cheap one! They're nasty, unregulated things, subject to line voltage variation etc.
I know they're more expensive, but get a properly regulated one - the guys at rat shack should be able to help you (if they're not just clueless store clerks). The output voltage accuracy from the regulated supplies is usually better than +/- 5%, they're thermally protected, and will usually survive all kinds of abuse without the smoke leaking out.

[Edited on 18-7-2005 by Twospoons]

IrC - 17-7-2005 at 19:31

I think the little regulated 4 amp supply rat shack carries would be better, then he could also use it for other stuff down the road. As to your comment "(if they're not just clueless store clerks)", this one he needs to worry about. I lived near one that had a Pearl Optical next door. Going in there and looking around, the guy came up trying to be overly helpful like they usually are there. He asked what I was looking for and I told him "contact cleaner". He stated "I'm sorry, that is the store next door". Getting pissed that I started ignoring him while I walked on around looking for the can of lubricating contact cleaner he came up and started arguing with me that they did not have contact cleaner. He must have had some other thing going on that day as he was getting real angry with me until I found it, picked it up and showed him the words clearly marked on the can "contact cleaner". So when in radio shack it is helpful to already know what you are doing and totally ignore the help..

Newnes - Power Supply Cookbook (2001)

Lambda - 17-7-2005 at 20:25

I just had to throw this one in here, but IrC may turn out to be right afterall. Will anyone actually read this book ?.

Newnes - Power Supply Cookbook (2001) has 278 pages of crystal clear PDF-scans and is 3.76 mb in size.

For those who will read this book:
http://rapidshare.de/files/3123636/Newnes_-_Power_Supply_Coo...

[Edited on 18-7-2005 by Lambda]

Archimede - 18-7-2005 at 13:39

Quote:

Originally posted by IrC
Ya think? Like the very first answer to the question back on page 1 of this thread. Does anyone actually read these threads?

Yes I was trying to point out that this is the best way to do it and why. After reading 2 more pages that shifted the project (IMO) in the wrong direction.
Sorry, next time I will quote the original poster of this idea.

[Edited on 18-7-2005 by Archimede]

BromicAcid - 22-7-2005 at 19:37

Hard to say this but I ended up ignoring everyone....

I wired 10 LED's in parrallel and hooked them to a variable power supply a friend of mine let me borrow and cranked it up into the 3 - 4 V range and viola the little LED's lit up powerfully, very nicely. It disconnected one and the amps put out by the power supply dropped accordingly. So, I scaled up. I took a wood salad bowl and put 75 LED lights into it. Good thing I noticed they have a positive and negative lead before hand.

Anyway, I wired them all in parralel and it looked really neat, now I hate soldering though, I ended up doing about 160 solder spots and messed up a lot. Then I hooked it up to the power supply and ::bink:: :mad:

One of the lights blew out and none of the rest were lighting, strange since they were in parrallel... So I looked at a voltage carefully and the needle was jumping all over the place, sometimes over 20 V. So, I started cutting wires and sectioning it off into smaller sections, and they would light up readily when current was applied until I got to one section. One of the lights would glow orange and none of the other ones would light. Normally when a LED dies it does so quitely, I don't know if this was a factory defective LED or it just chose some strange path to die on, but somehow it was taking all the current I put into the setup and canceling everything out.

So, a few days and frustrations later I had most of it re-wired along with a few blulbs that I burned out by holding the soldering iron on them too long. I kept the leeds long to act as heat sinks like was suggested earlier in the thread. Out of the 75 origional lights 6 died along the way so I have 69 left in the setup. I left it running for 30 minutes without it warming up or any more lights dying. The test run will be tommorow, weather and time permitting.

Attached are some pictures, 1) Inside of the bowl where all the lights point 2) Outside of the bowl with the wiring all in place 3) Picture of the thing turned on, everything else appears dark but it was taken in a lit room.

BTW, sunglasses are a must, UV radiation is of course bad for your eyes but even looking at them for a few seconds makes things fuzzy for a while afterwards.

UV.jpg - 51kB

12AX7 - 22-7-2005 at 19:39

Anyone wanna put down bets on when the whole set is going to go, completely?

Ten dollars says two hours run time.

Tim

BromicAcid - 22-7-2005 at 19:55

Oh, wasn't that a major vote of confidence. I'll have you know I've distilled nitric acid from drain cleaner and fertilizer in glass pop bottles with a propane torch as a heat source with no problem until I found out soda-lime shouldn't be heated, whereupon my next attempt ended in disaster, it's amazing how far ignorance will take you. And now you've ruined it, now I have a nay sayer and my experiment is more likely to fail

Honestly, if I take a light out of the setup the amp output of the supply drops as a result, the voltage remains constant throughout and nothing heats up. I'm not going to be running these things at 100% of the recomended voltage and I have purposely destroyed LED's with this new toy of mine and it takes me purposly bringing the voltage into the dangerous region for these LED's. I have this one experiment to do that should take a few hours and I see no problem in completing it with the contraption that I have made.

12AX7 - 22-7-2005 at 20:41

Well yeah, current drops because you're reducing the number in parallel, Kirchoff says so. That doesn't mean it's good, or for that matter the least bit stable!

Tim

BromicAcid - 23-7-2005 at 17:25

And 6 hours later not a single light had gone out

Ran my apparatus for photochlorinatoin today, it went well, except that towards the end the sulfuric acid was falling into the same area of the hypochlorite and it wasn't reacting so too much pooled there and the stopcock of the addition funnel was opened so it all started reacting at once, blew the stopper out and caused a rain of concentrated sulfuric acid. The cloud of chlorine gassed me more then I had ever been gassed before by that element....

As for how it worked, when I lifted up the chlorination cover the air and lights were the same temperature as the surroundings but the flask was hot, the chlorinated constituents were condensing on the inside of the flask it was so hot. Definately a good reaction pace (being that the heat must have been coming from the exothermic chlorination). Going to fraction the product later considering most of the products have markedly different boiling points.

Attached is a picture of the setup. I made one neat improvement though, the bubbler through the sodium hydroxide solution at the end was actually inserted into a sepretory funnel, into the stem of it. In the end a glass wool plug was inserted. In this way I could close the stop cock to stop the suck back when necessary and in addition there was a lot of water to suck back to fill the inside of the sepretory funnel.

Oh well, picture attached.

Setup.jpg - 110kB

Tacho - 31-7-2005 at 15:06

I agree with 12xy. The leds are likely to fail soon if you have too much current flowing through them. Measure it! It should remain bellow 20mA. Besides if you put too much current through a common green led, they go yellowish and stay that way. Maybe the same will happen to your UV. You may have a different wavelength and not notice it. Long UV?

Twospoons - 31-7-2005 at 17:14

Well, they're your LEDs, so do with them as you please. Just be extra careful not to knock the power supply, or heat the LEDs too much. You have no room for error.

IrC - 31-7-2005 at 21:33

He has some room for error, he isn't using them all so he has some left. Besides, that Hong Kong guy is cheap! I cannot believe what I have paid for those things in the past. Anyway, his statements tell me all I need to know. His experiment seems to have worked ok so he must have enough UV to do the job. Also he said they were not getting hot so the current cannot be too high even if he does not have a meter. Reaction working and leds not hot, it ain't broke don't fix it. I do disagree with the under 20 mA statement though, for these leds they just would not work very well down there.

As a side note, Bromic Acid and Twospoons got me going playing with them and I found out one of the aluminate phosphors I thought had failed actually did not glow at all until I hit it with the UV LEDS, then it glows great in the yellow. I can shine a 3 watt white LED on it and nada, hit it with a dozen UV LEDS which are way dimmer and great yellows come out after I kill the light. Very bright, and interesting but the powder does not seem to glow in the UV other than a slight white similar to a brightner. I think the MgO quenching I was playing with somehow made this batch only respond to UV, and then only glow in the yellow (the patent formula was only blue here until I added a quencher). So next I will stir some up in some glass and see if I can make a cool UV to yellow frequency shifting material. I don't think I'm off topic here as it did involve the UV LEDS.

Organikum - 1-8-2005 at 01:42

An interesting approach, but.......

Actually looking at your setup Bromic I would say that removing the UV-diode mantel and exposing the flask to the sun which shines so nicely in the picture will produce much better results. A simple mirror would boost efficiency.

The actual power of the diodes compared to plain sunlight is neglectable, compared to a simple 150W halogen lamp with the glass-UV-shielding removed it is still small.

Like this:

Remove glass, be careful the lamp is HOT. Worked perfect for me every time.

A dash of red phosphorus makes the UV obsolete - the shit scraped from matchboxes works.

/ORG

Mr. Wizard - 1-8-2005 at 08:36

I was wondering why nobody mentioned the bare halogen light bulb. Those tubes are supposed to be fused silica, so they will transmit UV.

IrC - 1-8-2005 at 10:18

A good idea but they do put out a large amount of heat as well. Robbing a copy machine of the exposure light and matching power supply is one idea, you also need a cooling fan. There also is the thought about duty cycle, 100 percent for a long time might not be a good idea. In any case handle the bulb with gloves and clean it well with some kind of solvent that has no denaturing or other additive that would leave a film. Any finger oils on the quartz envelope will cause dark spots to appear. The light also needs to be run a proper duration so that the halogen cleaning cycle will be completed, to ensure maximum life and light output. Like I mentioned, there will be the issue of a great deal of heat. That was why I mentioned the mercury bulb used in UV erasers earlier, lots of UV with little heat, but another mentioned they do not put out enough overall candlepower. I'm still not so sure about that though, if the LEDS work in the experiment I see no reason the eraser light would not also work.

Organikum - 1-8-2005 at 10:25

Sunlight is the first choice, the heat of the halogen lamp can be hancled with a 12V computer fan. The halogen is to handle with care when addition of compounds to the reaction flask is necessary, but me at least survived this too.

vulture - 1-8-2005 at 11:33

Organikum, do you have any indication as to which are the shortest wavelengths a halogen bulb will emit? Because 400nm for those LEDs isn't really UV, IMHO.

Twospoons - 1-8-2005 at 16:27

Near as I can work out halogens go down to 250nm, but the output is quite low.
The vast bulk of the radiation is in the IR. Only 1% of the output is in the UV range under 400nm.

Mr. Wizard - 1-8-2005 at 17:56

Even 1% of a 100 watt bulb is a terrific source of UV. The bulbs are dirt cheap and can be run with only convection cooling. If electricity is that expensive where you live, maybe the UV LEDs would make sense. Sunlight is even cheaper, but not as convenient, as many have mentioned.

IrC - 1-8-2005 at 19:00

I figured all along that Bromic just had these LEDS and wanted to find a cool use to play with them. If I were actually going to build the reaction setup mentioned I would use one of those 18 inch 20 watt fluorescent germicidal lights. Much more UV in the short end of the range, little heat and little electricity used, plus it can be on 24 hours a day if needed without worry of any failures or other problems.

Organikum - 1-8-2005 at 23:20

vulture, I don´t know the wavelengths emitted by the halogen lamps, but I know from experience that a 150W halogen without the shielding glass works the same or even better than a dedicated 60W UV lamp.
Don´t forget that shorter wavelengths are filtered out by the glass anyways.

Quartz Mercury Arc

Eclectic - 2-8-2005 at 06:43

http://www.aceglass.com/html/products/support_info/photochem...

Mr. Wizard - 2-8-2005 at 18:03

I have a few of these mercury vapor quartz lamps with fused quartz or silica protective tubes. GPH287T5L rated at 14 watts. They are called non ozone generating germicidal lamps. They have a big test tube surrounding them. Does anyone know how to tell if the big tube surrounding the discharge tube is quartz or just glass? I don't know what the spectrum is either, just that it's for 'germicidal use" and "non ozonating."

Mr. Wizard - 2-8-2005 at 18:15

Picture of lamps in previous post.

UV lamp tube ballast.jpg - 73kB

neutrino - 2-8-2005 at 19:07

Try heating the tubes with a blowtorch. If the glass gets red hot without softening or having any orange sodium vapor come off it, it's quartz.

vulture - 3-8-2005 at 01:37

If it's none ozone generating, it's a safe bet the wavelengths emitted won't drop below 280nm.

Mr. Wizard - 3-8-2005 at 05:55

I think the majority (95%) will be around 254 nm according to this site.
http://www.ultraviolet.com/whatis.htm
The ozone producing band is produced by the Hg vapor at 185 nm and is only a small percentage of the output. What wavelength forms the O3?
I'll test the sleeves I have to see if they are quartz. I'll clean them first to remove any surface contamination then get them red hot and see what happens.

[Edited on 3-8-2005 by Mr. Wizard]

Mr. Wizard - 4-8-2005 at 14:48

It looks like quartz. I cleaned it with tap water, wiped it dry.
1 warm up
2 full heat with propane weed burner
3 red glow, doesn't seem to glow much compared to other things.

DSCF0870.jpg - 21kB

Mr. Wizard - 4-8-2005 at 14:50

I haven't figured out how to attach two jpeg files yet. Here is the tube while under high heat. Notice there is no sodium glow.

DSCF0869.jpg - 22kB

12AX7 - 4-8-2005 at 16:01

Sounds good, anything else would've melted by then. 8)

Twospoons - 10-8-2005 at 17:56

My workmate and I are looking at buying a wafer of 400nm UV LED die from Cree (about 6000 die) and getting them bonded into TO3 size multi-cavity packages. This would put 300 die into a space of 1 square inch, with a UV flux of around 3W. I'm just curious if anyone here would be interested in such a device (we want them for UV glue curing), as we'll probably offer them for sale (couldn't guess a price though!).

12AX7 - 10-8-2005 at 19:37

Wow, hooking up a whole LED slice? That would be AWESOME, UV or otherwise!

d00d...can you imagine a full sheet of Luxeon Stars?

Tim

Twospoons - 10-8-2005 at 19:41

Our first cure lamp consisted of 3, just 3, of the 5W royal blue Luxeons (700mW output each). The result was scary. Those Luxeons are just incredible. Shame they don't do UV.

LED related schematics

Lambda - 13-8-2005 at 18:19

LED related schematics:
http://www.commlinx.com.au/LED.htm

UV Lamp

SAM4CH - 14-6-2006 at 23:50

I have UV lamp (TLD15W/05 PHILIPS 15W T8), and I like to make photochlorination reaction So how can I operate my lamp safty and what about its connections?

neutrino - 15-6-2006 at 02:16

Consult the product manual and / or manufacturer if you want to know how to run the lamp. I'm assuming that's what you mean by 'connections'.

Twospoons - 13-9-2006 at 21:53

On the subject of LEDs as UV sources, we've just acquired some excellent ones from Cree. In their Xlamp package, they put out 200mW @ 400nm , with a supply of 350mA @3.5V. With a solid heatsink I've pushed them to 800mA (

), almost doubling the light output. They cost US$9 each.

I'm also trying to get some 380nm, 150mW ones from Hanse.

vulture - 14-9-2006 at 05:27

Quote:

On the subject of LEDs as UV sources, we've just acquired some excellent ones from Cree. In their Xlamp package, they put out 200mW @ 400nm

I see alot of this...but 400nm isn't UV! A quick test in the analytics lab with a difraction grating showed that most people can still see light up to 350nm.

Twospoons - 14-9-2006 at 14:58

400nm is short enough to appear quite dim. And 200mw is uncomfortable to look at. I had 3 running yesterday, giving me 1 watt of 400nm. I have 455nm Leds that put out 700mW, and I can't look at them.
I think the definition of what contstiutes UV is somewhat arbitrary. At the IR end, 850nm is consider well into the IR, yet it is faintly visible in the dark.