Maker Pro
Maker Pro

confused over wall wart label

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
I have a wall wart that is labeled 9v 1000ma.
Do they mean it'll provide up to 1000ma at 9v OR is that actual 1000ma draw at 110v.

I want to power an LED with a resistor, that will ask 1000ma.
Basically:
Does that mean I'm drawing all that wall wart is rated for, and can't drive any more LEDs?
 

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
My guess is that it supplies 9V constant voltage, up to one amp of current. And yes, if your LED is drawing 1A then you can only run one with that supply.
 

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
Ok,thanks.
Maybe I need to get a bigger power supply!
 

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I presume you're using white LEDs, right? check the data sheets to see what voltage you should expect at 1A. I'm assuming 3.6V.

With the former supply (9V 1A you can run 2 white LEDs at 1A if you place them in series. You'd need a resistor of about 2 ohms (2.2 ohms is the closest) rated at 5W or more.

You can use the 5V 5A supply, but you'll need a resistor for each LED. Probably 1.8 ohm 5W for each.

Make sure you heatsink the LEDs or they'll have a short life.
 
Last edited:

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
Good point, I had assumed running at an amp it had a higher Vf. A lot of the really high power LEDs I see are actually an array of die in the same package, series and parallel connections. Sometimes the Vf is 6, 9, 10V etc. Depends on the model. :)
 

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
I have a 5v laptop supply

Using Cree XPG LEDs
data sheet:
http://www.cree.com/products/pdf/XLampXP-G.pdf

I want it to be bright
so to run at 1000ma I read spec sheet as saying 3.3 vfwd (correct me please)
so a 1.7 ohm (or 1.8) resistor at about 2watts

I plan to run 4 of these from a 4a amp switching computing supply each with a resistor, and switch them each on and off with a relay in series with the LED.

All make sense?
 

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
Aaah, I've worked with the XPG. Let me say: these things get HOT. ~3.5W of power in that tiny little package... you're going to need a well layed out PCB to mount them and a very well thought out heat sink. Even then, you'll only be able to just keep the junction temp down to decent levels.

What are you mounting these to? Are you utilizing that center pad? When cold, the Vf will be somewhere in the range of 3-4Vdc depending on the Vf bin you get. As they heat up, this will drop. If you're using a constant voltage input, this will raise the current through the LEDs. It doesn't take much... 200mV or so, and you've just shot up to 1.4A. Thermal runaway, etc.

There are many things you need to take into account and manage in order to keep these things alive longer than a month or two. I'm not an engineer, so I can't give you any specific advice. LEDs are getting a bad reputation in real world lighting applications for just these reasons. They are dying prematurely, they are being used in poorly engineered products.

Personally, I'd run these guys in series w/ a 1A constant current supply with a decent voltage window (12-24Vdc.) No resistors necessary.
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
Mitchekh is right. Your bigger problem is going to be thermal design. 3.5W is A LOT lot of power for that tiny of a package. You have to balance two competing requirements. Performance and reliability. If you use a resistor, you will have to choose one or the other. You will not be able to get both. If you want both, you will have to design a constant current supply.

You can design 4 constant current supplies and run from 5V (harder to do electrically, easier thermally) or you can put 4 LEDs in series and do a single constant current supply (easier to do electrically, harder thermally if you do not get exactly the right voltage power supply).

I too would recommend putting 4 LEDs them in series and running from an 18V power supply. Since most laptops use 18V, you may be able to take an old laptop power supply and cut off the plug.

Alternately, if you are willing to live with only 3 LEDs, you can use a 13.8V power supply (sometimes called "12V automotive" supply) . You can get them from Jameco, RadioShack, HSC and lots of other places.

As I had posted earlier, as of last summer, at least one of the big name LED manufacturers was considering restricting sales to "qualified" customers only because they were afraid of getting a bad reputation dues to failures caused by poor thermal designs. This is not poor design by hobbyists like you, but by "professionals" at companies that are in the business of manufacturing and selling products.

---55p
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Personally, I'd run these guys in series w/ a 1A constant current supply with a decent voltage window (12-24Vdc.) No resistors necessary.

I'd agree, except that I'd run them at 700 mA or so.

I'd also use a switchmode power supply designed to deliver a constant current so that my power supply didn't run *really* hot.

There are actually some relatively cheap constant current power supplies out there.

For a single LED from up to 240V I'd consider this or this. I would be aware of the normal Chinese quality control and safety standards too :-o (This probably falls into the category of KIDS, DONT TRY THIS AT HOME)

For a much safer 5V supply, I might be tempted to use something like this and place 4 of the LEDs in parallel, each with their own resistor (around 0.33 ohms) to balance the current.

edit: Sadly, they do not have a switchmode constant current driver suitable for the 19V that 55p suggests (incidentally running diodes in series using a constant current source is the best option. My suggestion for putting them in parallel is not optimal). If you've nod designed and built one of these before, you'd be far better looking at a linear design, say, something using a LM317. A relatively large heatsink might be in order.
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I plan to run 4 of these from a 4a amp switching computing supply each with a resistor, and switch them each on and off with a relay in series with the LED.

All make sense?

Here is the point at which we ask "Why are you doing this?" If you can give us a bigger picture, we may be able to tell you if it's OK.

For example, if this is a device mounted in a computer case, then using a computer power supply is quite reasonable. The 5V rails are usually able to supply lots of current, so that's fine too. I consider relays to be a pretty poor option, I would prefer to use a constant current source (instead of a resistor) because I could probably switch that on and off using far less current than a relay.

I also need to ask you "How will you be heatsinking these LEDs?"

And "Do you really need to run them at their full rated current?"
 

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
Magtech makes supplies that would fit the bill. 1A const current, 12-24V
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
For a single LED from up to 240V I'd consider this or this. I would be aware of the normal Chinese quality control and safety standards too :-o (This probably falls into the category of KIDS, DONT TRY THIS AT HOME)
Forget about quality control, I would worry about design control. Just looking at the pictures, the open chassis power supply is violating the several UL and CE safety requirements. There is no way it is meeting the 110V creepage requirements, not to mention the 220V ones. If the 100V capacitor is getting anywhere near 100V, they are supposed to be subject to creepage and isolation requirements and I do not see them meeting either. Scary. Just scary.

I have my doubts about the enclosed one as well, but they claim to have CE marking. I do know that the NB we use for our CE would never let us get away with putting that kind of a 110V connection.

---55p
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Magtech makes supplies that would fit the bill. 1A const current, 12-24V

If he wants to switch them on and off individually (a requirement added just recently), then that supply is probably not going to fit the bill :-(
 

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
Looking over docb's posting history, it looks like you guys have gone over all of this already with him.

I agree that in order to really help, we need to know the big picture of what exactly he is planning to do.

---55p
 

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
A lot of the history of what we talked about last month doesn't apply, because I an no longer using that whole design. Now it's just relays to switch the power to individual circuits. So far I plan to use 5 circuits, 5 power supplies if need be, and 5 drivers or resistors. I may add more later, up to a total of 8. Not sure what other info you might need.

The plan is for small flashing lights, so I have four relays, controlled by the computer. I already have this part working, so just need to get some LEDs. Each light can have it's own supply, or driver, as they are all independent.

These LEDs cannot be in series or parallel. They are now all separate. One supply or driver per LED.

I did in fact order a few of these
http://www.dealextreme.com/details.dx/sku.13552
that Steve mentioned, to test, so we'll see how that works. That's nice and easy to just switch the 110v with the relays, and be done. They take weeks to get, so I'm still waiting.

I can use whatever supply is best. I don't need to run at 1000ma, but want them to be bright, so why not? I have some 5v 2.2amp switching supplies that I thought I'd try.

I haven't decided on heat sink yet, but since these XPG's are used in flashlights, I imagine they can't get that hot. So far I plan to test with a vice grip clamped to it, and then try a aluminum wall switch plate per LED.

The supply thing will be mounted in whatever case I need, not in the computer. The LEDs outside that, all separate.

The tiny relays I used have 2a contacts, so I see no problem there.

If I was to use these:
http://www.dealextreme.com/details.dx/sku.20330
I'd just use these in place of the resistors, and still use my 5v supplies, yes? Other than some reduction in heat, is there any benefit?
 
Last edited:

Mitchekj

Jan 24, 2010
288
Joined
Jan 24, 2010
Messages
288
What say he ran one C/C supply, 4 LEDs in series. To turn individual ones off, he can use a relay to short a given LED, i.e. one relay connected across one LED each. You could only short three of the four at any time, to avoid shorting the whole supply, of course. The fourth relay could be used to switch the input mains to the supply?

Since the supply is C/C, it will auto adjust the Vout, to keep everything running at 1A (or as Steve said, 700mA, a better idea. Hell, lower the better.) Now, you'd need a supply with a lower end Vout of ~3V or so, but that may be findable/buildable?

I wouldn't go the relay route myself, but if he's already put the time and effort into that portion he's already got working...?
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I did in fact order a few of these
http://www.dealextreme.com/details.dx/sku.13552
that Steve mentioned, to test, so we'll see how that works. That's nice and easy to just switch the 110v with the relays, and be done.

I presume you're qualified to do mains wiring?

I can use whatever supply is best. I don't need to run at 1000ma, but want them to be bright, so why not?

only because heatsinking may be critical -- and by critical I don't mean that you have to have it, but that you have to have it RIGHT. A lot will depend on the duty cycle if the on times are short.

I haven't decided on heat sink yet, but since these XPG's are used in flashlights, I imagine they can't get that hot. So far I plan to test with a vice grip clamped to it, and then try a aluminum wall switch plate per LED.

Do you mean packaged like this? If so, read this. I'm not sure that vicegrips are rated in degC/W. You'll need a heatsink rated at about 15 degC/W to allow you yo use these at 1A in an environment up to about 40C. And I'm not sure how you mount that small package on a heatsink... If they're already mounted on a star, then that makes it easier, but adds a little to the thermal resistance.

If I was to use these:
http://www.dealextreme.com/details.dx/sku.20330
I'd just use these in place of the resistors, and still use my 5v supplies, yes? Other than some reduction in heat, is there any benefit?

No, they won't work if you need to turn the LEDs on and off individually. This would be far more appropriate (1 per LED).
 

docb

Feb 11, 2010
131
Joined
Feb 11, 2010
Messages
131
Whats so bad about using relays? Not very elegant perhaps, but I don't see any real downside.

If I use 110v drivers I need to switch 110v with the relays. If I use my 5 volt supplies with a resistor or a driver, I can switch the 5v out with the relays. (pre-driver, or just the 5v to the LED/resistor.

Yes, the on times will be short, and I can test with increasing on times and see if anything gets warm. The XPG is used in flashlights, so how hot can it possibly get? That's not a very big heatsink, and I will use a bigger one. Yes, they are already on stars.

If I use resistors, I can use 5 watt, and run at just 700ma for testing, and go up if it all seems to be working and staying cool.

>>No, they won't work if you need to turn the LEDs on and off individually. This would far more appropriate (1 per LED).

I don't get it. Why can't I just switch the input to any driver I want with a relay? What am I missing? Im staying with constant current in each case.
 
Last edited:

55pilot

Feb 23, 2010
434
Joined
Feb 23, 2010
Messages
434
If I use 110v drivers I need to switch 110v with the relays.
Working with 110V is a huge step up from working with 5V or 24V. Unlike 5V, 110V can kill you. A major screw-up with 110V can set your house on fire. It is much easier to screw up with 110V than with 5V. That is why switching 110V is highly discouraged unless you REALLY know what you are doing.
The XPG is used in flashlights, so how hot can it possibly get?
A couple of us have already pointed to this tangentially, without explicitly saying this. Now I will say it explicitly because it really needs to be said. Many companies that are designing LEDs into their products are doing a horribly bad job. They are violating a whole bunch of specs and not anticipating things that can and will happen. As a result, the LEDs in their products are getting too hot and failing. So just because you see someone do something in a product you buy, does not mean it is a good idea. If you copy them, your design will suffer the same bad consequences that that product is going to suffer. You can take that badly designed product to the manufacturer and exchange for a new one (if they are still in business). You can not take your design to them and ask for a refund because you copied their poor design.

---55p
 
Last edited:
Top