Chinese 3W LED review [With pics]

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
Hello all,
I just ordered a lot of these Chinese 3W Cool White high power LEDs straight from the manufacturer. Not only they're cheap, but they're reilable and color consistency is excellent.

Here, have a look at its die (or dies?>:sssh:)
IMG_3614.jpg


On careful observation, you see not one, not two, but a whooping 8 Bond wires on the substrate! I verified this by giving 2.00 V and each bond seems to have its own die!

The benefits are twofold, firstly, these dies are simple and easy to produce (I think these are the same dies with 5mm LEDs). And being in Parallel, I one die fails, other continue to function!:tinfoil:

IMG_3604.jpg


Here's a closer peek using my cam's SuperMacro and some super advanced Macro lens-

IMG_3594.jpg


Sorry, I don't have any scientific equipments to measure the light output and spectrum, but by personal analysis, they're BRIGHT!
atleast as bright as the Seoul p4 LEDs
Imagine that a just less than half the price!!!

One more thing i'm happy about is, very low vF.
They pull about 430mA just at 3.05V!


I hope to bring some beamshots tomorrow, stay tuned!

Update 1>

VF chart-
Voltage---------mA
2.99 370
3.01 460
3.20 690
3.33 700 (Sweet spot, and drives direct on PC SMPS 3.3v rail)
3.55 900

1> LED at very low power. We can see the dies glowing separately-


IMG_3624.jpg


2> At Slightly more power, the glow gets more and more homogeneous-

IMG_3624.jpg


Update 2.... Reserved for beamshots
 
Last edited:

znomit

Enlightened
Joined
Aug 1, 2007
Messages
979
Location
New Zealand
Sorry, I don't have any scientific equipments to measure the light output and spectrum, but by personal analysis, they're BRIGHT!
atleast as bright as the Seoul p4 LEDs
Imagine that a just less than half the price!!!

:whistle:
:popcorn:
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
Let me guess, these are the same ones being sold by Satistronics.

If they are, I've tested them. Good color and decent brightness, though not as bright as an average bin/flux Cree P4. Plus, optics optimized for Cree won't work with these at all.

Next, they have big problems when it comes to current regulated drivers. They need a fixed voltage source (or battery), or they smoke quickly.
 

avion23

Newly Enlightened
Joined
Feb 2, 2008
Messages
1
Next, they have big problems when it comes to current regulated drivers. They need a fixed voltage source (or battery), or they smoke quickly.
What are you talking about? No LED i know needs a voltage source. I highly doubt what you are saying.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
Avion23, are you using a Charles ****ens grammer checker or something?

No LED i know needs a voltage source

I don't know where to start with that one, but I'll leave it alone.

Here's the link to the Satistronic's neutral white 3-watt LEDs. They do look familiar.

I've got several dozen of these in all different colors running right now in various projects, and they all have one thing in common - they all die within a few hours (or minutes) of being placed on a current regulated driver such as a Xitanium, etc. Even if placed in series with other more robust emitters, they die. They run fine when run off a fixed voltage source like a laptop brick.

I could spend a lot of time trying to figure this out, but the fact is they cost less than $2.00 each and that should tell you something.
 

Norm

Retired Administrator
Joined
Jun 13, 2006
Messages
9,512
Location
Australia
Just guessing if all the dice are in parallel if one die dies the current to the remaining dice increases next die dies current goes up again sounds like a chain reaction. One weak die and it takes all the others with it.

If the voltage is constant the current drawn by each die will be the same even if a die dies.
[/GMO] (guess mode off)
N0rm
 

JohnR66

Flashlight Enthusiast
Joined
Aug 1, 2007
Messages
1,052
Location
SW Ohio
I haven't tried these yet, but my experience with multi chip medium power cheap Chinese/ebay LEDs has been poor:

1) Grossly overstate current. My rule of thumb is to multiply the the number of dies by 20ma and derate a little, so this LED probably should not see any more than 150ma.
2) even following #1 above, the LED will still degrade in ouptut within a couple weeks to a month.

Whoopee. So you saved a few $$ but what is gained? A LED that lasts days compared to ones that last years. Ones that are not as efficient and probably have crappy color tints.
 
Last edited:

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
Let me guess, these are the same ones being sold by Satistronics.

If they are, I've tested them. Good color and decent brightness, though not as bright as an average bin/flux Cree P4. Plus, optics optimized for Cree won't work with these at all.

I don't know the satistronics.
I got a batch to 500 straight from the noodle kingdom!

yeah, Color consistency is excellent.
And Atleast 95% as bright as a cree or Seoul.

Rock solid on Direct drive in My car (worst direct drive DC Platform).
 

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
One more thing....

The heatsink is only warm even after hours with these LEDs at 350mA.(It feels 'ouch' on the seoul p4) I don't know that for happy or hell.

Either-
1. LED are very efficient with low vF (less than 3v at 350mA)
2. Thermal path is exteremly poor. :(
 

winny

Flashlight Enthusiast
Joined
Apr 14, 2005
Messages
1,067
Location
Gothenburg, Sweden
Just because they have several bond wires does not mean that they have used several dies under that layer of YAG-phosphor. High power devices with big dies commonly use several bond wires to get even current distribution and to minimize I^2R-losses.
They might have several dies, but your conclusion is wrong.
 

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
Just because they have several bond wires does not mean that they have used several dies under that layer of YAG-phosphor. High power devices with big dies commonly use several bond wires to get even current distribution and to minimize I^2R-losses.
They might have several dies, but your conclusion is wrong.

Macro pics of Dies at very low vF are coming up...
and yeah, we are discussing! :poke:
Not concluding!:D

Winny, can you explain how more bond wires are better than one?
What are the pros and cons? i see they are more in number, but thinner than Luxeon or Seoul p4. I one or two bond wires fail, will the others carry on? or What will happen?
 
Last edited:

biggerrigger

Newly Enlightened
Joined
Jan 6, 2009
Messages
13
Location
Hillsboro Ohio
Avion23, are you using a Charles ****ens grammer checker or something?



I don't know where to start with that one, but I'll leave it alone.

Here's the link to the Satistronic's neutral white 3-watt LEDs. They do look familiar.

I've got several dozen of these in all different colors running right now in various projects, and they all have one thing in common - they all die within a few hours (or minutes) of being placed on a current regulated driver such as a Xitanium, etc. Even if placed in series with other more robust emitters, they die. They run fine when run off a fixed voltage source like a laptop brick.

I could spend a lot of time trying to figure this out, but the fact is they cost less than $2.00 each and that should tell you something.
Have you had any problems in ordering from this company in the past? Is it like ordering from KD or DE takes forever.
 

winny

Flashlight Enthusiast
Joined
Apr 14, 2005
Messages
1,067
Location
Gothenburg, Sweden
Winny, can you explain how more bond wires are better than one?
What are the pros and cons? i see they are more in number, but thinner than Luxeon or Seoul p4. I one or two bond wires fail, will the others carry on? or What will happen?

More bond wires means that the current has several different paths to enter and exit the die with a larger cross section area using all bond wires than just one. If you take a look at wikipedia, although the example is a FET and not a LED, it uses three bond wires for the source and only one for the gate due to the fact that larger currents have are running though source.
Larger total wire cross sectional area is one reason, the other is local hotspotting. This can be super critical in some transistors but perhaps not in LEDs where the additional bond wires actually obstruct the light shining of the die.

Yes, in a single die-multi bond wire LED, if one bond wire fail, the other will make sure the LED keep on shining, although with lower efficiency.
 

JohnR66

Flashlight Enthusiast
Joined
Aug 1, 2007
Messages
1,052
Location
SW Ohio
More bond wires means that the current has several different paths to enter and exit the die with a larger cross section area using all bond wires than just one. If you take a look at wikipedia, although the example is a FET and not a LED, it uses three bond wires for the source and only one for the gate due to the fact that larger currents have are running though source.
Larger total wire cross sectional area is one reason, the other is local hotspotting. This can be super critical in some transistors but perhaps not in LEDs where the additional bond wires actually obstruct the light shining of the die.

Yes, in a single die-multi bond wire LED, if one bond wire fail, the other will make sure the LED keep on shining, although with lower efficiency.

Most power dies I've seen use one or two bond wires and a spreader grid on the die. I have not seen power LEDs with a bunch of bond wires, although you could be right. I sure haven't seen every LED out there.

One way to check is to power the LEDs at low current and observe through a blue filter. Spots of bright areas around the bond wire indicates a die for each, otherwise it could be a single die or the phosphor is too thick to see through.

Chinese budget LEDs are often multichip.
 

Nil Einne

Newly Enlightened
Joined
Apr 15, 2010
Messages
82
Location
New Zealand
Yes I know this is an old thread but I'm wondering how it's going. Still happy with the LED? Or did it die/lose lumen output?
 

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
2 years update@!!

The LEDs are still going VERY well.
However, I am using a constant Voltage source so I can't say about Constant Current one.
The electrical/thermal parameters are still the same. I am dead impressed how such cheap LEDs can last as good as an SSC P4 or Cree.

When Driven at 3V ( i am using a series of 4 at 12v) they draw around 470~500mA which is very sweet indeed.
 

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
7 years update!
The LEDs are still going on very well. They see less use, but use is very harsh in Automotive Heat/Voltage Enviorment. I've added a resistor to Underdrive them somewhat. I cannot see any light output degradation yet!

The tech has moved on a lot and Most lamps are COB now. But I have NO qualms will these small babies. They work very well.
 
Last edited:

Nil Einne

Newly Enlightened
Joined
Apr 15, 2010
Messages
82
Location
New Zealand
7 years update!
The LEDs are still going on very well. They see less use, but use is very harsh in Automotive Heat/Voltage Enviorment. I've added a resistor to Underdrive them somewhat. I cannot see any light output degradation yet!

The tech has moved on a lot and Most lamps are COB now. But I have NO qualms will these small babies. They work very well.

Quite interesting update. Goes to show that even when you're buying (it sounds like) small quantities you can still get good stuff from China. You just have to put up with the weird quarks and sometimes unexpected or undesirable behaviour.
 
Top