Epistar/BridgeLux/noname: A red LED Under The Microscope - Real and Electronically

[Rasser]

Since I have easy access to the LED's in my module and it may become necessary
in the thread "Let's take this shitty led apart and let's design a better one!"
to change the broken ones out, and I wanted to try to measure the current and
just play with it, I turned on my soldering iron and garbed my solder sucker and removed it.

Watching the LED closely I remembered I had bought a cheap USB microscope in Aldi
some years ago, so my thought was lets see what we can see, and to my surprise
it worked out well, the USB micro cam is not worth much with it's 320x240 resolution,
but by putting my phone's camera up to the eyepiece and holding it frustratingly steady
it was possible to take some clear focused shots.

This was my first view ever of a power LED up close like that, quite exciting stuff to do when you stoned.

Taking a peep:

LED Module: With the Red test LED missing:

Dusting off the ALdi USB Microscope literately:

A peek down the scope with 16x4 magnification and using my mobile to take pictures:

Increasing the magnification to 16x10:

Making a Chart:

The next thing I wanted to do was make a chart over the current voltage relationship
in the LED that's in my grow light, I don't know the LED brand, and if it's a noname or
BridgeLux. Anyway here is the numbers.

Setting up the LED test bench:

Using strips to pressure the LED down.	The multimeter's used.

The finished result a chart of current voltage relationship,made manual in Open Office Draw, my life's first !

Note: I have named the LED in the the chart - Epistar - BridgeLux but I have no way of knowing if that's the case.

When using the microscope and taking the measurements I almost wished I had a white coat and
a clipboard to the notes, I'm like fuck I'm doing science now,

My conclusion is: Nice to see the LED under the microscope and see it's not a mess,
the chip it self and the wire mesh looks perfect(from a layman perspective)
and the current-voltage relationship points to a almost linear rise with no surprises along the way.

 

[Chronikool]

What chip is that under the microscope? I know it isnt a Epistar or a Cree. Maybe a Bridgelux or a Epiled?

 

[Rasser]

Don't know, but please tell how you can see it's not a Epistar or Cree.

Browsing BridgeLux and Epistar website it looks like, BridgeLux only makes blue/white LEDs
where Epistar makes blue and reds - but Epistar's red LED the highest power rating is 500mA.

Looking at EpiLED website there are making red LED's running 700mA, so i could be from there.
but no datasheets.

What I would like do is finding LED's that can be used as a replacement(top brand and low),
so if one broke I would have spare parts and maybe if I wanted to change color or lens angles.

The LED package is named P2 on Epiled's web site, don't know if thats a general package description.

Right now I'm looking at the Osram Hyper red voltage current relationship, and it's not much different
than my home made chart, so I can't have messed up to much, and my multimeter's seams to be ok.

I'm looking forward to inspection the white LumiLED I got, and the blue LED's from my unit
under the microscope and make a chart over that.

 

[Chronikool]

A cree chip looks like a window frame...4 squares inside a big square (generally) and a epistar....doesnt look like that...

Cree

Epistar

 

[Chronikool]

It could be a SemiLED...they make some quality chips....

 

[Rasser]

Thanks for the explanation.

I search Google images for 3W red LED's and the ones from
Shenzhen Hanhua opto Co., Ltd looks very close to the ones I got
with it's golder balls in the center.


Images: 1W & 3W

About:

Shenzhen Hanhua opto Co., Ltd was established in 2004. Our factory located on Shenzhen of China,
covers an area of 6000 square meters and has about 200 employees.
We are specializing in producing and designing High Power LED, DIP LED and SMD.
All the LED are made by Auto machine which imported from Singapore ASM.
We can make high power LED form 0.5W -300W in single, double and RGB color form...
Our DIP LED have different size in different shape: 3mm, 5mm, 8mm, 10mm in round, straw hat, bullet, flat, concave forms.

So I think I'm going to remove the Epistar / BridgeLux names from the chart.

 

[Rasser]

Grabbed some more pictures, this time with and without power to the LED's
I felt like writing 'no shit' above the ID text on the red and blue leds

So only the red and far red are looking the same, so maybe the other LED's are from someone else than Shenzhen Hanhua opto Co.

The red LED with power on looked very nice in the microscope but was not possible to capture on the phone.
The far red LED is basically invisible except when running full power.

 

[Rasser]

If I go to the Shenzhen hanhua opto co website and search for 660nm
besides having some juicy 100W plant grow LED's when i select a 3W 660nm LED with ROHS the information says:
chip number 1
chip arrangement 1 chip
chip brand CREE, Epiled, Bridgelux.

So maybe S.H.Opto is just mounting chips from others. They also have a '660nm 1w/3w high power led with golden supplier' what ever that means,
where the brand name says 'HHE' and the picture could look like the red chips i got.

Anyway it's nice to moving from Mercury to Gold.

 

[patrikantonius]

Very interesting thread Rasser, as usual if I may say
What would be excellent is to compare those chips with "branded" LEDs such as Osrams, Philips or Crees. Not only in manufacturing but more importantly in light output.
I think we would be able compare LEDs with a standard digital camera: if the wavelength is similar, doing shots at equal exposure and looking for the "brightest" could give a relevant idea on the differences in efficiency of each one of the different LEDs.
Eventually, doing measurements with real tools such as spectrometers would be even more interesting but this shit surely costs a lot of money.
Of course the different shapes of lenses should be taken into account as well.

I'm really looking forward to do this kind of tests myself, because those Chinese LEDs always remained as mysteries as there are no real datasheet available.

 

[Chronikool]

Good work Rasser.

There are a few other aspects to look at when trying to identify authentic, or branded chips, or indeed just quality of the overall light.

1/ degree of lead /silver in solder that is used to fix the chip to the PCB board.

2/ volume of 24K gold, in the wire that connects the chip to the PCB.(there should always be 2 small, even gold wires protruding from the chip)

3/ quality and thickness of the PCB board. (should be at least 1.5ml think)

4/ weight/design of aluminum in the heat sink

5/ quality of the lens that allows max. transmission of emitted light.

6/ Internal insulation. (Use of thermal gels and workmanship)
7/ The size of the chip itself. (The bigger the chip, the better really. There are various sizes)

There are literally thousands of copy cat makers out there. Various quality. I suspect i will see alot of 'branded' LED grow lights starting to fade quickly in the not too distant future. (just a hunch)

 

[Rasser]

Very interesting thread Rasser, as usual if I may say
What would be excellent is to compare those chips with "branded" LEDs such as Osrams, Philips or Crees. Not only in manufacturing but more importantly in light output.
I think we would be able compare LEDs with a standard digital camera: if the wavelength is similar, doing shots at equal exposure and looking for the "brightest"
could give a relevant idea on the differences in efficiency of each one of the different LEDs.
Eventually, doing measurements with real tools such as spectrometers would be even more interesting but this shit surely costs a lot of money.
Of course the different shapes of lenses should be taken into account as well.
I'm really looking forward to do this kind of tests myself, because those Chinese LEDs always remained as mysteries as there are no real datasheet available.

Thank you. My thoughts exactly on the brand vs. no-name's test's. I thought about using my PAR meter at a fixed distance
maybe inside a tube, since it's a kind of standard instrument.

Searching for cheap spectrometers I found this: Looks very interesting in a kind of fun way.
Public Laboratory: Build a $10 USB visible-light spectrometer

[video=youtube;IA5BTD-aelo]http://www.youtube.com/watch?v=IA5BTD-aelo[/video]


And this:Open Source Cell Phone Spectrophotometer

From this siteLooks very interesting...

 

[Rasser]

We properly all know that a prisms and a CD/DVD is refracting the light to it's individual components, working like this:

Following the idea from the video and using a cut out piece of DVD-R and mounting
it in front of my phones camera to make a quick test, I was able to take these shots.

It's was difficult to take the pictures from all the reflection and the light source it self
so putting it in a black box with small slit makes sense, but for a quick test it looks to be working.

I noted that it was easy the see dark-spots and hot-spots when turning around my self, a kind of sweeping the room fashion.

The black vertical strips is from the black duck tape holding the mylar.

Signal's clean. Range, 2.0 meters. They've found a way in, something I've missed. Hicks: We didn't miss anything. Hudson: 1.7 meters.
Checking the sweep but Something is under the floor, not in the plants, I don't know. Hudson: 1.5 meters. Hicks: Definitely inside the the grow room.
Hudson: 1.2 meters. That's right in side the container. Hicks, Vasquez get back. Vasquez: There right on us. Hicks: Remember, short controlled bursts.
Hudson: 0.9 meters. 0.7 0.6 0.5 That can't be; that's inside the plants. It's reading right man, look! Hicks: 0.001m What the hell?

The very bright strips on image one comes from a fold in the mylar(it's was not on a roll)
and the very bright spot on the 2nd. picture is right behind a running 250W HPS

Maybe someone else needs to sweep there room for infestation, I've seen some wild beast's on some of the pictures.

 

[Rasser]

I'm very impressed
I've just run the program Cell Phone Spectrophotometer and with out reading a single help file
I opened the program loaded my image of the daylight from the window as reference and the HPS as source.
Clicking on the images a pop-up appears and asks for blue or red end of the spectrum so it can make the adjustments.
Press make plot, select nm, and maybe adjust the height of the pixel from 1 to 10 to get a middle value.

First the HPS spectrum:

So we can recognize every peak on this calculated analysis.
Besides the saturation at the great peak, it's almost identical.

And overlaying the to images and stretching the x axis to fit.

Not bad from a cut out piece of DVD and a phone and some free software.

 

[PSUAGRO.]

^^^^^amazing work rasser.............+rep

 

[Chronikool]

You are making that spectrum your bitch Rasser....!

 

[Rasser]

Thank you both.

Just some add on notes to clarify:

The sweeping images was made while standing still, and then rotating between the shots.
When cutting out the DVD I did not split it, to remove the reflective surface as shown in the video
when taken the shots. Doing that makes the plastic with the refracting pattern transparent and it has to be in a dark box to work.

And it's appears that my HTC phone cam is very sensitive in the blue CCD chip.

I was very difficult to take clear shots of the LED grow lamp, properly because of the many light sources,
and by placing a white paper and measuring the reflection from that, the picture was captured.

I think I will later buy some different LED's that I'm sure the brand, bin and peak nm of, to test on.

 

[Gastanker]

Very interesting and that software is a ton of fun. This is what I ended up with for a single 3000k white LED (Ikea lamp) as well as from my cheapo 100w Chinese array.

 

[Rasser]

Nice, I had hoped that someone would make there own spectrum analysis for comparison.
What did you use as reference, and can we please seen some spectra shots

 

[foreverflyhi]

Holy fuck rasser! Rep+ all day! Ur def going to help make RIU led amazing with all those test ur doing, like u mention in your first post, is there some how we can combine our threads? Along with other peoples DIY led?
Seems like what ur doing is pretty much what I'm really trying to understand, I really want this RIU light up and running ASAP! Truth be told I want it ready by the time my orange bud grow is finished! So less then 3 months.
Anyways
Those test your doing are crazy mind botteling!
And whoever said marijuana makes people undetermined? Haha

 

[Gastanker]

No need to use a reference as that's just a feature to compare any two spectrum - simply overlays the two when calculating intensity.

Cool white 100w Array.

Warm White:

Ikea 3w LED: