Dead LED

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Dead LED

Post by Bigglez » Sun Oct 05, 2008 7:17 pm

After a couple of thousand hours an LED (D76) died on
my 64 LED matrix PCB. I wonder why?

The unit was deliberately left running after I completed
the project and it became an 'art piece' in our home.

Recently I though this LED was weak, and now its almost
dark. None of the other LEDs (on this board or others) have
this problem. The matrix driver IC would not affect a
single LED in any way that I can see.

Image

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Sun Oct 05, 2008 11:10 pm

Hi,


The only LEDs i have ever seen blow out are those that have been
driven by too much current. Even the older ones still light up
they just are not as bright and that is a characteristic of many
LEDs, that they go dim not blow out completely.

If it blew out then it could have been defective, or you may wish
to review the peak current rating and the drive peak current.
The peak current in a matrix can not be greater than the peak
current rating either, and the average current can not be
greater than the average current.

Another factor that affects their long life is their proximity to
one another. If they are too close together and the current
causes them to heat up, they heat up neighbors which could
blow them out because of high heat.

So you should check...

1. Average current drive, and LED rating.
2. Peak current drive, and LED rating.
3. Heating effects.

If all this is ok, then you got a bad LED to begin with.

Physical shock could play a part but since it's only one LED
i would not suspect this unless more blew out too.
LEDs vs Bulbs, LEDs are winning.

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 12:28 am

MrAl wrote:So you should check...

1. Average current drive, and LED rating.
2. Peak current drive, and LED rating.
3. Heating effects.
1. SMT 1206 LED: Rated 80mA peak, 30mA DC
2. Driver IC: Rset = 22k 1%, Iset = 23.6mA (peak), 3mA (Avg)
3. Free air at room temp.

Quite possibly this LED was overheated during hand
soldering. Only failure (so far) in 196 LEDs.

Dean Huster
Posts: 1263
Joined: Wed Dec 05, 2001 1:01 am
Location: Harviell, MO (Poplar Bluff area)
Contact:

Post by Dean Huster » Mon Oct 06, 2008 3:57 am

After 20+ years of teaching electronics, I've noticed that LED in the students' digital trainer parts kits would often glow in colors other than the original red -- maybe more orangish -- and if I looked close, sometimes saw some black deep down inside. It was very common for a student to operated the LED directly from the 5v supply with no current-limiting resistor. On the original trainer (The E&L Pencilbox LD-1), the power supply was limited, so didn't do a lot of damage to LEDs (it limited the current somewhat itself!), however, I later modified the boxes with 1A regulators and robust wall-warts and the LEDs did suffer more.

So, as far as bad LEDs, I agree with MrAl, and in my cases, I could see that they had been stresses, so your observation of a possible future failure was not far off.

I remember years back getting "grab bags" of LEDs that were supposedly all the same type. Lining them all up with appropriate (and equal-value) limiting resistors, I noticed that they're light output varied all over the board. I ended up sorting them into 4 or 5 categories just so I wouldn't have weird-looking displays.

Dean
Dean Huster, Electronics Curmudgeon
Contributing Editor emeritus, "Q & A", of the former "Poptronics" magazine (formerly "Popular Electronics" and "Electronics Now" magazines).

R.I.P.

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Mon Oct 06, 2008 4:50 am

Hi again,


Dean:
You reminded me of a bad experiment i did one time with an LED
flashlight and a 4v li-ion cell i had just purchased.
The flashlight was a converted MiniMag with a single LED, and it
was originally powered by three N cells, which totalled 4.5v nominally
and i had built in a current limiting resistor.
Well, when i got the new 4v cell, i stuck it in there without thinking
and i already had a 1.5v cell still in there and no current limit resistor.
Within about a second the white LED turned BLUE ha ha, a very dim,
bleak, blue. Nasty. I could fix it i guess but those light sort of
went out of vogue so i never did fix it.

Anyway...

Another failure mode could exist with a multiplexed array if the
start up algorithm isnt quite right or there is some start up delay.
As the circuit starts up, it is possible for one LED to get banged
very hard with a huge current as the oscillator starts or initialization
routine progresses. The peak current is then the full current that the
driver can deliver assuming the scan is stalled and so the LED gets
a static drive rather than a dynamic pulse drive. The failure
symptom would vary depending on how high the drive can go
during this period, but i would bet it would look like an LED
burns out every so often after a time length much shorter than
its normally predicted life. This would mean we could be on the
lookout for another LED that burns out sometime soon in the
future, or if this isnt the cause then no more LEDs will burn out
anytime soon, but dim out as expected over a long time period.
Geeze, at 3ma im surprised any of them burned out.
Now that i think about it though, if 23ma is the true peak then
it's even a little hard to believe that if the scan stalled 23ma could
take out an LED, unless of course at that time the current could
somehow go higher.
LEDs vs Bulbs, LEDs are winning.

User avatar
dacflyer
Posts: 4472
Joined: Fri Feb 08, 2002 1:01 am
Location: USA / North Carolina / Fayetteville
Contact:

Post by dacflyer » Mon Oct 06, 2008 6:44 am

biglezz >> what sort of display was this.. i cannot resist led art..lol
you mentioned that the display became an art piece for ya.

just currious..

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 9:09 am

MrAl wrote: Another failure mode could exist with a multiplexed array if the
start up algorithm isnt quite right or there is some start up delay.
As the circuit starts up, it is possible for one LED to get banged
very hard with a huge current as the oscillator starts or initialization
routine progresses.
Are you speculating or do you know something that MAXIM
(who designed and fabbed the IC) doesn't?

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 10:09 am

dacflyer wrote:biglezz >> what sort of display was this.. i cannot resist led art..lol
you mentioned that the display became an art piece for ya.

just currious..
Originally the project was to attach 64 LEDs in an array
along with a MAXIM driver IC, a PCB was used to ensure
the LEDs were fairly accurately placed compared with hand
soldering a protoboard (and I'm lazy...)

To get more use out of the PCB I started adding other
features before fabbing the board. I had not used the
MAXIM IC before and was looking for a simple driver
to operate a large array of LEDs.

The PCB operates with a PC and MAXIM's demo software,
and also as a standalone with an AVR uC and some other
hardware features (RTC, rotary encoder, RS-232 I/F).

Once I had reached this point I didn't feel like junking
the project, so I casually attached three of them (Red,
Green, and Yellow LEDs) to a chunk of wood and stuck
them on a display shelf in our family room.

By this time I'd figured out how to command the LEDs,
so it wasn't much more effort to create annimated
geometric patterns and add the ASCII character subset
for numerals and uppercase letters. Viola! Instant art!

While this is a good candidate for YouTube, here's a few PIX:

Image
Image
Image
Image
Image
Image
Image
Image
Image

User avatar
Dave Dixon
Posts: 436
Joined: Thu May 01, 2003 1:01 am
Location: Wichita, KS
Contact:

Post by Dave Dixon » Mon Oct 06, 2008 10:11 am

Not sure about these LED's, but the white LED's we use are ESD sensitive. They can show "latent" failures months after installation/ESD event. Doubt this is the case this time, but thought I'd throw in my 2 pennies worth! Dave

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 10:29 am

Dean Huster wrote:I remember years back getting "grab bags" of LEDs that were supposedly all the same type. Lining them all up with appropriate (and equal-value) limiting resistors, I noticed that they're light output varied all over the board. I ended up sorting them into 4 or 5 categories just so I wouldn't have weird-looking displays.
Branded LEDs from top flight suppliers are "binned" by
brightness. Multi-segment displays are also binned and
the data is stamped on the actual part so that assembly
can be from the same group for uniform displays.

Video screens with LEDs require other means to get
uniform brightness, as binning would be hard to yield,
say, a million pieces all matched. Plus, colour video
LED screens require colour temp adjustment through
software. The ICs that drive LEDs in this case have
registers to hold colour temp and LED match data.
(sometimes the data is lost or corrupted, so its easy
to see a block of brighter, dimmer, or off-colour LEDs).

Drivers on Interstate 580 near 880 (East Bay) often
see a cropped windows logo on the giant LED display
'cos the PC driving it crashes...

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 10:31 am

Dave Dixon wrote:Not sure about these LED's, but the white LED's we use are ESD sensitive. They can show "latent" failures months after installation/ESD event. Doubt this is the case this time, but thought I'd throw in my 2 pennies worth! Dave
Aha! That could be it. When I have time I will replace
the dead LED. Any suggestions for post mortem testing?

User avatar
Dave Dixon
Posts: 436
Joined: Thu May 01, 2003 1:01 am
Location: Wichita, KS
Contact:

Post by Dave Dixon » Mon Oct 06, 2008 10:53 am

Surprisingly enough... we just tossed suspected failures in the trash. If you knew the guys I work with - - That is quite unusual. They usually will take any excuse to spend days on end dissecting/analyzing/"learning" from these kind of things !!!
Maybe someone else has some ideas on what to look for. Good luck!
Dave

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 11:32 am

dacflyer wrote:biglezz >> what sort of display was this.. i cannot resist led art..lol
you mentioned that the display became an art piece for ya.

just currious..
I have a love-hate relationship with LEDs, I'd rather
work in hot glass and neon...

I was inspired by the lightshow used by Radiohead.
(My wife is a fan - I'm not, she dragged me to a show).

Here's a POV video (for what it's worth) Turn the sound OFF!

Here's a POV snapshot

Also, here's a blog (scroll down to "LEDs lighting the way")
from the band's website about their innovative LED
stage lighting project.

Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Post by Bigglez » Mon Oct 06, 2008 11:41 am

dacflyer wrote:biglezz >> i cannot resist led art..lol
Here you go:
BB4. BB7, Satellites - www.i-pix.uk.com
Versa Tubes - www.elementlabs.com
Colourblasts - www.pulsarlight.com
iwblasts, iflex - www.architainment.co.uk
Video Screens - www.nocturneproductions.com
Custom Deployement System - www.specialz.co.uk
Lighting Control and Rigging - www.negearth.co.uk

User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Post by MrAl » Mon Oct 06, 2008 12:28 pm

Hello again,

Bigglez wrote: Drivers on Interstate 580 near 880 (East Bay) often
see a cropped windows logo on the giant LED display
'cos the PC driving it crashes...
No doubt the PC is running Windows (ha ha).


I dont have any specific info on that chip, i am just throwing
some ideas out there on how an LED in a matrix can fail.
Since you said the peak current is limited to 23ma though
i cant see how it could get a peak that was too high, unless
there was some loophole where the current could go higher
as when a single resistor 'programs' the entire array current.

Another possibility is that the reverse voltage when the LED
is 'off' is too high. This is something that could easily be
checked with a scope. I've seen kill time delays after high
reverse voltages in LEDs already, but then i would think
more than one LED would go out if this was the case too.
LEDs vs Bulbs, LEDs are winning.

Post Reply

Who is online

Users browsing this forum: No registered users and 39 guests