Exceeding LED ratings

[Pimpom]

If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta <30 deg C

Current = 60 mA peak at 2Hz, 50% duty cycle
One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

 

[Baron]

Pimpom Inscribed thus:

If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta <30 deg C

9 rows @ 20ma = 180ma Ok there !

Current = 60 mA peak at 2Hz, 50% duty cycle

Current de-rated by a third, times 2 ! (2Hz = twice a second)

One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

Infinite lifetime expected !

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

Very good.

 

[Pimpom]

Pimpom Inscribed thus:


9 rows @ 20ma = 180ma Ok there !


Current de-rated by a third, times 2 ! (2Hz = twice a second)


Infinite lifetime expected !


Very good.

Maybe my description of the arrangement was misleading. The 9x9
array is just the physical layout. The circuit could be all
series or series-parallel, but for the purpose of my question,
the full 60mA is to be pulsed through each LED.

That's a peak of 3 times the assumed max rating, 1.5 times
average. Certainly not good for long-term use especially since
the pulse frequency is low. OTOH, each pulse train lasts for only
4 secs with a few minutes of idle time in between each train. And
the whole thing has to last for only two days.

 

[Pimpom]

Small LEDs have been pushed very hard for medical use, and
although
failures happened, most survived.

However, using LEDs with decent output within spec would be the
more
usual option.

Yes, of course. For brevity, I didn't mention my reason for
wanting to push the LEDs in my opening post. The requirement came
at a time when I'd badly depleted my stock of red LEDs. I live in
a remote place where there's no way to replenish my stock or get
higher-powered types on short notice. I ran a test with the LEDs
operating at 30mA peak but they weren't quite bright enough. I
was wondering if I could get away with running them at twice that
level. A couple blowing during the two-day period is manageable
as I'll be in a position to do field replacements.

 

[Don Lancaster]

If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta<30 deg C

Current = 60 mA peak at 2Hz, 50% duty cycle
One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

LED's are binned into intensity groups to make such use even remotely
possible. Not having a tight enough bin causes all sorts of problems.

--
Many thanks,

Don Lancaster voice phone: (928)428-4073
Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml email: [email protected]

Please visit my GURU's LAIR web site at http://www.tinaja.com

 

[Pimpom]

LED's are binned into intensity groups to make such use even
remotely
possible. Not having a tight enough bin causes all sorts of
problems.

I'm afraid I don't get your point. What does intensity binning
have to do with surviving a higher-than-normal drive current?

 

[DonMack]

"Pimpom" wrote in message
If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta <30 deg C

Current = 60 mA peak at 2Hz, 50% duty cycle
One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

------------------

Without knowing the probability to surviving a single session it's hard to
know the probability of surviving the two days. You are already using about
2x the maximum current(50% duty is about 1.5 continuous). But your session
rate and flash rate is very low so it should be fine. (this is based on the
assumption that the LED's are 20mA)

If this is something serious:

1. Run them all in parallel if a single dead LED or two can be tolerated.
2. Do some testing and graph the results. It shouldn't be too hard to
extrapolate the lifetime of a single LED since that is all that is needed.
Start with some excessive currents to reduce the lifetime significantly(it
should be exponential). You can do this for continuous and pulsed or just
pulsed and scenario based.

For example: If you take a few LED's and pulse them at 100mA continuously
for 2 days or more and they all work fine(and say have the same luminosity)
then you should be good to go. You could start from a lower current to a
higher current for an added safety margin on the same LED test set but this
could result in a skewed failure rate(sooner than later).

In my estimation it will work fine for 2 days or so.

 

[[email protected]]

If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta <30 deg C

Current = 60 mA peak at 2Hz, 50% duty cycle
One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

My guess, and that's all it is, is that your chances of success are fairly
high. You're likely only 2x over the continuous rating and your duty cycle is
pretty low.

 

[Baron]

Pimpom Inscribed thus:

Maybe my description of the arrangement was misleading. The 9x9
array is just the physical layout. The circuit could be all
series or series-parallel, but for the purpose of my question,
the full 60mA is to be pulsed through each LED.

That will teach me not to make assumptions.

Based on the "60ma" through each led, I would reduce the cycle time to
1/3rd of a second. ie 1/3rd on, 2/3rd off. If you do put them in
parallel, there is a risk that one or more will current hog and
deteriorate rapidly with the others having to adsorb a greater amount
of current.

That's a peak of 3 times the assumed max rating, 1.5 times
average. Certainly not good for long-term use especially since
the pulse frequency is low. OTOH, each pulse train lasts for only
4 secs with a few minutes of idle time in between each train. And
the whole thing has to last for only two days.

I've seen led displays run at 50hz with a 3:1 current ratio. But they
were effectively single devices.

 

[John S]

If I were to use a group of LEDs as described below -

LED type = ordinary 5mm red diffused TH. No data but assumed to
be ~20 mA.
Several LEDs from the batch used for a few years as indicators at
a few mAs. None has failed.
Mounting: 81 LEDs evenly distributed (9x9) on 4.5"x4.5" PCB
Environment: Outdoors, LEDs unenclosed, direct free air, shaded
from direct sunlight, Ta<30 deg C

Current = 60 mA peak at 2Hz, 50% duty cycle
One flashing session = 4 sec, 8 flashes
Session frequency irregular, roughly every 2-10 minutes
To be used for two days at ~6 hrs/day

The arrangement is for temporary emergency use, to be
dismantled/discarded after the said two days. What are the odds
of all LEDs surviving to the end? Informed opinions please.

I agree with the general consensus that your are safe.

What is the price of failure?

Cheers,
John S

 

[John S]

Perhaps 1000 deaths on an Indian railway


NT

In that case, don't chance it.

John S

 

[MrTallyman]

I reckon you should be fine, but I'd set them up in such a way that
one dead LED doesnt extinguish another, which is very easy.

It would be more than a little stupid to string them in series for more
than a few reasons. Might even be more stupid to have even suggested it
possible, but then again... we are talking about homo sapiens.

 

[ItchyGato]

Perhaps 1000 deaths on an Indian railway


NT

The Indians made railways? Does a little guy sit on top of the
'engine' with a rug, making smoke signals?

 

[John S]

It would be more than a little stupid to string them in series for more
than a few reasons. Might even be more stupid to have even suggested it
possible, but then again... we are talking about homo sapiens.

Well, not when it comes to YOU! You have not demonstrated that you are
sapient.

OTOH, AlwaysWrong, by putting them in series, the current is the same in
each one. Since you are AlwaysWrong ( [email protected] ), you lose
(as you are so fond of saying).

You are stupid beyond belief! Meaning, you are not sapient.

 

[MrTallyman]

OTOH, AlwaysWrong, by putting them in series, the current is the same in
each one.

Idiot. It also means that a FAIL open is a catastrophic failure, not
that you, being the result of one, would know that term.

 

[MrTallyman]

You are stupid beyond belief!

You obviously mis-read what I wrote.(haha) You are busted being stupid
yet again.

You are stupid, and considering your retarded bloodline, it is quite
believable!

 

[[email protected]]

The Indians made railways? Does a little guy sit on top of the
'engine' with a rug, making smoke signals?

Um, wrong "Indians". ;-)

 

[Pimpom]

I agree with the general consensus that your are safe.

What is the price of failure?

Just a slight embarrassment and inconvenience unless a mass
failure occurs. As I said in one of my posts, I'll be on hand to
replace an occasional blown LEDs.

 

[Ralph Barone]

It would be more than a little stupid to string them in series for more
than a few reasons. Might even be more stupid to have even suggested it
possible, but then again... we are talking about homo sapiens.

I was under the impression that LEDs failed short (at least under
reasonable overload conditions). If you whack them so hard that the
wirebond to the die evaporates, then certainly you'll get an open
circuit, but my Christmas tree lights (wired in series) appear to work
just fine with a couple dead LEDs.

 

[Pimpom]

It would be more than a little stupid to string them in series
for
more than a few reasons. Might even be more stupid to have
even
suggested it possible, but then again... we are talking about
homo
sapiens.

OK, I'll bite. Why is it not _possible_ to use 81 LEDs in series?
We use 230V AC here. Not everyone lives in the US.