Absolute Maximum Ratings

[Cubyte]

Hi
  Just wanted some advice regarding led's Absolute Maximum ratings.
  looking at the Datasheet and we have a disagreement in the design team
  about the term  ABSOLUTE Maximum Rating.

  I believe it indicates the limits that a device can tolerate, but not operate at and under no circumstances should it exceed.
  and hitting more than 1 limit at a time would be suicidal!!

 or

 does it mean I can run the device permanently at this limit without damage or early failure even allowing tolerance issue which may push the limits further. All the
 limits can be maxed out would also not be an issue.

 I know this is a bit of a daft question but  would apricate your thought's and comments
 https://lumileds.com/wp-content/uploads/files/DS237-luxeon-sunplus-2835-line-datasheet.pdf

[Siwastaja]

Your first interpretation is closer to the truth, at least on average.

But the actual truth is pretty complicated, and sometimes you can reliably run at some Absolute Maximum Ratings without any issues.

There is no generic answer, and this is because manufacturers use the same language to mean different things. Some manufacturers give "recommended" operating conditions. For some, absolute maximums might mean some "reduced lifetime" ratings, but this is not a standard practice. For others, those ratings might mean that non-destruction is guaranteed, but operation according to specs (like current consumption, noise, etc.) is not. This is a moot point, too, because all the other specifications are defined in some specific operating point anyway (like Vcc=5.0V, Ta=25degC).
 
If you are having a fight over this, it's a telltale sign that neither party fully understands the problem. Of course, admitting your limitations in understanding, and erring on the side of caution when uncertain, is your best bet.

[Siwastaja]

And, regarding the LED, I assume someone suggested you can run it at Tj=125degC and abs. max If, at the same time. Did I guess correctly?

Now sometimes manufacturers give a derating table which discusses this. This datasheet fails to have it.

OTOH, the footnote #1 is quite clear: no such derating (drop to lower I already before max Tj) exists. It's enough to derate just to maintain max Tj. If you know what it means and how to do it, it's well possible these LEDs are fine as long as you satisfy abs. max. ratings. But the practical problem becomes, how are you going to measure and feedback Tj? You can't, at least accurately, in-application. So instead, you measure heatsink temperature, and calculate Tj based on some thermal modeling. A lot of source for error here; so you'd aim at maybe Tj_max=100degC, and start derating If already at maybe Tj=90 degC.

Figure 2a is a good hint why you should be derating Tj anyway: for efficiency!

Finally, LEDs are notorious for failing under temperature and/or mechanical stress; bond wires crack and so on. Derating is sensible to do, even if it results in a bit more LED cost.

[wraper]

Depends on what it the rating, and device in question. If the limit is current/power/temperature, in most case device will have a reduced reliability or lifetime. Limits must not be exceeded, it's the absolute maximum, any tolerances must fall below that. Not to say if you operate LED close to the maximum current, efficiency drops quite a lot.

[SeanB]

Means this is the maximum you can run it at, but with a corresponding short life. IE run at max current and max temperature, life is 1000 hours, which is a little over a month, 41 days roughly. Derate and you get longer life. You never want to run at anything approaching this point, only thing is that this is the point where the manufacturer knows they will not fail, or at least under 1% will fail, but exceed this, even briefly, and they know it will be over roughly 5% that will fail.

[Siwastaja]

Means this is the maximum you can run it at, but with a corresponding short life. IE run at max current and max temperature, life is 1000 hours

No, there is no such general 1000-hour rule, or similar. (Except for electrolytic capacitors, and then it is well defined.)

You really need to have reliability data, and manufacturer might have it available if you are a large enough customer. Failing to get that, add a fudge factor "just in case". I know, it's handwaving.

Some part might actually misbehave when ran at abs. max ratings (even if never exceeded), another part might suffer for short life say 1000 hours, yet another is fine for basically unlimited time but just not with guaranteed performance numbers.

Power LEDs are easy though: you don't want to run them at max ratings, for efficiency reasons alone, and difficulty in thermal management. The question is not if you need to derate from abs max ratings, the question is by how much you want to derate. And here, derating Tj_max instead of current is more sensible, because it forces you to go through at least some thermal modelling. It's better to run at 90% rated current while being sure Tj is just 80 degC, than to assume running at 50% rated current is always good, and accidentally run at Tj=120 degC because of insufficient heatsinking.

[emece67]

.

[Cubyte]

Thanks for your comments I've been asked to investigate failure's on these LED's which fail in a little over a month very close to Sean's Comment.
It also contains another led which is rated at 30mA Ta=25degC they are operating it in ambient temperature of 30 degC  (they don't last long).

[Siwastaja]

It also contains another led which is rated at 30mA Ta=25degC they are operating it in ambient temperature of 30 degC  (they don't last long).

Sounds like no one's engineering properly. Your actual ambient temperature and datasheet Ta are irrelevant. What is calculated Tj? I guess no one knows. Even a simple napkin calculation is better than nothing.

[Gyro]

The standard wording used by most semiconductor / IC manufacturers may help...

Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability.

[wraper]

Thanks for your comments I've been asked to investigate failure's on these LED's which fail in a little over a month very close to Sean's Comment.
It also contains another led which is rated at 30mA Ta=25degC they are operating it in ambient temperature of 30 degC  (they don't last long).

Ambient temperature where? Ambient temperature somewhere outside the device is irrelevant, what matters and what the LED is rated for is a temperature around the LED and it probably gets much higher than 30^oC. And even if it does not, designing a device without any margin whatsoever is completely stupid, unless you don't care at all about what will happen in the field.

[Cubyte]

I would have thought the ambient temperature is important since the junction temperature is part of the overall thermal resistance calculation.
My thoughts was the led was be EOS according to the datasheet where 30mA is indicated at 25 degC Figure 1 page 2 since the site location is a steady 30 degC  the LED current should be lower.
https://www.vishay.com/docs/82556/vlmu3100.pdf

[TimFox]

The semiconductor does not know the ambient--only the junction temperature.
Calculating the rise of junction above ambient can be a complex process, as discussed above, especially if the complete unit is not specifically specified for thermal resistance from junction to ambient.

[Siwastaja]

From Ta to Tj through RthJ-A is the trivial part. The problem is, the Ta is this equation is the local Ta around the part.

When you have single parts with no other heat sources, without casing, hanging in free air on lab table, you can assume actual Ta = local Ta.

[KT88]

I recommend to use the highest temperature on the surface of the part as Ta...