Results 1 to 10 of 10

Thread: Diode Life vs. Power

  1. #1
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default Diode Life vs. Power

    Aside from the relatively low cost diodes that are frequently used in RGB projectors and that we often drive at several times their rated power, is there any reasonable relationship between % of rating and lifetime? For some pumped lasers the benefit (or cost) of over driving (or under driving) is far from linear and may be worth it. So, what I am looking at is say for a 20,000 hr. MTF diode run @ 110% the lifetime may be expected to drop to 10,000hr?
    Last edited by planters; 07-18-2013 at 18:16.

  2. #2
    Join Date
    May 2009
    Location
    UCSB
    Posts
    715

    Default

    It is more complicated than that, the life of high power diodes is rarely determined by a failure but rather by the rate at which its output power drops.

    Consider a 1w dpss laser with a perfect cavity so you need 1.25w of pump to power it. If you buy a '1.25w' diode, and run it at the rated 1.25w for its entire life, you can keep turning up the current until you get to the maximum that it is rated for (or heck, the maximum that your power supply is rated for, if you can handle the extra heat and a vaporizing diode won't damage your cavity when it finally melts), which if the manufacture did their calculations right should be about the lifetime of the diode.

    Alternatively, you can buy a 1w diode, and immediately start running it at close to or above its max rated current (its your diode, you can run as much current as you want through it) and keep increasing the current from there, however the diode is going to age faster (due to the increased current/optical power compared to its spec) and presumably the point at which you run out of current/it explodes will come sooner. But, on the bright side, because the lower power diode will have a lower threshold current (for similar technologies) the overall heat load should actually be lower (certainly to start) so maybe its worth it to live with the shorter life.

    Yet another plan would be to buy a 2w diode, and run it at half power, and slowly keep turning up the current. Operating like this should give you an nice long life (things like telecom applications do this, where you have to go replace the laser up on top of a power pole or something). You have to deal with a higher heat load (since the 2w diode is going to have a higher threshold current), but short of you killing the diode or a manufacturing defect which causes it to fail (ex, die exposed to water, back reflections, esd, indium migrating into the active region, ions moving around in the semiconductor in ways they shouldn't, etc) it should last a very long time.

    The issue is that the relationship between current, temperature, optical power, environmental conditions (for non-hermetically sealed diodes), etc are different depending on the exact manufacturing process, and even the manufactures do not have a perfect model which says 'if you increase the current 10% the lifetime will decrease by xxx hours', although if they actually do life testing on their diodes they may be able to tell you if you ask them (although these things are usually somewhat of a trade secret).

  3. #3
    Join Date
    Sep 2006
    Location
    Netherlands
    Posts
    1,438

    Default

    In most graphs you can find on optical power vs time, you can see or deduce an exponential relationship between junction-temperature and output power.
    So for each xx degrees of temperature rise, output-power decreases twice as hard.

    What doesn't show in any graph that I know of, is the effects of thermal cycling on life time.
    I feel that what happens in laserprojectors is not particular good for lifetime.
    Modulation, irregular duty-cycles, dropping bias current with stand-by-beamsuppression, etc, will probably wreck more damage in 1000 hours than nominal dc-current @ ambient T, during 10.000 hours.

  4. #4
    Join Date
    Dec 2007
    Location
    Nottingham, UK
    Posts
    2,849

    Default

    [snip ]
    Last edited by danielbriggs; 07-20-2013 at 04:36.
    - There is no such word as "can't" -
    - 60% of the time it works every time -

  5. #5
    Join Date
    Dec 2006
    Location
    Essex, UK
    Posts
    8,647

    Default

    a short chat me and briggs just had on facebook -

    me - that pdf you just posted
    me - could you provide an executive summery?
    me - if you don't know an ex summery is, its shorter than a normal summery


    briggs - Cold shit, at lower currents = longer life shit.


    now that just saved you all reading 7 pages...
    Eat Sleep Lase Repeat

  6. #6
    Join Date
    Dec 2007
    Location
    Nottingham, UK
    Posts
    2,849

    Default

    Con...
    Such a chop.
    There's a bit more to it than that; but I figured if I put it in "Con speak", all would be good "treacle"
    - There is no such word as "can't" -
    - 60% of the time it works every time -

  7. #7
    Join Date
    Dec 2006
    Location
    Essex, UK
    Posts
    8,647

    Default

    piss my shit !
    Eat Sleep Lase Repeat

  8. #8
    Join Date
    Feb 2011
    Location
    New Hampshire
    Posts
    3,513

    Default

    Actually, I enjoyed it until I looked away for a minuet and saw the Two Ronnies giggling in the corner. But, something bart mentioned about cycling stresses is indeed suggested in the paper as a source of accelerated failure. It might be interesting to run two identical diodes side by side at a relatively high power with one CW and the other @ say 50% duty cycle with a nice harsh square wave @ say 10kHz and compare lifetime. If it turns out to be similar or worse for the modulated diode then maybe we should consider after the fact AOM.

  9. #9
    Join Date
    Dec 2007
    Location
    Nottingham, UK
    Posts
    2,849

    Default

    I would suggest modulated diodes have higher MTTF than CW driven ones, as degradation is a thermally driven process; i.e. in the 50% duty cycle example, power is halved to the junction; that is unless, like Bart suggested, degradation from thermally induced strain effects play a greater effect.

    For 652nm high power diodes:


    [SNIP]
    Last edited by danielbriggs; 07-20-2013 at 04:35.
    - There is no such word as "can't" -
    - 60% of the time it works every time -

  10. #10
    Join Date
    Sep 2006
    Location
    Netherlands
    Posts
    1,438

    Default

    Quote Originally Posted by planters View Post
    Actually, I enjoyed it until I looked away for a minuet and saw the Two Ronnies giggling in the corner. But, something bart mentioned about cycling stresses is indeed suggested in the paper as a source of accelerated failure. It might be interesting to run two identical diodes side by side at a relatively high power with one CW and the other @ say 50% duty cycle with a nice harsh square wave @ say 10kHz and compare lifetime. If it turns out to be similar or worse for the modulated diode then maybe we should consider after the fact AOM.
    I would suggest:
    Run #1 @ 100% nominal power @ 0,01 Hz @ 90% duty cycle
    Run #2 @ 100% nominal power @ 100 Hz @ 90% duty cycle
    Hypothetical: #1 will suffer more than #2 because a big temperature trajectory run once causes more strain than a small trajectory run many times.

    I expect that thermal load is the general cause of optical degradation, while thermal cycling (and the dT/dT at which it happens) is the general cause of catastrophic failure.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •