Author Topic: Life time of an optocoupler  (Read 4776 times)

0 Members and 1 Guest are viewing this topic.

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Life time of an optocoupler
« on: December 04, 2019, 06:44:35 am »
Hi,
I want to isolate a digital IO, So I used EL817 as the main optocoupler, due to speed requirements, I have drived the input of the opto @ 10mA and the output transistor also is driven @ 10mA, so my question is, Does this affect the life of the input LED? how many hours or years of work should I expect? generally what's the life of an optocoupler?
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline Cliff Matthews

  • Supporter
  • ****
  • Posts: 1911
  • Country: ca
    • General Repair and Support
Re: Life time of an optocoupler
« Reply #1 on: December 04, 2019, 01:50:40 pm »
MTBF is the term you are looking for. Google gives PDF's and about 40K results on "MTBF optocoupler"
« Last Edit: December 04, 2019, 01:53:33 pm by Cliff Matthews »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Life time of an optocoupler
« Reply #2 on: December 04, 2019, 05:06:27 pm »
For whatever reason, optocouplers are not as reliable as other common parts.  They all seem to degrade over a period of decades with a loss of current transfer ratio.  I suspect this has to do with loss of output from the GaAs LED.

 

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: Life time of an optocoupler
« Reply #3 on: December 04, 2019, 05:40:05 pm »
The MTBF number isn't directly useful here.  You aren't concerned with failure, but rather the transfer ratio dropping below your specific threshold.

The light output of the LED decreases over time due to several factors.  The dominant one is high LED current.  To compensate you can increase the LED current.  Which obviously becomes counterproductive.  The alternative is to select parts with a higher CTR and run them as gently as you can, or use a digital isolator.

 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Life time of an optocoupler
« Reply #4 on: December 04, 2019, 05:47:23 pm »
A typical LED's operating time is in the order of thousands of hours; they rarely die suddenly (unless the conditions are very severe), but they will lose brightness over time.

I would expect something similar for optocouplers. Thousands of hours may sound like a lot, but if your device is operating 24/7, it's really not. 10000 hours, for instance, is just a bit more than 1 year of continuous operation. Of course, in practice, even under 24/7 operation, the opto's LED won't be ON at all times, thus the operating life will be a multiple of this.

 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: Life time of an optocoupler
« Reply #5 on: December 05, 2019, 07:02:15 am »
Thanks, I will use it for transferring 24/7 USART data, so If i manage to make their life about 10 years,I would use them,otherwise I should go to digital isolators which are sadly pricey! |O |O
Also if you could suggest sub 0.08$ optocouplers with long life, it can help a lot,

10ma as the LED current is considered high, since it can handle up to 50ma and the recommended is 20ma (EL817 part)?
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline neilhao

  • Regular Contributor
  • *
  • Posts: 58
  • Country: us
    • My Personal Website
Re: Life time of an optocoupler
« Reply #6 on: December 08, 2019, 02:39:49 am »
Thanks, I will use it for transferring 24/7 USART data, so If i manage to make their life about 10 years,I would use them,otherwise I should go to digital isolators which are sadly pricey! |O |O
Also if you could suggest sub 0.08$ optocouplers with long life, it can help a lot,

10ma as the LED current is considered high, since it can handle up to 50ma and the recommended is 20ma (EL817 part)?

I believe the the MTTF can be used to evaluate the optocoupler life time, however the thermal analysis should also be performed.

As DBecker indicated that the light output of the LED decreases over time mainly due to current. This is totally correct, think a step further, in the most of case, the current rises the junction temperature.

See Page 8 of this paper from Philips: http://www.climateaction.org/images/uploads/documents/Philips_Understanding-Power-LED-Lifetime-Analysis.pdf

I used EL817 B and C version, the temperature of EL817 rose 1C to 5C (depends on your layout and industrial design) when the current was 5ma and the environmental temp was 26C.  Thus I said that I believe the temperature of optocoupler can be used to evaluate its aging life. I did the MTTF using HIOKI LR8400 Multichannel Logger and High and low temperature alternating humidity and heat test chamber.
Notes about my technological project: https://uniteng.com
RF Store:
https://shop.uniteng.com
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: Life time of an optocoupler
« Reply #7 on: December 08, 2019, 06:14:56 am »
Thanks neilhao, I managed to lower the input led to a lower level, Now it can work with 2ma! does it mean I should bother about aging?
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15154
  • Country: de
Re: Life time of an optocoupler
« Reply #8 on: December 08, 2019, 09:47:41 am »
Besides reducing the LED current, there is also the way of having the receive side build so that it can tolerate a 50% or maybe even 80% drop in CTR.  I would not lower the LED current so much, as this leaves less room for still working with reduced CTR. The temperature only goes down a little less than linear with the current.

My suggestion would be to make the prototype work with 1-2 mA and than set up the current by a factor of 4 or 5 to account for variations and give the receiver a chance to also work with less light due to aging.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Life time of an optocoupler
« Reply #9 on: December 08, 2019, 04:49:47 pm »
Besides reducing the LED current, there is also the way of having the receive side build so that it can tolerate a 50% or maybe even 80% drop in CTR.  I would not lower the LED current so much, as this leaves less room for still working with reduced CTR.

Agreed. Obviously don't lower the LED current down to a state in which "it just works in the lab". Give it some decent margin.
 

Offline neilhao

  • Regular Contributor
  • *
  • Posts: 58
  • Country: us
    • My Personal Website
Re: Life time of an optocoupler
« Reply #10 on: December 08, 2019, 05:49:44 pm »
Thanks neilhao, I managed to lower the input led to a lower level, Now it can work with 2ma! does it mean I should bother about aging?

Actually you will get the reasonable aging life when the current is about 5mA to 10mA. The concept had been well explained by Kleinstein. Of course, the optocoupler temperature may be rosen by the 5ma current about 1C to 5C, but compares to the affection of other heat sources in your system, the affection to the aging life may be omitted.

According to my experience, the major factor which affects the aging life of optocoupler is the environmental temperature near the optocoupler. This temp can be roughly measured by attaching a thermocouple to the optocoupler case. Low temp = long aging life, however for my design, a ACDC, the major heat source is not the optocoupler, but mosfet. The optocoupler temp is risen up by the heat exchange between the optocoupler and the heat sources. Therefore, try to keep the sufficient  clearance from  the optocouplers and the heat sources. The heat sink may also necessary.

I used EL817 in a ACDC design. The high temperature aging life was more than 11.4 years when the current was 5ma.  You will notice the major heat source is the Switcher IC which contains the Mosfet.
« Last Edit: December 08, 2019, 06:31:37 pm by neilhao »
Notes about my technological project: https://uniteng.com
RF Store:
https://shop.uniteng.com
 
The following users thanked this post: edavid

Online coppice

  • Super Contributor
  • ***
  • Posts: 10034
  • Country: gb
Re: Life time of an optocoupler
« Reply #11 on: December 08, 2019, 09:36:15 pm »
The ageing of LEDs in a pain. I did a couple of things where pulses from a LED along a fibre were used to report measurement results. These designs had to operate for at least 25 years. The only way to keep the LED output reasonably consistent for that long was to use a high lifetime LED, a variable current drive for it, a light sensor, and some MCU software to use the sensor to steadily adapt the LED current over the years.
 
The following users thanked this post: neilhao


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf