LED DRIVER FAILING

[trai]

Hi,
I work at a lighting company as a design engineer and recently we had failures of LED drivers onsite. Almost 15 LED drivers failed in 2 months but not all at the same time. At first 5 failed then in 10day another 3 failed and so on. We suspect that there are more drivers to follow.

The driver we used was an AC Electronic Ballast driven at 1050mA.

LED that we used was XP-G Cree with 7 of them in series on a board.

Can anyone help me in determining the cause for driver failure.

All fixtures are indoor rated controlled by occupancy sensor.

Specs for LED driver and LED are attached.

 

[IRstuff]

Seems to me that this is a question for the supplier. Failure is typically due:
> over stress
> random fails
> infant mortality

Has this part worked in the past? Has anything changed? What temperatures do the parts wind up at? Is this only a single site that has the failures? What's the proportion of parts relative to all parts shipped?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[trai]

I have already asked supplier for trouble shooting. I'm waiting to for their response.

This is the first time i tried this ballast. It passed the bench test and they had been functioning perfectly since March. Since Mid June they started failing. Yes it is a single site that has failures. There were 135 fixtures shipped out of which 15 has failed already. They are on 2 different circuits.

On the original order customer requested dimming option based on which we adopted this driver. But they are not using any dimming onsite. They have just capped off the dimming wires.

Could their be an issue with forward voltage of driver and LED?

When we bought these ballasts, the output voltage on spec sheet was 22-29V. But actual driver has output voltage marked as 24-29V, which is the correct one. Based on my calculation, forward voltage for LED is between 22-23V. Could this difference cause LED driver to fail?

 

[IRstuff]

Theoretically, no; but practically, possibly. One rarely attempts to buy capacitors rated at 50V for a 45V application; designers invariably design for margin, picking a 100V part for such applications to allow for life-degradations, temperature, etc. Theoretically, the part you bought should operate to spec, at the value cited, but that's the slippery slope to premature failures, for whatever reason. The fact that they are failing at a single site only, however, suggests that there's some environmental or operational stress being applied uniquely at this site. Either line voltage is too high, or environmental conditions cause the operating temperature to be too high, etc. The datasheet values for life are probably the least reliable pieces of information on the sheet. Nevertheless, if the part is operated at >90C, there could be drastic consequences.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[trai]

Quite a possibility. I totally understand. They had some other fixture in same circuit which were down light LED (By others)which are working fine due to which i ruled out the possibility of issue with site.

Can occupancy sensor cause any issues with driver?

Response from Project Engineer Onsite regarding occupancy sensor.

"The lighting load is landed directly to the relay card and the low voltage components (switches, occ sensors, photo sensors, etc.) are landed on the low voltage channels on the logic board, which mounts in front of the relay cards.

There is a CKT16 per floor and these panels are networked together and control all of the lighting in the building."

 

[mcgyvr]

How do you know the drivers "failed"?
Have you attached a known good driver to the LED strip and does it work?
Have you opened the drivers?
Are they operating in the required ambient?

Your Vf of the string is under the specs for the driver?
Have you tried to add one more LED to the string?
Have you actually measured the Vf of the string?
LED's Vf changes with temperature.

 

[trai]

How do you know the drivers "failed"?
We sent them replacement driver and it worked with the existing LED.

Have you attached a known good driver to the LED strip and does it work?
Yes we did and it worked.

Have you opened the drivers?
No, we didn't open the driver. We sent it back to manufacturer for testing and waiting to hear back from them.

Are they operating in the required ambient?
Yes, as they are mounted Indoor.

Your Vf of the string is under the specs for the driver?
Yes it is. When we bought these ballasts, the output voltage on spec sheet was 22-29V. But actual driver has output voltage marked as 24-29V, which is the correct one. Based on my calculation, forward voltage for LED is between 22-23V.

Have you tried to add one more LED to the string?
I can't. It is designed board with 7 LED's.

Have you actually measured the Vf of the string?
No I haven't. i just went with the spec sheet for XP-G.

LED's Vf changes with temperature.

 

[mcgyvr]

If you are looking for a replacement driver (I sure as heck would be no matter what info the vendor comes back with) I highly recommend looking at Meanwell. I've used a bunch of those and NEVER had a failure.
I really like the LPF series but there are plenty others.

Do you know the stabilized junction temperature of the LEDs in your design?
With that close of a margin in Vf you really can't "calculate" it. And of course the Vf will drop with temperature and datasheets usually post the best case Vf assuming an infinite heatsink.

 

[trai]

Well normally we do use Meanwell, but for this particular design fixture we had to go with High perfection (Magtech) Ballast.

Stabilized Junction Temp for LED is 85°C.

I'm not 100% sure about this but may be you can guide better in this.

On the original order customer requested dimming option based on which we adopted this driver. But they are not using any dimming onsite. They have just capped off the dimming wires. Could that be an issue for driver failure? Although my experience and research shows, that it won't be the cause.

 

[John2025]

Shorting the dimmer leads together is what I would be inclined to do, but that spec sheet doesn't give you much of a clue. As long as you're talking to the supplier, ask them.

 

[MacGyverS2000]

Some comments...

You cannot calculate the forward drop of a string of LEDs, you can only measure... the drop given for a specific LED is a nominal voltage, not guaranteed, and it will actually vary by several tenths of a volt for most LEDs.

Second, a proper LED driver will run via current-mode, not voltage-mode... any variance you see in the measured voltage of each string is due to the varied drop, as mentioned above. The LEDs should receive the proper current based upon the spec. Find out if the dimming is via PWM or current control... PWM will keep the proper light color temperature over all dimming levels, but current control is used in cheaper drivers to save a few pennies during manufacture.

Third, large variances in power conditions at the installed site can kill drivers in short fashion. Make sure the line voltage at the site is not subject to nasty variations that can blow the input side of the switcher.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[trai]

On the original order customer requested dimming option But they are not using any dimming onsite. They have just capped off the dimming wires. Could that be an issue for driver failure?

 

[IRstuff]

"They have just capped off the dimming wires. Could that be an issue for driver failure"

No, the worst that can happen would be random fluctuations in light level because the input is picking up noise, because the input should actually be shorted for proper noise mitigation.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[trai]

Ok IRstuff. Thanks for clarification. I guess i will wait for the ballast manufacturer to provide their test result for failure of ballast.

 

[IRstuff]

What exactly is the failure mode?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[trai]

EVERY 2-3 WEEKS, 3-4 BALLAST FAILS. Since early June 13 ballasts have failed.

 

[itsmoked]

trai; You have a tiger by the tail ..

A couple of thoughts I have are:
1) You say indoor ambient. The drivers could still be seeing a really high ambient mounted up against an uninsulated flat roof ceiling.

2) I had a VFD (motor speed controller), running in my living room. They are notorious for creating electrical disturbances in local wiring. Anyway, no CFL lamps lasted more than 30 days in my entire house. It dawned on me months after the VFD was no longer in use that all my CFL bulbs were now staying alive and I have not had one fail in years.

The facility your drivers are failing in could have some device causing electrical power disturbances that no human can see or detect but that your drivers are seeing and dying over. Often product makers don't test their products under those conditions and things like this happen.

3) If any of the 'replaced' ballasts have failed, you may well have one of the above problems.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[trai]

Thanks itsmoked for your input. I totally understand what you are saying but it will be very hard for me to determine this.The circuit has no heavy fan or motor attached to it. So i ruled out that option before.

Till now, none of the replaced driver has failed. But if this happens for sure we have one of the above problems.

 

[IRstuff]

But, HOW do they fail? What are the symptoms? Is anything functioning at all? Did you take one apart yourself?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[trai]

The LED stopped working. We asked onsite electricians to swap the working fixture driver with the bad one and LED were functioning. Based on this we sent them replacement drivers and now all of the fixtures are working. Bu again we got a complaint that 3 more fixture stopped functioning. So I'm so eager to find the root cause of issue.