Switching Power Supply Failures

So once again I get to replace failed switching power supplies in LED fixtures (first photo).
Fixture: Elation Opti-Tri PAR LED

10 units out of 12 failed after a power disruption at a local city auditorium last week the night before the final performance.

In 2014 12 units failed in 100 watt Elation ELED Fresnel II fixtures and then in 2017 13 units failed following a power line disruption in a university TV studio (second photo)

The power supply in both fixtures was the ERP (brand name) UHD160-1009 which is a small 2” x 4” 36vdc 4.4A switching unit. Looks like the exact same failure, as described by ERP in their “failure analysis report” after I decided to send a couple of units in after the 13 failed.

Hopefully Elation is not using this particular power supply any more, but I wouldn’t bet on it. Interesting to note that when I went to order replacements for the Opti-Tri PARs from LightParts, they sent a 24vdc 6.66A unit after having verified with Elation that it was a proper replacement. Unfortunately, haven’t been able to locate someplace that would do component-level repairs on the failed units (for a reasonable cost that is).

Interestingly, both installs also had Altman LED fixtures (Chalice houselights at the auditorium and SpectraCycs at the TV studio) which were unfazed.

I wonder how often we have these kinds of failures to look forward to as LED fixtures become the norm?

We've seen quite a few of that same failure on those exact same fixtures locally. Almost universally in places that have intermittent power failures on a regular basis. I haven't seen it on the bulk of Elation fixtures, but those Tri Pars and Quad Pars seem to all have used that same power supply.

Out of interest, were the fixtures that failed normally left powered up 24/7 or powered down after each performance?

00AVD said:

Out of interest, were the fixtures that failed normally left powered up 24/7 or powered down after each performance?

Click to expand...

In both cases the fixtures were left powered up 24/7, which of course is not ideal but several factors were involved in that. And also the ETC ColorSource or Foundry relay systems hadn't been invented when these were installed!

That's interesting. The fault in the SMPS was probably a minor one then. Most of them have a small capacitor on the primary side that is responsible for getting them started. Over time this degrades and become ineffective, but is not noticed until it is required at the next start up. Think of it being like the starter motor on your car failing while you are driving. You don't notice until the next time you switch off and restart. So the upshot is that these power supplies may be able to be repaired with only a few minutes work.

00AVD said:

That's interesting. The fault in the SMPS was probably a minor one then. Most of them have a small capacitor on the primary side that is responsible for getting them started. Over time this degrades and become ineffective, but is not noticed until it is required at the next start up. Think of it being like the starter motor on your car failing while you are driving. You don't notice until the next time you switch off and restart. So the upshot is that these power supplies may be able to be repaired with only a few minutes work.

Click to expand...

Here's the fault description from ERP's "Failure Analysis Report":
Severe power line disruption on the power supply caused the bias Zener diode Z4 to get damaged and became leaky. With this, Vcc of the PFC controller IC2 was unable to reach the turn-on threshold voltage and hence, was unable to turn on. The startup circuit continued to try to turn on IC2 and eventually damaged startup resistor R74. Power supply returned to normal operation after replacing the failed components.

Failed Components: Z4, R74

Yes should be an easy fix for someone but probably not me as the zener is surface mount and not sure I can even discern the value. May give it a go sometime.... if it's already broken what can possibly go wrong?

I've not seen that failure mode before on an SMPS so good to keep it in the back on my mind.

microstar said:

Yes should be an easy fix for someone but probably not me as the zener is surface mount and not sure I can even discern the value. May give it a go sometime.... if it's already broken what can possibly go wrong?

Click to expand...

You'd need an bill of materials or schematic to determine the value, but I suspect someone with a hot air rework station and Digikey could easily fix it.

Let this be a lesson to other users: switch your units off with an air gap (relay / switch / breaker / pull the plug) when not in use.

Let this be a lesson to engineers: use robust power entry protection on your power supplies.


Oh, and it sounds like your location is pretty prone to "power line disruptions" - a lot of lightning there?

Sounds like people using "technology" because it exists. 160 watt 36V poser supply: 1 small transformer, one silicon bridge rectifier, one electrolytic & done.

JonCarter said:

Sounds like people using "technology" because it exists. 160 watt 36V poser supply: 1 small transformer, one silicon bridge rectifier, one electrolytic & done.

Click to expand...

Yep, doubt it would have taken that type of PS out and sure would have simplified things.

JonCarter said:

Sounds like people using "technology" because it exists. 160 watt 36V poser supply: 1 small transformer, one silicon bridge rectifier, one electrolytic & done.

Click to expand...

Ah, but what would it's life be when powered 24/7? Wouldn't it have a much higher minimum draw? How about overall efficiency & cost?

Don't you turn things off when you're not using them? Talk about wasting!

JonCarter said:

Don't you turn things off when you're not using them? Talk about wasting!

Click to expand...

I put everything except the exit signs, a few convenience receptacles, and the architectural processor on a relay; and try to get it programmed to turn off at some late hour. Must save power supplies.

JonCarter said:

Sounds like people using "technology" because it exists. 160 watt 36V poser supply: 1 small transformer, one silicon bridge rectifier, one electrolytic & done.

Click to expand...

160 Watts is a fairly large and expensive transformer. Manufacturers love switch mode supplies because they are small, light weight, and cheap. Unfortunately, switching supplies have made power transformers less common and more expensive.

FMEng said:

Manufacturers love switch mode supplies because they are small, light weight, and cheap.

Click to expand...

...and energy efficient, and have good voltage regulation, and work with any voltage and frequency found globally without modification (so one doesn't need to produce multiple variants of the actual device for various markets--just toss in a locally suitable power cord and instruction manual).