Why does a high voltage luminaire have a more expensive enclosure?

[treez]

Hello,
A lighting designer showed me an offline LED luminaire that he designed.

It comprised 224 LEDs altogether, in 28 paralleled strings of 8_LEDs_in_series.
It comprised four LED PCBs, and each LED PCB contained 56 LEDs (obviously in 7 paralleled strings of 8_LEDs_in_series)

This 224 LED “bank” was supplied by a 230VAC input, 2.8Amp output, Offline Switch Mode Power Supply, and since the LED strings were only eight LEDs long, the maximum output voltage was just 32V.

The designer told me that since the maximum output voltage Of the offline 2.8Amps power supply was just 32V, then the enclosure for this LED luminaire could be made very cheaply (since no high voltages were present that had to be sealed safely away from prying fingers etc.)

If I had been the designer of that luminaire, I would have implemented it as four strings of 56_LEDs_in_series. I would have a high voltage, buck converter to supply each of the four LED strings from the PFC output rail (390V). (so four buck converters in total, one on each of the four LED PCBs.)

However, my solution would have meant having the 390V PFC output voltage on those LED PCBs. –Would this high voltage have meant that the luminaire enclosure would have necessarily been more expensive to make for my design, compared to the low voltage (32V maximum) design?

 

[betwixt]

Re: High voltage luminaire has more expensive enclosure?

When you say "offline" SMPS, are you suggesting it is external to the enclosure? If it's internal, the same voltage will be present anyway on the mains side of the SMPS so I can't see the cost should be any different.

Brian.

 

[treez]

Re: High voltage luminaire has more expensive enclosure?

The offline smps is in its own enclosure, inside the overall enclosure of the lamp. So the actual mains is not exposed.

 

[betwixt]

Re: High voltage luminaire has more expensive enclosure?

I'm no expert in safety regulations but I wouldn't think there would be any additional costs even so. All enclosures are supposed to be 'finger proof', even if their contents run on only 32V. The heat dissipation should be the same regardless of wiring topography. The only difference I can think of that may account for cost increases is heavier grade insulation on the connacting cable but that shouldn't be much.

Brian.

 

[FvM]

Re: High voltage luminaire has more expensive enclosure?

All enclosures are supposed to be 'finger proof', even if their contents run on only 32V.

But for enclosures involving hazardous contact voltages, additional mechanical strength requirements apply, e.g. drop and impact tests have to be passed. For some fancy lamp designs, this might be a problem.

Obviously, this a question of mechanical detail engineering and applicable safety standards which can't be answered by general considerations.

 

[treez]

Re: High voltage luminaire has more expensive enclosure?

Perhaps it is the safety during production test which makes the low voltage version of this lamp so worthwhile?
Its just that I cannot understand what makes the risk of so many strings of parallel leds worthwhile...just because it keeps the voltage low.

 

[betwixt]

Re: High voltage luminaire has more expensive enclosure?

There are reliability factors to take into account as well. If one goes open circuit, 8 LEDs failing to light in 228 is preferable to 56 out of 228. Also consider that in countries where 115V line voltages are used, the PSU design may have to be significantly different which, for reasons of economy may make it less econmical to produce. It may even be that the design has to work on 48V DC as used in millitary and aviation supplies.

Regardless of possible reasons, I would be suprised, given the similarity of designs, if two different enclosures were economical to design and make. I would have thought a 228 LED mounting on four PCBs with sufficient safety isolation (even the LV version would have risk of power line leakage current on it) would be best using a single 'do all' design.

As FvM suggests, there may be factors we don't know about.

Brian.

 

[treez]

Re: High voltage luminaire has more expensive enclosure?

Perhaps i can help if i show the kind of luminaire that this is...

http://www.ledlightsense.co.uk/product-category/led-light-panels/
...its LLS-P0606

...thats not the one, but thats what it looks like......that area is split into 4 pcbs

 

[FvM]

Re: High voltage luminaire has more expensive enclosure?

These lamps aren't portable, so the mechanical strength requirements are limited. Compared to fluorescent lamp that gives acces to high voltage when you break the glass, they are quite safe by design, I think.

 

[treez]

Re: High voltage luminaire has more expensive enclosure?

Since you have said its safe, I am now baffled as to why they have implemented this lamp as 28 strings of eight_LEDs_in_series......I mean surely 4 strings of 56_LEDs_in_series would have been a more robust solution, with a current regulator for each of the four strings...

eg four strings with each strings current controlled by a HV9861A
**broken link removed**

 

[betwixt]

And if one LED goes open circuit....

Brian.

 

[treez]

yes but that's unlikely..the manufacturing process for leds these days is surely top quality

 

[betwixt]

Don't be so sure. I built a light with 12 Avago 'Jade' 3W LEDs, in a PWM dimmable circuit around 18 months ago and it has died. Even at 100% PWM the current was less than half their rating and they are adequately heatsinked. The trouble is it is mounted 8' under water and 150,000 litres have to be drained to reach it. On a second installation in an identical environment I added 3.9V Zener diodes across each LED. In both, one LED has gone open circuit, the one without the Zeners is completely dead, the other is missing just the bad LED and the Zener is keeping the rest of the chain alive.

I'm seeing a pattern of failures, not only the Avago types but from other manufacturers too. The ones that die most are exposed to relatively high ambient UV levels, perhaps it causes premature failure of the junction. I even have samples of 5mm white LEDs here that randomly flash on and off despite their forward current staying almost constant!

Brian.

 

[treez]

That's interesting to hear of Betwixt.
Onsemi make some "LED failure componets" for cheap which can do the job of that zener diode but cheaper and better...I forgot the name of them now, but they are very very cheap and give about 1V voltage when the led they are parallel with goes open.

It'd be interesting to know whether or not LEDs that are properly heatsinked and never_overcurrented ever fail *short* circuit?

 

[betwixt]

I don't think I've ever seen a shorted LED but I have seen completely open circuit ones. Of all the failures I've seen, at least on 3mm and 5mm types, probably the most common is corrosion of the leads. Mind you, some of my applications involve very harsh environements. Where chemical vapour or liquid contact is likely, I now wipe LED wires with silicone grease then shrink 1.2mm sleeving over them, it seems to be the most effective way of keeping nasties out of reach.

A few years ago, I made a one-off special 'port hole' LED light, it was two concentric rings of 2.5mm copper wire with 36 white 5mm LEDs, each with it's own series resistor wired between them. It made a self-standing structure which was placed in a tube, about 15cm diameter by 15mm deep, on a glass sheet base. It also had a 1N4007 diode wired between the rings for reverse voltage protection and a 3A diode in series with the power cable. In other words, it was very well protected against reverse voltages. After 48 hours of testing at LED maximum current (1A @ 12V) the tube was filled with clear resin, allowed to cure then tested for 48 hours again. The resin, being very similar to the plastic used in the LED body, made the body disappear, making it look as though the glowing junction was floating in a disc of glass. This was intentional for the required appearance and so the light gave a wider beam.

So it has a glass front, the LEDs and all the wiring are encased in deep resin, it has 4 days of testing at full power and is electrically bomb proof. After a few months one of the LEDs went out.....Argggghhhh.
It had to be scrapped and another rebuilt from scratch as there was no way it could be dismantled.

Brian.