Problems with Dr. LED Lights

[colemj]

Interesting. In a past thread on LED lights, I had posted that we bought a bunch of these for $7/ea at a local hardware store and were very happy with them (still are - they have worked perfectly for 3 years now): Amazon.com: Array 1-Watt (10W) T3 Plug-in Base Warm White Indoor LED Flood Light Bulb Item#324349 Model#AAC10G4WW UPC#852580003007: Kitchen & Dining

These are like you describe - designed for a single voltage and controlled with resistors and diodes (maybe a simple linear regulator). They produce no appreciable heat and have no measurable rf interference. They haven't dimmed or lost performance in 3 years. Their efficiency is such that turning on every single one on the boat barely moves the ammeter - 1A for 10 lights.

Conversely, our experience with Sensibulbs, which are switch-mode controlled, was that they got very hot, used a lot of power and had rf issues. And they were too big because of the heat sink needed.

The consensus on that thread was that only switched regulators like Sensibulb should be used on boats because of the voltage swings and efficiencies. It was said that these linear regulator or resistor-controlled lights would dim and burn out quickly.

Mark

[Paul Elliott]

Originally Posted by colemj

[...]
The consensus on that thread was that only switched regulators like Sensibulb should be used on boats because of the voltage swings and efficiencies. It was said that these linear regulator or resistor-controlled lights would dim and burn out quickly.

I don't recall anyone saying that a linear regulator design would have any problems. Perhaps it would be slightly less efficient than a switcher, but it should work fine (the well-regarded Bebi lights used a linear current-regulator).

Even a resistor-controlled bulb can be reasonably bullet-proof, as long as it is designed for the voltage range seen on boats. This will be less efficient, but in many cases this loss is minor.

What I and others have said is that some resistor-controlled bulbs are designed for a very limited voltage range, and will burn out at the high voltages seen during battery charging. This was a long time ago, and perhaps this style of bulb is no longer being manufactured.

[DeepFrz]

Perhaps the LED industry has gained a lot of experience with building lights for cars and trucks and they have improved the tolerance for voltage fluctuations. I used to see quite a few vehicles with some of the LED's in the taillights not working. I haven't noticed that lately, although we are just entering the "dark zone". I wonder if others have observed the same thing?

[Cpt Pat]

Here are spectragraphs of the RFI noise from the red Dr. LED I tested.

On the left, the baseline (background) noise levels (LED powered off) taken inside the screen room. And on the right, the RFI noise level from the Dr. LED.

Placed near your VHF antenna, this noise level from the LED lamp would practically obliterate your VHF reception.

I do wish there were RFI standards applied here in the US to these lamps. Most sailors don't have a screen room and a $20,000 spectrum analyzer handy.

[Rick]

All good LED power supplies use constant current drives to the LEDs. Those drive circuits need to tolerate a variation in input voltage and no engineer would think about using a linear regulator if the input variation is significant.

For example, a 5 Watt LED driver having to tolerate 10 to 16V (or more) input variation would have to dissipate about 3 Watts at the upper end using a linear regulator. Such a loss is intolerable for any claim to high efficiency and, therefore, a switch-mode supply design must be used.

Low noise switch-mode supply design topology is well known and must be implemented at the cost of more dollars and product volume and weight. Therefore, high quality LED products used in marine dc systems are expensive yet worth the cost, in my opinion.

[CarlF]

Mark,

I have the same Array LED and love them. Unfortunately the company seems to be gone. Not one out of about 20 I installed have failed. If I ever sell this boat, I plan to keep the bulbs.

I think there's more to this design than voltage regulation. Each of the 12 tiny LEDs has a small heat sink wire attached. Seems to run cooler than the chunky heat sinks like Sensibulb or the 'no heat sink" of many cheap varieties.

Too bad the company didn't make it (and that I didn't have the foresight to buy a 1000 bulbs at $7 and resell them at $18)

Carl

[Cpt Pat]

Originally Posted by Rick

All good LED power supplies use constant current drives to the LEDs. Those drive circuits need to tolerate a variation in input voltage and no engineer would think about using a linear regulator if the input variation is significant.

For example, a 5 Watt LED driver having to tolerate 10 to 16V (or more) input variation would have to dissipate about 3 Watts at the upper end using a linear regulator. Such a loss is intolerable for any claim to high efficiency and, therefore, a switch-mode supply design must be used.

Low noise switch-mode supply design topology is well known and must be implemented at the cost of more dollars and product volume and weight. Therefore, high quality LED products used in marine dc systems are expensive yet worth the cost, in my opinion.

I agree that high quality LEDs are worth the cost. But without expensive lab equipment, how does one determine the quality? I've got a spectrum analyzer and a screen room to test in (and 40 years of commercial radio communications experience). But the average boater has to test "in situ" by listening on the radio, and the interference will shift in frequency with supply voltage and temperature. They aren't using crystal controlled oscillators in those lamps. Channel 16 may be fine while channel 15 is completely clobbered. Wait 5 minutes and the situation will be reversed.

There are no standards. I'm waiting for the day a manufacturer states their LEDs have passed the FCC Class A test at an RF testing lab. Just about everything else has to pass the FCC tests.

I'm not sure I agree with the voltage range quoted. If your batteries are at 10 volts, you've killed them. They will never again accept a full charge. Time to buy new batteries that don't have permanently sulfated plates. If they are at 16 volts, there's steam coming out of them. And old-fashioned 12 volt incandescent lamps won't last long at 16 volts. Voltage transients can be suppressed with electrostatic discharge (ESD) diodes that don't add RFI.

A reasonable design voltage range is 11.7 to 14.7 volts. With a purely analog design, LEDs won't dim any more than incandescents over that range. And there's no switching noise to wipe out the radios. Are LEDs with switching supplies better? Yes. But only if they are well designed to be quiet. How can one be assured they are quiet? You can't. They aren't tested for RFI.

Boaters need to demand RFI specs., and I mean hard numbers. I'd demand less that -80 dBm at 1 foot from 2 to 170 MHz as a bare minimum. And less if I intended to install the LED at my masthead near my VHF antenna.

Dr. LED told me that their LEDs produce practically no RFI, and the one I tested was an all-band radio transmitter. I've rejected half the LEDs I've tested from other manufacturers. The good ones (Hela Marine, Signal Mate) are very quiet, but the others (some that were more expensive) wiped out even the FM broadcast receiver on my boat. That's a transmitter radiating 100,000 watts only 5 miles away! The problem is: there's no testing and no standards. Which can be dangerous. I don't expect the average boater to recognize that his navigation lights ruin all reception on his VHF radio. That could happen at a critical time when he really needs that radio.

[colemj]

How about that? I looked on their website and it says they bought a company called SeeSmart and transferred everything to them. Then I went on that website and they don't sell any products from Array at all. Must not have been competitive in the G4 style.

Luckily, I picked up several more just because, so I have some spares. Haven't lost one yet, though. These things have been great and inexpensive - better than all the others we tried at 3-8x the price.

Mark

[Lake-Effect]

Originally Posted by Rick

All good LED power supplies use constant current drives to the LEDs. Those drive circuits need to tolerate a variation in input voltage and no engineer would think about using a linear regulator if the input variation is significant.

For example, a 5 Watt LED driver having to tolerate 10 to 16V (or more) input variation would have to dissipate about 3 Watts at the upper end using a linear regulator. Such a loss is intolerable for any claim to high efficiency and, therefore, a switch-mode supply design must be used.

Where would you be using a 5W LED on a boat? That would be the equivalent of a 50+ W incandescent. So even in that application, losing 3W to linear regulation still produces lighting that's far more efficient than incandescent. But you're correct that the switched regulation will be more efficient.

Linear regulation is fine for lower-power LED lighting, especially for the cabin interior; the problems with linear are 1) heat dissipation as you mention, and 2) linear regulators usually require a voltage differential of over 2 volts (in to out) to regulate effectively. Typical linear regulators can easily handle the swing from 10 to 16v, but realistically the expected range will be 12 to 14.5 volts.

For our interior I've either used LED strip lighting or put handmade LED circuits into small fixtures, with dropping resistors. I also add a reversed diode for transient protection. This has been fine for us.

[transmitterdan]

Originally Posted by Cpt Pat

Boaters need to demand RFI specs., and I mean hard numbers. I'd demand less that -80 dBm at 1 foot from 2 to 170 MHz as a bare minimum. And less if I intended to install the LED at my masthead near my VHF antenna.

Exactly right. I've been railing about this for at least 3 years. Practical Sailor took a swing at the problem but their testing was not very scientific.

It's not easy to design a switching regulator for a light that will work in any installation without causing interference to VHF, SSB, WiFi or commercial broadcast radio. So the interference is there but 90% of cruisers are probably blissfully unware their new LED lights are causing any problems to themselves or their neighbors.

In the US the FCC should care about this but either they don't know about it or just don't care. So what can we do about it?

[Rick]

Addressing Cpt Pat and Lake-Effect:
Having designed alternator voltage regulators, battery monitors, inverters, and various industrial electronics (including 40 Watt LED panels for photography) I can assure you that the actual design spec for nominal "12V" marine, RV and automotive applications require a minimum tolerance of 10 to 18 V range. The lower limit is not a "dead" battery value it is the value UNDER LOAD that can easily be 10V or less even though the battery at "rest" may be above 10V. (as a note, lead-acid batteries, which include AGM and gel-cell can very well survive a 10V standing voltage depending upon many factors). Sophisticated charging systems can approach 16 V when using the Amp-hour Law charging algorithm and, therefore, lights must tolerate that value. I have routinely used that algorithm on my own boat with AGM and Gel-cell batteries with long life (15 years and more) with many deep discharges and rapid recharging under high currents).

Regardless, even with a constant dc input to an LED driver a switch-mode regulator is necessary to convert that voltage to a constant-current driver to guarantee meeting the LED manufacturer's spec for light output and lifetime. Yes, it is possible to do that with a linear regulator yet, again, the losses are intolerable if one is minimizing any heat-dissipation requirement which would often meet or exceed the cost of a good switch-mode design.

I contend that we WILL see more products that meet agency specs for noise emission if for no other reason than products sold into the Europe markets must adhere to higher standards than FCC standards here.

Cpt Pat has a very good example of why low noise drivers must be used in an LED masthead light near a VHF antenna. That example is a particularly nasty proximity problem because ANY switch-mode supply will raise the noise floor over a wide frequency range. However, there have been designs since the 70's that have generated what can be understood as being tolerable for RFI in the proximity of receivers, yes, at some cost.

As I write this I am using a 3.2 Watt LED to illuminate this local area for good writing and reading. The light has a 5000 degree Kelvin color temperature and replaces a 20 Watt quartz-halogen light with much improved illumination. I love this technology!

[DeepFrz]

Its good to see you posting again Rick.

[Dockhead]

Originally Posted by svadagio

Cal40John
I cut a Dr LED apart and found only a dropping resister. Do a little math and you will find that does not meet the requirements. It does not limit the swing in voltage. It only reduces the voltage. You can not reduce the voltage enough to light the LED at the low end and limit the current at the top end. It is probable better than nothing but still a short life.
As someone else said a constant current driver is the way to go. I do not know which manf. do this so I went with a voltage reg that holds the voltage at 12 volts. Now I can use anybody's bulb.
Bob

A good way to be sure that an LED G4 unit or fixture is buck regulated is to buy only ones rated for 10-30 volts.

I don't have any choice anyway since I'm on 24v power.

My boat is one giant test bed for different LED units. I've bought from at least a dozen suppliers.

My remaining Suoerbrights have all burned up. Never buy from them again.

The very expensive Sensibulbs, top rated some years ago, continue to disappoint, dim, and bulky. One of them has lost its light output progressively to the point that it is no more than a night light.

The others (from Imtra, Dr LED, Marinebeam, Boatlamps, UltraLEDs, and others I can't remember) are still all going strong.

I have discovered inexpensive downlighting fixtures which replace the horrendously crappy Cantapuli G4 halogen fixtures which were part of the original build. The take MR16 format units which are available in much greater light output than G4. So I've been putting these in in places where I need good light (over the galley stove, over the salon table, etc).

My favorite supplier has become Bedazzled.co.uk, although lately I've even bought a few from Farnell. I think these LED suppliers are generally just a guy who has ordered a few crates of LED units from China and ships them from his garage on a 700% markup. But I have been pleased with Bedazzled - quality stuff, pricing not as rapacious as others, very good advice over the telephone. Their stuff is all fused, buck regulated, and RFI suppressed, and has worked flawlessly, and giving more pleasant quality light than others.

[Cpt Pat]

Originally Posted by transmitterdan

In the US the FCC should care about this but either they don't know about it or just don't care. So what can we do about it?

As in so many other areas, the US has lost its leadership role in stewardship of the electromagnetic spectrum -- which is public property -- in the interest of Laissez-faire, unregulated economics (i.e., "foist any crap you want on the public"). The FCC Commissioners are lawyers and vulture capitalists who are more interested in making money for the federal coffers by auctioning off public spectrum than in regulating responsible use of the public spectrum. Any regulations aimed at reducing RFI from what will soon become ubiquitous LED lamps is apt to produce screams of "jackbooted government thugs interfering with free enterprise."

The Ham community is already up in arms about LEDs in traffic signals wiping out HF and VHF bands used for emergency communications. The FCC has disappointingly failed to pass tough regulations on RFI. Hopefully, if those communications are needed in a disaster, the traffic signals will be out of service.

I predict that, without strong certifications, the lousy quality LEDs being implemented in homes will make HF and VHF bands unusable in urban areas of the US. Regulation may come after the broadcasters and wireless cellular companies start to see the impact. But given their long life spans, once the problem becomes severe, it'll take a very long time for quieter LEDs to be installed.

I'd look for Electromagnetic Compatibility (EMC) certification from the European Union.

[Cpt Pat]

A general description of the RFI issues with LEDs can be seen here: Radio interference from LED lighting | EMC and Regulatory Compliance

Europe is taking the lead on mitigating the problem. Not much help for us here in the US, but if the LED lamp carries the "CE" (Conformité Européenne) mark, it should be quieter.