RFI from Cree LED bulbs?

[Eric Salop]

I bought some unbranded LED bulbs and gave some to our son to use in his kitchen. Although low lost, they were brighter than the halogens they replaced and a lot more efficient.

We visited him the following week and I noticed that the light output had significantly reduced on 2 bulbs.
Three days later he rang me to say one had caught fire. Fortunately he was in the room at the time to see the flames and turned everything off before any serious damage.

Lesson learned - we'll stick to LEDs from companies who have a reputation to keep up, even if that does mean the products cost a little more.

 

[k0ua]

If there's RFI, it will most often be from the LED's driver circuitry, not from the LED itself. Since a LED is basically a diode that's forward biased, there has to be some way to limit current so the diode doesn't overheat (thermal runaway) and die. In the "old" days, a simple current limiting resistor was all that was needed as LEDs were low power and drew maybe .020 to .030 amps (20 to 30 milliamps).

With high power LEDs, drawing several amps, a current limiting resister would have to be physically large to dissipate heat and would waste power. Thus high power LEDs will often use a "LED driver"circuit, built into the housing (or sometimes external when sourcing parts, driver, LED, and heat sink, separately on eBay). Depending on source AC or DC, either an AC to DC driver or a DC to DC driver is used. Sometimes, if source voltage is out of range, a boost or buck boost driver is used (this allows say a 6 volt source to drive a 12 volt LED).

LED drivers use a high frequency DC switching circuit as higher frequencies allow smaller (cheaper) circuit components. DC input is switched to form a square wave and this is then fed through a toroidal transformer to step output voltage up or down as needed, The AC is then rectified to DC and a regulated voltage and current output. Unfortunately, the fast rise and fall time of a square wave is rich in harmonics and tends to radiate RFI. The interference problem can often be mitigated with good circuit design, shielding, or distance.

Excellent explanation of the LED driver circuitry. It is basically the same thing as the modern switching power supply that we find in all of our modern electronic devices. Instead of the traditional linear power supply, most manufactures have went to the "switcher". No heavy laminar iron core transformer of course with its larger size and weight and cost, just switching components and a small torroidal transformer as described above to "transform" the AC voltage to a much lower DC voltage and better regulation too, than the simple transformer, bridge rectifier and capacitor filter. BUT these early "switchers" were often "nasty" RFI wise. Often rich in harmonics up into the VHF spectrum. They can be pretty clean when manufactured with proper filtering, but in some simple designs this extra money in not spent.

I had one of these supplys on a simple router/access point in my home network, and never had any trouble with it until one day I sat down to do some operating on my amateur equipment and was greeted by a tremendous noise signal from 300 Khz to 60 Mhz. Wall to wall noise. I started looking for the culprit and started flipping of circuit breakers in the house. Found it pretty quick and replaced it with another supply to solve the problem. I suppose some filter components had failed inside of this sealed unit, while its prime function of supply the quoted DC voltage was still working.

I was amazed by the amount of RF coming out of this tiny module and I kept it as a curiosity. I have often thought by sticking a proper resonant antenna on its output leads and keying the AC input with an old Morse Code key, a person could use it for an emergency CW (morse code) broadband transmitter. It basically covers all frequency's at once! :shocked:

But so far all of the Cree LED bulbs I have tested are "clean and green":thumbsup:

 

[k0ua]

I bought some unbranded LED bulbs and gave some to our son to use in his kitchen. Although low lost, they were brighter than the halogens they replaced and a lot more efficient.

We visited him the following week and I noticed that the light output had significantly reduced on 2 bulbs.
Three days later he rang me to say one had caught fire. Fortunately he was in the room at the time to see the flames and turned everything off before any serious damage.

Lesson learned - we'll stick to LEDs from companies who have a reputation to keep up, even if that does mean the products cost a little more.

Dang, that could be very very bad! Yeah I think I will stick to the Cree brand. So far so good.

 

[Eric Salop]

:confused2:Wonder how many of our other members are like me and have almost no idea what you're talking about.:laughing: But it's undoubtedly good information for some.

Sometimes it's the language engineers use to give names to things they work with regularly that takes a bit of getting used to.
I'll try and explain some of this in a way I hope anyone interested can understand.


The main thing with LEDs is that we have to limit the current they take, or they go pop.

Older, not very bright LEDs, have the current limited by what engineers call "Linear" circuits, such as a simple series resistor.
With linear circuits, the current stays steady all the time. A steady current does not produce any radio frequency interference (RFI) .

New, brighter LEDs need more current.
A linear circuit, like a resistor, would still work fine to limit the current, however it would have to be physically much bigger to handle the extra amps and it would get very hot.

Instead they use a "switching" circuit, which is far more efficient.
An electronic switch turns on the current, let's it build up, then before it get's too high, switches if off again.
After a short pause it turns back on again, then off, on, off, on, off, on, off ...
All this on-off-on-off switching is very fast, many thousands of times every second, so the human eye is not aware that anything is happening.


Have you ever thrown a power switch on an appliance or lighting circuit and noticed that occasionally this causes a flicker on the tv screen or a click on the radio ?
Throwing the switch changes the current. The more rapid the change in current, the greater the RFI it produces.

Imagaine now how much potential there is for RFI from an LED lamp that is internally switching on and off thousands of times every second.
Cheap and/or poorley designed LED circuits can generate an awful lot of RFI.
Better designed circuits limit the abruptness at which the switching takes place and so get rid of most the RFI.

 

[bumperm]

:confused2:Wonder how many of our other members are like me and have almost no idea what you're talking about.:laughing: But it's undoubtedly good information for some.

I'm with you Bird . . . I have no idea what I'm talking about either :c)

bumper

 

[Bird]

I'm with you Bird . . . I have no idea what I'm talking about either :c)

bumper

:laughing::laughing::laughing:

Incidentally, I've been hunting some short, preferably globe shaped, LED bulbs with chandelier bases for the lights over the mirror in our master bath. There's 6 of them and each one has a seashell shaped cover or shield that never touched the little light bulbs themselves. I haven't found any in the stores here, but this morning, I bought 6 Sylvania 25W replacement bulbs from Lowe's. They were only $7.98 each plus tax. They work just fine, BUT . . . I've noticed that when I turn on the light switch it's a fraction of a second before the lights come on; haven't noticed that with any of the other LEDs. And if I tighten the long screws that hold the seashell shield it will actually touch the bulb. Does that matter?

 

[Sodo]

After a short pause it turns back on again, then off, on, off, on, off, on, off ...
All this on-off-on-off switching is very fast, many thousands of times every second, so the human eye is not aware that anything is happening.

If this on/off/on switching happens more than 3 times per second and less than 3 billion times per second, it produces radio noise considered to be within the Radio Frequency (RF) range.

For a very simplistic example, if the LED light switching circuit in your LED happened to switch at the rate of 970 kHz (970,000 times per second,,,) then it would make noise that you could receive at "970Hz" on your AM radio. If you like to listen to a station at 970 then you would hear only radio noise from your LED light, instead of your favorite blabbermouth.

The RF range is very wide (3 times per second, to 300 billion times per second) consequently almost any oscillator creates RF and may create noise on many frequencies.

 

[BobRip]

:laughing::laughing::laughing:

Incidentally, I've been hunting some short, preferably globe shaped, LED bulbs with chandelier bases for the lights over the mirror in our master bath. There's 6 of them and each one has a seashell shaped cover or shield that never touched the little light bulbs themselves. I haven't found any in the stores here, but this morning, I bought 6 Sylvania 25W replacement bulbs from Lowe's. They were only $7.98 each plus tax. They work just fine, BUT . . . I've noticed that when I turn on the light switch it's a fraction of a second before the lights come on; haven't noticed that with any of the other LEDs. And if I tighten the long screws that hold the seashell shield it will actually touch the bulb. Does that matter?

This is OK. The small base probably allows smaller parts with less power available and longer time to get up to the required voltage. I doubt if touching the sea she'll has an impact. You could remove one seashell and see if that makes a difference or try in another place.

 

[Bird]

Thanks, Bob.

 

[KubotainNH]

I just got done testing a Cree "4flow" LED 40 watt bulb for RF interference. I could not find any from 200Khz to 59Mhz. Now I did not check every frequency in that range, but I do have a spectrum monitor on the receiver and did not see or hear anything. This is the new Cree design without the heat sink, that has little slits in the plastic "bulb" that allow for a chimney effect of cooling air to flow through the bulb. I was a little worried that without the shielding effect of the metal heat sink these might be dirtier than the othe Cree's but that does not seem to be the case. So far so Good on the Cree Brand. I am starting to get on this LED bandwagon. CFL's I think were just a stopgap measure. We shall see how reliable and long lasting these "newfangled" bulbs are, but if they will do what they say, then yes, by all means they will save me some money in the long run. Not to mention the ability to run them at least for a short time from my UPS's. And the reduction of unwanted heat in the summer will be great.. Not so great in the winter. Also one other downside, my cat liked to park directly under the lamp on my table to sleep, and now she doesn't seem to care to do that. Not enough warmth to interest her! Sorry kitty!:cat:

Cree Model # BA19-04527OMF-12DE26-3U100 Internet # 205597080 Store SKU # 1001085533

Interesting, not bad for $8 I see. The one LED bulb I have is Cree but not this model. I do like it but haven't burned all my old school stock pile yet.

What I do wonder is if new houses will eventually be wired for DC LED's and have a power supply elsewhere to power them. Imagine the savings in house wire if all ceiling lights ran DC instead. Just have a power supply or 2 near your circuit box. Interesting thought anyway.