Jump to content

LED Light Interference


Recommended Posts

Things are very quiet here these days. I thought I'd post some stuff that is sort of relevant and may be interest.

My Jayco caravan came with LED lights that makes the FM /TV 6M, 2M and 70cm Ham bands (50,144 and 430MHz) unusable. I now a combination of incandescent and LED lights in it.

Purple = Peak / Max Hold
Yellow = LED Turned Off

Jayco LED Mag Probe.jpg

 

I checked out the LED light using my Spectrum Analyser with my poor mans Officeworks probes ;-) Probes M and E.JPG

DIY M and E Probes.
M Probe is as per the bottom Left using semi rigid coax.
E Probe is as per the bottom top right using RG174 (?)
emcprobes.jpg

 

Link to comment
Share on other sites



The tick system is designed useless. Doubly so for anything that falls under series 6 or 7 (household appliances & lighting; AS/NZS CISPR 14.1 & 15). Those basically amount to:

  • Creates interference <30MHz? Don't care.
  • Creates interference >30MHz? Don't care much...

It's quite possible that these lights actually meet the standards, particularly since they're not mains-powered (so EN61000-3-2 & 3 don't apply).

Link to comment
Share on other sites

The switchmode power supplies are the source of the interference and they are connected to the power mains. If the filtering is inadequate can feed interference back into the mains as well as  feeding it into the wiring to tle LEDs. How much interference comes from an LED torch which is battery powered! None!

Alanh

Link to comment
Share on other sites



On 22/12/2016 at 3:19 PM, alanh said:

The switchmode power supplies are the source of the interference and they are connected to the power mains. If the filtering is inadequate can feed interference back into the mains as well as  feeding it into the wiring to tle LEDs. How much interference comes from an LED torch which is battery powered! None!

Alanh

Spoken like a man who has never had to chase down & deal with RFI from a battery-powered lamp.

Yup, I'm sure the 12V T10 base Jayco caravan lamps Col has tested above are connected to the power mains, and aren't battery powered at all... </sarcasm>

(p.s. you obviously didn't bother to read your own links & the documents they refer to. If you had, you'd have known that the EMC standards AS/NZS CISPR 14.1 & 15 apply to both mains-powered and battery-powered appliances & lighting...)
 

Link to comment
Share on other sites

On 25/12/2016 at 10:11 AM, alanh said:

Juar shows how little you know of electronics.

Hey, I know those things often use switchmode converters. After all, it wasn't me who said ...

On 22/12/2016 at 3:19 PM, alanh said:

How much interference comes from an LED torch which is battery powered! None!

That was you.

You seem to have a problem with remembering who said what - even when it was you who said it!

 

On 25/12/2016 at 10:11 AM, alanh said:

LEDs themselves run at around 2 V depending on the colour. The 12V has to be reduced to that voltage and another switch mode power supply is used!

Back in the real world, those little T10 panels (and many battery-powered LED lamps) fall into 2 types:

  • Ones with 12~15 LEDs, and dropper resistors feeding each parallel string of ~3~4 series-connected LEDs. Many of these also include a bridge rectifier so they work regardless of source polarity. These produce no RFI.
  • Ones with 15~16 LEDs, and a boost converter/constant current LED driver that boosts the 12v to ~18~38v (depending on the current - remember, LED's are current-driven devices, not voltage-driven). The boost converter usually incorporates it's own reverse-polarity handling. These produce anything from minimal to lots of RFI, depending on how cheap and nasty they are.

By both visual inspection of the back of the board shown and the EMI generated, CMatten's lamps are obviously of the latter type.

It's possible I guess, in theory, that someone makes them with a buck converter to reduce the 12v supply to ~2v like you say (implying that the LEDs are all wired in parallel). All I can say is I've never seen any like that in real life, have never seen any like that for sale anywhere, and an extensive search online has found none. They're all either the dropper resistor or boost converter types I outlined above.

Oh, and I think you'll also find most white LEDs - either RGB triplets or phosphor-converted - run around 3v~4v Vf, not 2v. So even LED torches - which are usually single or double cell (i.e. 1.5v or 3v)* or 18560 (i.e. ~3.7v) - also use a boost converter / constant current driver to boost the source voltage enough to overcome the LED Vf.

(* decent ones, that is - not the chinese cheapies that use 3 or 4 AAA or AA cells in series...)

Looking forward to what bit of your technical ignorance you'll expose in your next post or comment...

Edited by Malich
(clarification, in blue)
Link to comment
Share on other sites

  • 2 weeks later...

Not at all,

LEDs themselves do not generate EMI as I stated, after they are a light emitting Diode. The example Malich used uses a switching buck inverter which takes the 12 V DC and chops it up so that only a small percentage of the time the output is switched on, the result is fed in to a capacitor to average it out. It is the high speed switching which causes the interference. The above process is identical for mains powered LEDs as well, but with the addition of a rectifier to convert the AC mains into DC first.

If you are using a battery to power the LED you don't want to use dropping resistors as they waste power, making the above circuit the most efficient.

As for the voltage of LEDs, the voltage as I said depends on the colour. The real world is a LED is a semi-conductor junction whos turn on voltage is controlled by the doping of the semiconductiors. It is upto you or the manufacturer as to whether they are connected in series or paralell.

Alanh

Edited by alanh
Link to comment
Share on other sites



Malich has obviously never soldered in a single LED http://www.altronics.com.au/p/z0700-red-8mcd-3mm-led/ as I said about 2 V, so is 2.2 V close enough! Green is also 2.2 V, orange 2 V, Yellow 2.2 V, blue 3.5 V, White 3.5 V. No where near 12 V.

I was however talking about the ones with AAA cells not car batteries!

By the way the switching in these devices is usually performed by a transistor.

 

Link to comment
Share on other sites

For the reader,

Malich has clearly demonstrated operation of light emitting diodes (LED's) possibly from a designer's perspective and certainly from a perspective of having used them.

For anyone choosing to use LED's in a circuit, always base your design on current controlled/limited source. Despite what some have written here, do not parallel LED's directly, although their breakdown voltage for any particular device will fall into a tight voltage range, they will never share current equally and reliability will be jeopardised. Series your LED's with current limiting on each string, then you can safely parallel these strings without risk.

CMatten's tests are quite telling, I particularly liked your probes and backs up what is known to now be all too common a problem, switching with uncontrolled rise times, poor snubbing and insufficient pre and post switching element filtering whether it be bipolar or MOS based.

James

Link to comment
Share on other sites

Dear Reader

The poster above has kindly included a link that describes an impact of incorrect use of LED's through paralleling non current controlled strings. Most kind, although it's content appears to have been misinterpreted by this poster in this case.

For clarity, it is worth reading for yourself.

James

Link to comment
Share on other sites



Dear Reader

When one reads what another writes for any particular endeavour, you can tell if they are simply a reader of that endeavour or actually involved in it, by that I mean some things are understood and others need to be stated.

" It also says how to connect them in parallel so that a single LED failure will not cause total failure" should be an example of the understood.

I note also that the article quoted by the poster above having merit in terms of it describing the need for current limiting/control, I do not advise using the circuit they describe. I suggest equally simple circuits that permit individual string limiting independent of their neighbouring parallel strings is possible and superior and available even at the time of this quite old reference.

James

Link to comment
Share on other sites

This discussion was about RF interference from LEDs.

The Jayco light is a very bright one which is connected to the caravan 12 V "battery" supply via a bridge rectifier, making the power supply almost identical to one run from the mains.

If you look at the Jayco light it is hardly a torch, its 24 LEDs on a board.

Now it is proven that the LED is not the source but the high speed switching to limit the current.

James you can suggest all you like, but as soon as you use a parallel string, it is the same problem as a pair of LEDs in parallel. Each string will now have different currents. It all depends on the number of LEDs involved, and the voltage of the power supply.

The characteristics of parallel and series circuits will never change. In simple circuits resistors are used to limit the current which will also produce no RFI. You can also use transistors to be analog current limiters which will also produce no RFI. It is only when you use switching current limiters you get interference.

Perhaps a simpler reference might be better http://www.instructables.com/id/LEDs-for-Beginners/step8/Wiring-up-multiple-LEDs-in-parallel/

 

Alanh

 

 

Link to comment
Share on other sites

Dear Reader

It really should be easier than this but I am a half glass full person and so I will treat this as an opportunity to see where it first went awry for the poster above using quotes from his contributions.

" This discussion was about RF interference from LEDs." 

Well actually it is not, the original discussion is about RF interference from LED lighting which of course includes the LED itself and all the associated circuitry for operation from its power source. Unfortunately and quite inexplicably, despite CMatten, Malich and others clearly knowing the interference was a result of the regulation circuitry not the LED itself, the poster above decided to declare that the interference was not from the LED itself and cited that a LED torch operating from DC created no interference as ultimate proof. Since the poster above was apparently unaware that all LED lighting, be it DC or AC has switching regulation it sort of made his explanation and proof just a little silly.

 

"James you can suggest all you like, but as soon as you use a parallel string, it is the same problem as a pair of LEDs in parallel. Each string will now have different currents. It all depends on the number of LEDs involved, and the voltage of the power supply."  

Not choosing to be too critical but several here have attempted the have the poster above understand that these are current driven devices.  By having a current limiter on each string of seriesed LED's you can parallel as many of these strings as you choose, the beauty of this is that the power supply voltage can vary all over the place so long as the minimum voltage is greater than than required by the sum of the breakdown voltages of the LED's in the series plus the voltage drop of the regulator. Of course the power supply should not be permitted to go so high that the voltage rating of the regulator is exceeded.

In essence if we go back to CMattens light which I understand is 12vdc, it should be designed to operate from 10-16vdc quite happily. Regarding the poster above's statement on problem paralleling, I dare venture to say that CMatten is likely to have a second light in his caravan and depending on the level of opulence possibly a third, fourth or fifth. I believe that I can state on CMatten's behalf that he does not cringe in anticipation of switching on the second, third, fourth light in the van in fear of it disturbing the first light. In case the penny still hasn't dropped for the poster above, oh yes, turning more than one light on in the caravan is paralleling seriesed LED lighting.

The poster above then notes this " The characteristics of parallel and series circuits will never change. In simple circuits resistors are used to limit the current which will also produce no RFI. You can also use transistors to be analog current limiters which will also produce no RFI. It is only when you use switching current limiters you get interference."  BINGO !!! Dear reader there is light at the end of the tunnel, the poster above has something right.

 

Finally after the burst of light just above (excuse the pun on LED lighting) this is stated"

" Perhaps a simpler reference might be better http://www.instructables.com/id/LEDs-for-Beginners/step8/Wiring-up-multiple-LEDs-in-parallel/ "

We can only hope so, no doubt the poster above will let us know.

James

 

Link to comment
Share on other sites

OK...

So CMatten was kind enough to send me the part # & link to the datasheet of the chip on his LED light (he actually did that a couple of weeks ago, but I was on a work trip to Perth & then on holidays & mostly offline). Turns out it is a buck converter / constant current driver - a type I haven't seen before in this sort of lamp (as I said all have been either simple dropper resistor setups, or boost converter / constant current drivers). So Alan was at least partly correct.

I'm stilll not sure how or why Alan came to the conclusion that anyone was claiming the interference came from the LEDs themselves. As far as I can see, no-one - neither myself nor anyone else - has said anything like that. If somehow anyone did get that idea from what I wrote, I apologise - I asumed it was obvious and self-evident that the interference comes from the switching regulator driving the LEDs, not the LEDs themselves.

But on to a couple of remaining specifics:

On 10/1/2017 at 2:21 AM, alanh said:

I don't want one LED failure to stop all the LEDs. This was the problem with fillament Christmas tree lights which were all in series.

Then you'll be absolutely shocked to learn that in the vast majority of cases - battery powered or mains, doesn't matter - they are in fact connected in series strings (with multiple series strings in parallel). There's some obvious reasons for that, which I won't go in to - anyone with a basic understanding of electronics should be able to figure them out.

For safety-critical / high reliability applications (e.g. vehicle brake lamps, emeergency lighting, etc), they often use shunt protection devices - as described here, or simlar devices from Littelfuse, ON Semi, etc. - across individual (or small groups of) LEDs. They ensure the rest of the series string continues to operate if an individual LED fails.

 

On 11/1/2017 at 4:12 PM, alanh said:

I know Instructables are usually pretty terrible, or at least simplistic to the point of facile. That one borders on useless when it comes to explaining LEDs...

(Ah, I've figured it out now. It's the first result when you Google "wiring LEDs"...)

Edited by Malich
Update, in blue.
Link to comment
Share on other sites

Malich,

Apology accepted.

The shunt protection article is 5 years old. Its disadvantage is that it make the other LEDs in the string more vunerable to failure.

The way to reduce the current variation in each LED junction is to make them all on the same substrate, using SMT. So all junctions will be at the same temperature and created with the same doping concentration. If not integrating a constant current source for each LED is also possible. Consider the huge large LED displays at entertainment venues which have around 8 million LEDs.

Alanh

Link to comment
Share on other sites



31 minutes ago, alanh said:

Malich,

Apology accepted.

Thank you. You still owe everyone quite a few though...

 

32 minutes ago, alanh said:

The shunt protection article is 5 years old. Its disadvantage is that it make the other LEDs in the string more vunerable to failure.

I only posted the article for you as an introduction to the kind of devices used in high-reliability applications. You might like to follow up the manufacturer pointers I gave & read other, more detailed documentation to understand why they don't "make the other LEDs in the string more vunerable to failure" (hint: the point has been made repeatedly in this thread that LEDs are current-driven devices and, in such applications, should be fed by a proper constant-current source...)

Link to comment
Share on other sites

On 14/01/2017 at 11:48 AM, Malich said:

If somehow anyone did get that idea from what I wrote, I apologise - I assumed it was obvious and self-evident that the interference comes from the switching regulator driving the LEDs, not the LEDs themselves

I keep being deleted, puzzling, is it the terminology used? Should I write this a different way? Here goes....

Malich, your apology evoked a single acceptance, you must have been quite surprised.

James

Link to comment
Share on other sites

13 minutes ago, James T Kirk said:
On 14/1/2017 at 11:48 AM, Malich said:

If somehow anyone did get that idea from what I wrote, I apologise - I asumed it was obvious and self-evident that the interference comes from the switching regulator driving the LEDs, not the LEDs themselves.

I keep being deleted, puzzling, is it the terminology used? Should I write this a different way? Here goes....

Malich, your apology evoked a single acceptance, you must have been quite surprised.

James

Yes, I was ;)

Edited by Malich
Link to comment
Share on other sites

  • 4 weeks later...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...
To Top