Coast Guard Safety Alert: LED lights interfere with VHF & AIS radios

[continuouswave]

Originally Posted by Lake-Effect

....you can get your fill of FM detection theory here....

Thanks for that pointer. Here is a much more direct URL to a section on Phase-Locked-Loop FM Detection that I found by following a redirect from the link you gave

https://www.radio-electronics.com/in...modulation.php

The above gives a good description of the method of FM demodulation using PLL techniques, which is what I believe PAT was trying to say when he used the phrase "PLL-based receivers."

One problem: I don't see anything in there about "coherent signals." Perhaps there is more to this detector than is explained in that link.

Also, I would still like to know when the VHF Marine Band radios made by Standard-Horizon began to employ this technique. The best demarcation of this epoch would be to denote the model designators of radios that have employed the PLL FM detector.

[Cpt Pat]

Originally Posted by continuouswave

Pat--Yes, your reply got rather long. It failed to answer the question I asked:

When did the Standard-Horizon radios entered the epoch you named as "newer" with regard to how their squelch circuit was changed so that it only opens to "coherent signals"?

Also, I will take up your challenge to shorten your answer:

The change in FM receiver design you wish to mark (with the phrase "PLL-based receiver") is in the method of FM detection. You describe the method very briefly. Then you end with suggesting readers look elsewhere for more information, but you did not point the readers to a source of information.

Please give us a link to a URL where we can learn more about this new FM detection method. Thank you. Maybe an application note for a chip that performs this function would be illustrative.

ASIDE: I am a "radio person."

As suggested above, here is a good place to start: https://www.radio-electronics.com/in...modulation.php

> Then you end with suggesting readers look elsewhere for more information, but you did not point the readers to a source of information.

Beyond the link above, I'd wouldn't know where to start. It's been 45 years since I began my career as a communications engineer in broadcasting. I learned the technologies as they developed, I have no idea what curricula exist today.

The knowledge base is very broad. "Seek and ye shall find."

> I would still like to know when the VHFMarine Band radios made by Standard-Horizon began to employ this technique.

I'm not a historian for Standard Horizon. I suggests asking them.

[Cpt Pat]

Originally Posted by SoonerSailor

Unfortunately most PWM LED drivers are delivering a square wave to the LED and back into the power leads/wiring. Usually this frequency is on the order of a few kHz at the most, but square waves are rich in third order harmonics, and they extend far up in frequency in ever diminishing power - though often not diminishing fast enough to not cause a problem with our radios. Careful design of the driver circuits to mitigate this is possible but rarely applied rigorously enough to prevent interference to a radio via nearby antenna. One can drive LEDs with passive resistive networks which are noise free, but that does come at the cost of lower overall power efficiency than a good active LED driver circuit can deliver. Here is to mandating better design!

Excellent summary of the root issue!

[continuouswave]

Originally Posted by Cpt Pat

... I'm not a historian for Standard Horizon. I suggests asking them.

Pat--I am a bit confused. You introduced Standard-Horizon "newer" radios as having a special feature in their squelch circuit. "Newer" is a term of relative history. I was just curious to learn when "newer" occurred. "Newer" could mean quite different time periods.

Also, I am still unclear how use of a PLL FM demodulator changes the action of the squelch circuit. The goal of all FM demodulators is, generally, to ignore amplitude variations in the carrier signal and to only look at frequency variations. Random noise is generally a variation in amplitude hence an amplitude modulated signal. Also, the squelch circuit usually occurs after demodulation of the signal.

If there is some special interaction of the PLL FM demodulator and a squelch circuit, I would be interested to know the basics of that mechanism.

Your remarks seem to suggest that there is something about PLL FM demodulation that might make a squelch circuit unnecessary, as I think you are suggesting the PLL FM demodulation method won't produce any output signal until it sees a frequency-modulated carrier. If that were true, then I could see how squelch would not be necessary. But as best I can tell, Standard-Horizon VHF Marine Band radios continue to have a squelch control, and if you open the squelch you still hear noise output. That seems to contract the notion that the PLL FM demodulator won't produce noise output.

Also, congratulations on your career in broadcasting. I have been working in that field myself for five decades.

[NRG]

Some important notes I feel compelled to add or clarify here:

An LED doesn't create radiated RF interference!

The noise is created by electronics that controls the intensity of the LED. The intensity is controlled by PWM (pulse width modulation) of the power source to the LED. However, the PWM frequency is usually in the sub-kilohertz range which is out of band to cause RF interference in the Mhz range.

So, what causes the noise? The source of the noise is caused by the switching transient time or what is also referred to as slew rate (the time it takes for the power signal to change from one state to the other, on, off). The amount of radiated power produced by the switching signal is also proportional to the current that is being driven. So long arrays or clusters of multiple LED's consume more power and therefore produce more radiated emissions. And finally, something that is not mentioned much in these discussions is the cable that is carrying the LED power from the driver to the LED is very likely to be a primary radiated antenna unless it is properly shielded.

Note that LED's the can change color actually consist of tiny Red, Green, and Blue LED's (some also use White) combined to appear together when illuminated. The various apparent colors are created by varying the relative intensity of the primary color sub LED's. This is accomplished with PWM.

Some conclusions can be drawn from this.
1) LED's that are not intensity controlled DO NOT RADIATE. For example, Fixed deck lights that you just turn on or off would not radiate because there is no PWM driving them.
2) Any LED the can change colors or intensity must be PWM driven and therefore can Radiate but the amount of radiation is caused by the driving electronics to which it is connected, not the light bulb itself.
3) Many LED noise issues can be eliminated by ground shielding the drive cables.

LED's don't produce radiated emissions. The driving electronics may, however.

[SoonerSailor]

NRG - most of what you felt compelled to clarify was covered in previous posts. However, just because an LED module isn’t designed to be dimmable does not mean that interference causing PWM is not used in its driver circuitry. Voltage/current still needs to be limited in some way and this is still frequently done using PWM as part of that scheme. Meaning that a non-dimming LED module is still capable of generating wideband RFI if not designed to mitigate this, or current limited with a resistive network of some type, which typically results in lower power efficiency and less constant output with varying power supply voltage - which is why that scheme is not usually used.

The masthead tricolor LED units are not designed to be dimmable, and yet are still often a source of driver generated RFI.

[Lake-Effect]

SoonerSailor is correct. The offending RFI emitter is usually the switching current regulator within the bulb or fixture itself.

[RaymondR]

Originally Posted by SoonerSailor

NRG - most of what you felt compelled to clarify was covered in previous posts. However, just because an LED module isn’t designed to be dimmable does not mean that interference causing PWM is not used in its driver circuitry. Voltage/current still needs to be limited in some way and this is still frequently done using PWM as part of that scheme. Meaning that a non-dimming LED module is still capable of generating wideband RFI if not designed to mitigate this, or current limited with a resistive network of some type, which typically results in lower power efficiency and less constant output with varying power supply voltage - which is why that scheme is not usually used.

The masthead tricolor LED units are not designed to be dimmable, and yet are still often a source of driver generated RFI.

You can drop the voltage with a series regulator and current limit with a resister, at the milliamp load currents LEDs draw the power losses are pretty miniscule.

[SoonerSailor]

Originally Posted by RaymondR

You can drop the voltage with a series regulator and current limit with a resister, at the milliamp load currents LEDs draw the power losses are pretty miniscule.

Breadboard the circuit as you describe, and measure the power into the circuit, and the power into the LED. You will find efficiency is poor compared to circuits that incorporate a switching topology to limit current. And PWM is a switching topology.

Linear voltage regulators, which you are referring to, are in a sense highly sophisticated resistors. Any resistor used to limit current is going to dissipate power. Any power dissipated in a resistor is power lost.

[RaymondR]

Originally Posted by SoonerSailor

Breadboard the circuit as you describe, and measure the power into the circuit, and the power into the LED. You will find efficiency is poor compared to circuits that incorporate a switching topology to limit current. And PWM is a switching topology.

Linear voltage regulators, which you are referring to, are in a sense highly sophisticated resistors. Any resistor used to limit current is going to dissipate power. Any power dissipated in a resistor is power lost.

The efficiency of linear regulators varies widely with the voltage differential across the regulator and can be in the high 90s at very small voltage differences. Assuming an overhead of 1.5 volts we could regulate a 12V system to 10V output of the regulator (We want the lights to stay on until the flat battery 11.7 volts is declined to)

We want to pull 20ma from the regulator and it being a series circuit 20ma is the current limit through every series device. With a fully charged battery of say 14.4 volts that is a voltage reduction of 4.4 volts. 4.4 volts X 20/1000 amps is 0.088 Watts.

Now, I'm going to build a DIY anchor light using white 30 degree 5mm LEDs so I'll need 12 of them for 360 degree coverage. Since they have a VF of 3.2 volts I can only put 3 of them in series before I hit my 10V limit, however this means I will only need to drop 0.4 volts with my series resistor so E/I=R is 0.4/.020 = 20 ohm resistor and 0.4V X 0.020 amps = 0.008 Watts.

I now need to calculate the heat loss across the LEDs. 3.2V X .020 A = 0.064 Watts/LED X 3 = 0.192 Watts.

Summing the losses: 0.088 for the regulator, 0.008 for the resistor, 0.192 for the LED string = 0.288 Watts per LED string X 4 strings = 1.152 Watts total for the anchor light.

Neither my maths nor my logic is in that good a shape these days but if my calculations and logic is sound I'm quiet happy to sacrifice 1.152 Watts of wastage power to an anchor light if it's one which does not trigger the squelch on my VHF radio or AIS receiver (Be mindful that the largest Wattage loss is generated by the LEDs and one has to have them anyway even with a switch mode power supply)

[SoonerSailor]

RaymondR -

I think your math is fine. Your scheme gives about a 66% efficiency compared to roughly 85% efficiency of a typical switching regulator. This is not a large sacrifice, particularly for an LED module that will be in close proximity to a radio antenna. In fact, I argued just as you have argued for such a scheme in this thread 5 years ago: http://www.cruisersforum.com/forums/...or-115311.html

Why this is not a more commonly used method of driving LEDs for a masthead tricolor navigation light is puzzling. As long as the LED is bright enough to meet required standards at the minimum voltage (and from a safety at sea perspective, what would that be?) and not in danger of overdriving and thermal runaway at the highest experienced voltage (15V?) it should be fine.

[Lake-Effect]

I also custom-made LED assemblies for the nav and anchor lights on our little boat a few years back. I used LED arrays that dropped about 8 to 9 volts, then added series resistance to drop another 4v at the desired current. I also added a reverse diode to protect against transient reverse voltages which will kill LEDs. Yes I'm "wasting" roughly a third of the consumed power, but it's still much more efficient than the equivalent incandescant.

[Caveat - yes, homemade nav lights aren't USCG approved. Ours have proven to be quite visible to near a mile, so I'm not particularly worried]

Originally Posted by SoonerSailor

Why this is not a more commonly used method of driving LEDs for a masthead tricolor navigation light is puzzling. As long as the LED is bright enough to meet required standards at the minimum voltage (and from a safety at sea perspective, what would that be?) and not in danger of overdriving and thermal runaway at the highest experienced voltage (15V?) it should be fine.

The main reason is that using a series resistance in this fashion means that the current (and LED brightness) will vary noticably as the supply varies from 11v to 14v. If you choose a resistance to be bright enough at 11 v, the LEDs could be over-bright and have shortened lives at the more normal 12.5 to 13v.

Anyway, it is possible to make efficient switching current regulators that don't emit significant RF interference. I believe that some of the better makers of LED tricolours already have this licked.

[john61ct]

I would think at these low currents, a "preconditioning" DC converter could pretty efficiently and cost effectively stabilize output onto a circuit dedicated to the all the boat's safety lights.

Choose whatever voltage works well for X in series, not so low that wire gauge / V drop causes problems.

Can the remote LEDs then be used without any PWM nor resistors required out at the edges?

[RaymondR]

If you really want to get down to saving a few milli ampare hours there is a way to do it at a realtively low cycle rate which will probably not trigger the squelch on you VHF receivers.

About the simplest oscillator you can build is a relaxation oscillator using a programmable unijunction transister and a capacitor. Feed the output from this to a decade or hexadecimal counter and connect each LED to ground from the outputs. The UJT will clock the outputs one at a time causing each of the LEDs to flash. If you common earth them through a single current limiting resister the current through each of the LEDs is limited.

At any rate above about 16 cycles per second, 160 for the decade counter clock pulse, all the LEDs will appear to be on all the time due to vision latency however the current through the circuit will only be that required to turn on one LED and not say all ten connected to the decade counter.

[Lake-Effect]

Originally Posted by john61ct

I would think at these low currents, a "preconditioning" DC converter could pretty efficiently and cost effectively stabilize output onto a circuit dedicated to the all the boat's safety lights.

Choose whatever voltage works well for X in series, not so low that wire gauge / V drop causes problems.

Can the remote LEDs then be used without any PWM nor resistors required out at the edges?

Interesting thought. If one was to use an efficient switching DC regulator to regulate down to say 9v, then use that for LED nav lights that were designed to run on a guaranteed 9v... You would still need some form of current restriction in each LED assembly (eg a resistor to drop maybe 1 or 2 volts), but it would still be more efficient and you'd not have varying brightness thanks to the steady 9 v.


The two problems are:

1. system complexity - a separate regulated DC source for specific lights. What happens if the reg dies?
2. you still have a switching regulator, that needs to not radiate RFI

From the perspective of simplicity, the best solution overall is for the marine LED industry to just solve their RFI problems (as many already have)

Originally Posted by RaymondR

If you really want to get down to saving a few milli ampare hours there is a way to do it at a realtively low cycle rate which will probably not trigger the squelch on you VHF receivers.

About the simplest oscillator you can build is a relaxation oscillator using a programmable unijunction transister and a capacitor. Feed the output from this to a decade or hexadecimal counter and connect each LED to ground from the outputs. The UJT will clock the outputs one at a time causing each of the LEDs to flash. If you common earth them through a single current limiting resister the current through each of the LEDs is limited.

At any rate above about 16 cycles per second, 160 for the decade counter clock pulse, all the LEDs will appear to be on all the time due to vision latency however the current through the circuit will only be that required to turn on one LED and not say all ten connected to the decade counter.

Very interesting. Hmmm.

Just an observation that in this day and age the simplest oscillator is a programmed microcontroller. Tiny, cheaper, precision, no capacitors or ring counter required to do what you suggest.

(runs off to patent office )


OK I'm back. the problem here is that lighting a few identical LEDs sequentially will not produce more light than lighting just one of those LEDs continuously. Marine LED 'bulbs' have multiple LEDs because you want the sum of all the LEDs to produce the desired light level.