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

[joehersey]

Although I retired from the Coast Guard some five years ago, I've been helping them part time and was the one that drafted the USCG Safety Alert. Rlevin73 and I are attempting to address this problem.

Cpt Pat is absolutely right about the test procedure in the Alert not working with a radio using a PLL squelch, and has the right work-around procedure (tuning to a weak NOAA Weather Radio station, or for those outside the US, any weak continuously broadcasting station will do).

We're concerned about the lack of applicable standards and want to address that. FCC Part 15 exempts devices on boats, but Part 15 limits are insufficient anyway. European CE requirements are better, but are also insufficient. The examples of problems listed in the USCG Marine Safety Alert all came from France, who alerted USCG to the problem. UL to my knowledge does not address RFI. The best applicable maritime standard is IEC 60945 (EN 60945 in Europe), but while that standard is more stringent than Part 15 and CE limits, it was never designed for on-deck sources of RFI. IEC 60945 assumes the source of interference to be from the bridge of a ship and assumes the bridge is separated by at least 15m from antennas. IEC 60533, the companion standard to IEC 60945 states “The emission limits of this International Standard may not, however, provide adequate protection against interference to radio receivers when other ship equipment is used closer than 3 m to the receiving antenna”. Interference from IEC 60945-compliant LEDs will be non-existent if separated by 15 m from antennas, and minimal if separated 3m from antennas.

The automotive industry has adopted a stricter standard that seems to have worked well for them, and fortunately for us many LED drivers already meet that stringent standard. That standard is CISPR 25 Class 5. If you can find LEDs meeting that standard, you should have no problems with LED interference. Although CISPR 25 is currently only applicable to boats less than 15 m and assumes the power source to be DC, I am hoping to find a way to apply such a standard to above-deck sources of RFI on any ship.

[john61ct]

Fantastic!

So let's start listing products that comply with CISPR 25 Class 5 !

Or for those that want to test on a bench
https://www.batterfly.com/shop/tekbox-tboh01

[Lake-Effect]

Ed Sherman is the ABYC's electrical guy. Here is a recent article from him on RFI, including some detection tips.


It would be helpful if the ABYC would look into RFI and interference from LEDs and establish a spec that manufacturers could aim for.

[joehersey]

The secretary for IEC Technical Committee 80, the group responsible for IEC 60945, told ITU that in response to this LED problem that "IEC TC80 will therefore generate a new preliminary work item concerning requirements for LED navigation lights and their installation to be developed in the committee ". TC80 next meets Oct 2019. I'm not exactly sure what he has in mind. We are considering proposing they open 60945 just enough to apply something like CISPR 25 Class 5 to above decks equipment capable of causing RFI, or even just to LED lighting.

[TLSparks]

You posted a very Good write up and alert for cruisers. The Coast Guard does not mention MF and HF. It can also be impacted. Unfortunately, the signals are radiated like an transmitted signal. Since that is what receivers are looking for, it is very difficult to eliminate that signal without reducing the receivers signal unless it is coming in via the power supply, 12V DC. If one knows the exact frequency of the LED, a device called a trap could be used, but I suspect that the LED frequencies are not constant.



The only way to really escape the issue is to buy good quality LEDs.


Merry Christmas everyone.

Originally Posted by Cpt Pat

I'm not sure if this has been posted before, but for anyone who hasn't heard, several LED manufacturers are skimping on the radio frequency interference suppression in LED lights. This could wipe out your VHF and AIS reception (you can still transmit - but you may not be able to receive). That is happening, and the Coast Guard is now warning mariners about the problem.

See the USCG safety alert for details about LED interference with AIS & VHF: https://www.dco.uscg.mil/Portals/9/DCO%20Documents/5p/CG-5PC/INV/Alerts/1318.pdf?ver=2018-08-16-091109-630.

The USCG procedure in that alert won't work with PLL-based receivers - like Standard Horizon's because the squelch will only open with a coherent carrier - not with random noise. It's better to tune in the weakest NOAA weather station, or some other weak station, and then turn on your LEDs - one at a time. If the NOAA station vanishes or is buried in noise, you have a noisy LED. I've been in contact with the USCG people who are investigating this problem and informed them that their test procedure won't work with all comms radios. I've also reported a few tests to them.

This problem is especially bad if you have a noisy LED lamp near your antenna - like a bad tri-color at the masthead. But the RFI can radiate over the supply wires throughout the boat. Installations are so variable that defined countermeasures are impractical.

A little background: it's not the LED lamps themselves that are at fault, but rather the cause is the internal pulse duration switching power supplies - intended to allow the lamps to operate over wide voltage ranges. Those of you with a radio background will know about old spark-gap transmitters. They were outlawed because they broadcast across a wide RF spectrum. Well, these defective LEDs are doing the same thing. Since the lamps are usually encapsulated, the problem isn't field repairable. The only practical fix is for the manufacturer to reduce the slew rate of the switching supplies (which will make them slightly less efficient). More time in the switch transfer range = more power dissipated as heat.

This shouldn't besmirch all LEDs. After spending hours in a Faraday cage with a spectrum analyzer, I found that some are very quiet. I tested many lamps and found the ones with the European "CE" seal were generally the best. To get that seal, the lamps have to pass a lab test for RFI. Here in the US, it's the wild west. Caveat Emptor. The worst lamps I tested were the incandescent substitutes intended to fit in the incandescent lamp housing.

If you're sailing an untested boat, and you can't receive a station, turn off all your LED lamps. It may solve the problem (if you don't mind being in the dark with no navigation lights).

You might try RF chokes, shielded cables, parallel bypass capacitors, etc. But some of the RFI is radiating right off the lamp housing, and you may not have any success. If it radiates RFI, return it, and report the make and model to the US Coast Guard: https://www.navcen.uscg.gov/?pageName=contactUs, regardless of whether you are in the US. They will appreciate the report. They are trying to compile a list of bad lamps and makers.


73
N8QH - out

[Rough Magic]

Following

[brownr377]

@joehersey Thanks! That's really good info, been a while since I had to test to 60945, but you're absolutely correct its scope is for bridge mounted equipment, the problem we had as labs was, it was all we had for any kind of marine electronics regardless of where they got mounted in a ship. Let alone small boat (and I include most of our boats in that category)...

For people wanting to buy bulbs that are qualified to CISPR 25 yeah that's a great idea, however, as a consumer it's not so straightforward to figure out what standard was applied to claim compliance to the EMC directive, and apply the CE Mark.

Best bet is to buy bulbs that were aimed at the automotive space that is CE marked. Lighting meant for residential or commercial applications will be qualified to CISPR 15 which won't help you in a marine application. If you really feel like digging you can pull up the declaration of conformity for the model of interest, it may list specific standards, it usually doesn't. After that, you have to find the compliance department for the manufacturer and make a direct inquiry. In my experience, they will be quite forthcoming, but it takes a bit of work on the users part.

[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!

[SoonerSailor]

Err, I meant to say “odd harmonics” not third order harmonics.

[Wotname]

Originally Posted by SoonerSailor

Err, I meant to say “odd harmonics” not third order harmonics.

Yep, that's better

[continuouswave]

Originally Posted by Cpt Pat

...USCG procedure in that alert won't work with PLL-based receivers...like Standard Horizon's because the squelch will only open with a coherent carrier - not with random noise.

Please give more details about when the epoch of "newer" began for Standard-Horizon VHF Marine Band radios, that is, when their squelch circuit no longer just operates on noise. I would think the PLL-receiver era began when crystal controlled oscillators went away.

Also, if a Standard-Horizon VHF Marine Band radio has a PLL-based receiver, does the squelch EVER open if there is no signal present?

[Cpt Pat]

Originally Posted by continuouswave

Please give more details about when the epoch of "newer" began for Standard-Horizon VHF Marine Band radios, that is, when their squelch circuit no longer just operates on noise. I would think the PLL-receiver era began when crystal controlled oscillators went away.

Also, if a Standard-Horizon VHF Marine Band radio has a PLL-based receiver, does the squelch EVER open if there is no signal present?

This could be a very long reply. The challenge is making it short. I can only make it so short and still answer the question.

PLL receiver design is a subject separate of frequency synthesized local oscillator design. In fact, most older FM receivers, such as FM broadcast receivers, used neither crystals oscillators nor frequency synthesizers. They used analog tunable oscillators (variable capacitors) and coupled RF preamplifiers tuners (another ganged variable capacitor). Anyway, this crystal controlled versus frequency synthesis discussion is a separate topic from squelch circuits.

Receivers throughout the early to late 20th century used superheterodyne designs (superregenerative designs died early) where a local oscillator signal was mixed with a limited band of amplified signals from the antenna. The mixed signals were passed to an intermediate frequency (IF) filter and then amplified. Any signal plus any noise that fell within the bandwidth of the IF filter was then sent to either a diode detector (for amplitude modulation) (for SSB, a second frequency was injected here to simulate the missing carrier) or a frequency discriminator (for FM modulation). If the IF filter had a pass bandwidth of, for example 20 kHz, any signal or noise at all in that bandwidth was amplified and treated as a signal. That "signal" includes whatever random noise might be present. That random noise (plus any desired signal) level was detected and sent to a detector and converted into a strength-representative voltage where the user could set an adjustable threshold voltage called "squelch." If the noise plus signal - or just noise - was above that threshold the speaker would be unmuted. This design dates back to the vacuum tube era because it could be implemented with as few as 5 tubes for AM and 7 tubes for FM ("valves" for our British friends).

Phase-locked loop (PLL) receivers became more common in the late 20th century using advanced integrated circuitry in FM receivers. It functions by locking onto the carrier frequency and following it as the frequency is swung with modulation. These receivers detect and lock onto a carrier to mute and unmute the speaker. No FM carrier signal - no lock - no speaker output. Random noise - no lock - no signal output. This design provides enhanced noise rejection because random noise isn't detected. Most of the noise in these receivers is internally generated noise. When listening to a weak signal the has some noise mixed in, instead hearing rushing and crackling noise -- you hear only a hiss. The squelch threshold then adjusts the sensitivity of the lock detection. The squelch circuit isn't binary, there needs to be a user adjustable range. In these receivers, if random noise covers up a desired signal (such as a noisy LED), instead of hearing that signal along with the noise (as you will with a superhterodyne receiver), the PLL has no carrier signal onto which it can lock, and you hear nothing. The signal is "blocked' because the PLL cannot lock onto a carrier. That's why listening for the squelch to unmute as a detection method of noise with PLL receivers will fail.

Whew!


I'll refer any reader who has follow on questions to study materials on receiver design because if I go past this level I'll have to write a book with schematic diagrams and mathematics.

Radio people please forgive my omissions of subcircuitry, capture effect, advanced designs, etc. Also, feel free to shorten that answer if you can.

[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."

[Rough Magic]

Cap Pat, I for one, quite enjoyed your reply. I'm not a 'radio person' but understood your explanation. Thank you.

[Lake-Effect]

Interference is interference, and can screw up your reception, more or less, regardless of detector type. LED systems should not cause any discernable interference on any marine radio or other receivers, full stop.


If you're not watching parades or football today, you can get your fill of FM detection theory here.