4
\$\begingroup\$

I have a speaker (Logitech Z623) which was driving me crazy with a low intermittent hum. In order to get to the root of the problem, I attempted multiple solutions. In the process I discovered that my UPS is making the exact same hum when on AC mains.

In order to rule out the UPS (APC SUA1500I), I plugged both of them directly into an extension cord and both of them hum in sync with each other (the speaker is not plugged into the UPS). I tried plugging the speakers in different outlets across the house and they still make that exact same hum (same tone, duration etc.)

Initially I assumed that capacitors in my speaker as well as the UPS are degrading but there's no explanation for the exact same hum (and of the same duration) in both the devices.

So this leaves my electrical wiring as the sole suspect. Is this possible? What should I go looking for next?

What I have tried so far -

  • monitored the voltage (230 V), frequency (50 Hz) etc and everything seems to check out.
  • tested my speaker inputs and the hum is still there without any inputs connected.
  • tried turning off all heavy electronics in my house
  • tried turning off the lights in the room
  • turned off pretty much everything in my house
  • eliminated the building elevator as a probable source. There's no discernible pattern between the two.
  • flipped all breakers in the house except the one powering the UPS and speakers. The hum is still there.

P.S. Adding some more details -

  1. Plugging speaker into UPS and switching off the mains eliminates the hum.
  2. UPS has a 3-pin plug while the speaker is only 2-pin.
  3. The hum goes away in the middle of the night. This has happened 3 nights in a row. I can't help but feel that is coming from outside the house. Well, I spoke too son but the duration between the noisy events is way longer in the night than during the day. 1-2h compared to 3-5 mins in a day.

PPS. Attaching a recording. This is a very faithful recording to what I'm hearing. Tested it across the speakers, multiple headphones and phone speakers. The recording is made at 96 kHz and 24-bit (the highest possible setting in my Zoom H4n)

Update - Recording #2 - https://www.dropbox.com/s/isquus515alyz3f/Speaker%20noise%202%20-%20STE-000.wav?dl=0

Starts at 24s and ends at 55s.

OLD RECORDING - https://www.dropbox.com/s/1q0v0ajv1f8dq7g/STE-000.wav?dl=0

ocrdu has provided with an analysis of the audio file:

enter image description here

These are two periods of the noise; the periods are repeating at 100Hz with slightly varying DC offsets. From the comments: The spectrum shows a 5kHz signal with its harmonics (10kHz and 15kHz).

\$\endgroup\$
1
  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$
    – Voltage Spike
    Commented Nov 10, 2020 at 16:11

3 Answers 3

6
\$\begingroup\$

Load waveform in audacity, normalize. Buzz is present at the beginning and absent at the end, allowing comparison. There is a buzz signal around 5k with harmonics, and 50Hz hum. FFT on a representative interval:

Buzz absent: 50Hz -37dB

Buzz present: 50Hz -37dB

Therefore 50Hz hum is probably from the microphone and irrelevant to the problem. I remove it by applying 500Hz highpass filter to the whole waveform. Result:

enter image description here

enter image description here

-> Peak around 5500 Hz plus harmonics.

Clues from waveform:

  • No change across the recording, stops abruptly

  • 5k buzz present with duty cycle about 60%, absent in rest of half cycle

These two are in contradiction. The duty cycle would mean a phase cut dimmer or something that is modulated on AC cycle, with a noise generator downstream. For example a triac dimmer and a bulb with a noisy switching driver. But the latter would run at a much higher frequency than 5 kHz. And the 5kHz frequency indicates something heavy, like a motor VFD, for example in the elevator. But it stops abruptly at the end, which means it's probably not a heavy motor VFD like the building elevator.

If you had a recording of when the buzz starts, that would give more information.

Here's what I'm thinking about to explain both the 5k frequency and the 100Hz modulation:

Faulty LED/CFL bulb or electronic ballast with a switching supply that hiccups at 5kHz, coupled with a triac dimmer.

Or a motor VFD that runs at a low switching frequency, maybe an inverter fridge or heat pump.

Or the offending device is powered from three phase AC to DC rectifier without power factor correction, and only pushes noise into the phase powering your appartment when corresponding diodes conduct.

You could pick up the Logitech speakers and plug them into the wall socket closest to the main electrical panel, then flip all the breakers one by one until you no longer hear it.

EDIT from your recording of when it starts:

enter image description here

So, it starts abruptly, and then the noise pattern remains the same, which means it's probably not a motor. The noise pattern would change as the motor accelerates.

Note your first signal (top) is slightly different from your latest upload:

enter image description here

\$\endgroup\$
14
  • \$\begingroup\$ This is very interesting. Thank you. I have added another recording here - dropbox.com/s/isquus515alyz3f/… Starts at 24s and ends at 55s. I have eliminated the building elevator as a potentional source because there is no discernible pattern in the hum and start/stop of the elevator. \$\endgroup\$ Commented Nov 10, 2020 at 14:51
  • \$\begingroup\$ I will attempt your suggestions and report back. Thanks! \$\endgroup\$ Commented Nov 10, 2020 at 14:54
  • \$\begingroup\$ I flipped every breaker except the one powering the UPS and speaker and the hum was still there. I don't have any CFL bulbs. There are plenty of Philips LED lights and LED strips. There are no dimmers though. There is no heat pump. There is an inverter fridge but I switched it off at the breaker and that did not solve the problem. \$\endgroup\$ Commented Nov 10, 2020 at 15:28
  • \$\begingroup\$ Maybe no dimmer then. The 34s duration is suspicious. Does it always last around 34s or similar duration? Thinking about the building lights on a timer... \$\endgroup\$
    – bobflux
    Commented Nov 10, 2020 at 15:29
  • \$\begingroup\$ Not always but a lot of times it is 34s. I have seen durations ranging from just 2s, 12s upto 34s. 34 seconds has happened a lot of times though \$\endgroup\$ Commented Nov 10, 2020 at 15:33
2
\$\begingroup\$

The speaker may indeed be exhibiting 120Hz supply hum (not 60 Hz, since after rectification the peak and null each exhibit identical peaks). On the UPS, it's quite likely caused by a loose lamination in the transformer itself vibrating as the magnetic field peaks, or by the coil itself vibrating. They both sound the same, but one is converted to audio by the amp and speaker, and one is a mechanical vibration which produces the same effect due to the same cause, but without any transducers explicitly designed to convert the electrical signal to audio. (Although the case could be made that the transformer isn't that much different that a speaker topologically...)

\$\endgroup\$
2
  • 1
    \$\begingroup\$ The hum is not continuous. It starts and stops and this happens in sync on the UPS and the speaker. My electronics skills are pretty limited but can the above be responsible for the intermittent hum? FYI, the frequency in my country is 50Hz. \$\endgroup\$ Commented Nov 10, 2020 at 9:39
  • \$\begingroup\$ In that case, I'd check to see if the UPS still exhibits the problem without the speaker attached. If so, it would seem that the UPS is the source of the issue, which is dirtying up its output enough to affect the speaker. It's also possible the problem is further upline and is affecting both the UPS and speaker similarly. If you have another UPS around, you could try swapping them out. \$\endgroup\$ Commented Nov 12, 2020 at 21:26
2
\$\begingroup\$

When you turned off the mains and the hum went away, you completed the first step to finding the problem. Now continue the process of elimination, first by circuit breaker, then by testing each item on the offending branch.

\$\endgroup\$
15
  • \$\begingroup\$ I have shut off all electronics in my house and the hum is still there. I cannot unplug every item from the socket but can definitely switch off the appliance. Should I try by switching off every single circuit breaker one by one? \$\endgroup\$ Commented Nov 10, 2020 at 13:08
  • \$\begingroup\$ You said you shut off the mains and the hum disappeared. How did you shut off the mains? Yes, I would try individually shutting off each breaker. \$\endgroup\$ Commented Nov 10, 2020 at 13:25
  • \$\begingroup\$ By shutting off the mains, I meant to say that I turned the switch to the off position. Turning the switch to the off position is sufficient to remove the noise. I'll try individually shutting of each breaker in some time. \$\endgroup\$ Commented Nov 10, 2020 at 13:27
  • \$\begingroup\$ Sorry. Can you be very specific about which switch you are turning off to turn off the mains? Wall switch? Appliance switch? Breaker switch? Whole house switch? Surge protector switch? Knowing would help, thanks. \$\endgroup\$ Commented Nov 10, 2020 at 13:36
  • \$\begingroup\$ The wall switch next to the socket into which the device is plugged in. \$\endgroup\$ Commented Nov 10, 2020 at 13:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.