1
\$\begingroup\$

I've recently revived a pair of speakers by replacing the class D amp with a new separate module (plus a simple opamp active EQ) and keeping the stock SMPS which seems to be working fine. The unit is powered by a figure 8 mains cable hence it being a Class II earth equipment (enclosure is wood/plastic).

The original amp had a 3.5mm aux input socket other than the Bluetooth (now disabled), my plan was to use this as main stereo audio in. Everything works well when playing music from the phone but connecting to any Class I equipment would generate a pretty gnarly buzz.

Also connecting to my laptop would generate a hum if the charger is plugged in (the charger has an earth pin but can't actually measure continuity from earth to any of the pins).

My assumption was that since the amp's ground reference would be floating, then connecting to a Class I device would align its ground plane to the Class I device earth without any ground loops. But it seems that there is somewhere an exchange of current which creates this buzz.

Interestingly if I connect only the ground "sleeve" of the aux input to earth I have no noise at all or a reduction in noise due to shielding, and when I connect the "tip" to a class I device output (and so lowering the input impedance of the amp's input) I start to hear the hum, which continues even if I short the input to ground.

I could see in the original design of the amp the ground/shield from the aux in was connected directly to the ground, so if they've made it work there must be a way I'd assume. I fancy the idea of using inductors but not sure where.

Added block diagram and schematic of EQ section.

Block Diagram

block diagram and input/EQ schematic

\$\endgroup\$
3
  • \$\begingroup\$ Just added, hope it's understandable. Thanks \$\endgroup\$
    – Alex P
    Commented Nov 16, 2022 at 12:45
  • \$\begingroup\$ Neat, there's a ground loop right there in your drawing. Could it cause the things you hear? \$\endgroup\$
    – Justme
    Commented Nov 16, 2022 at 13:54
  • \$\begingroup\$ @Justme I've tried cutting the shield/ground between the EQ and the Amp module, issue is still the same. \$\endgroup\$
    – Alex P
    Commented Nov 16, 2022 at 14:28

1 Answer 1

1
\$\begingroup\$

Yes this is ground loop hum, even though there is no classical DC-connected loop here. The "isolated" SMPS is allowed to have up to a few mA 50Hz leakage to the DC side.

When you connect to floating devices, this 50Hz current cannot flow. The whole circuit will merely float around at a large 50 Hz common-mode voltage of ~100VAC without much differential signal voltage.

When you connect a device with Earth ground to the Aux jack, the SMPS leakage current will now flow back to earth through the AUX cable shield, it will produce a 50 Hz signal voltage because of the shield resistance. AUX connection is a failed design that uses the shield also as signal return, making itself sensitive to ground loops (which are agood and useful thing when properly signal transmission systems).

You can validate this the following way: Take multimeter and set to VAC. When Aux is not plugged it, you should measure ~100 VAC between Earth and any node on your amp (doesn't matter which). When you connect Aux to floating device, nothing changes..When you connect Aux to a Class I device, the large VAC on your amp will vanish because it is converted to the buzz-causing current.

\$\endgroup\$
8
  • \$\begingroup\$ Thank you for the explanation, I've done a quick test and the VAC between the amp and the earthed interface it was 100.1VAC, off by 0.1V! One thing I'd love to understand is how could I solve the issue without going the way of using input transformers. Also trying to understand, why if I connect to the earthed device just through the sleeve I don't get any buzz? \$\endgroup\$
    – Alex P
    Commented Nov 16, 2022 at 13:28
  • \$\begingroup\$ if you connect only the sleeve, then there is no signal current through the AUX cable. The local pulldown resistors in your EQ input will keep the signal at ~0 with respect to the local EQ ground. \$\endgroup\$
    – tobalt
    Commented Nov 16, 2022 at 13:38
  • \$\begingroup\$ @AlexP Remedying ground loop in AUX wiring is hard. Isolation transformers is one way. Another is differential signalling, but this would require either different cable/connectors or changes to your signal input circuit or both. Too extensive for this question. If you are interested it is better to make a new question about it. \$\endgroup\$
    – tobalt
    Commented Nov 16, 2022 at 13:42
  • \$\begingroup\$ I wasn't to keen on changing the type of input as that would require modifying the weirdly moulded front panel. I did have this lying around so I gave them a go. ebay.co.uk/itm/… Even though they look like open circuits to the multimeter connecting them has the same effect as when connecting to the source directly. Not sure how that's possible \$\endgroup\$
    – Alex P
    Commented Nov 16, 2022 at 14:13
  • \$\begingroup\$ @AlexP Even if you're trying to fix it with the transformer, it would be quite different focus to how your question was originally phrased. Creeping question focus is a bad practice here, because it constantly invalidates answers already written. So even about the transformers it is better kept in a new question (which should link to this one to show what is already known) \$\endgroup\$
    – tobalt
    Commented Nov 16, 2022 at 14:21

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.