6
\$\begingroup\$

I have been facing an issue that has been a mystery to me for quite some time, and I wanted to see if someone here can potentially give me more insights on this so I can understand why it is happening. Please bear with me since electrical engineering is not my specialty and I am doing my best to understand this phenomenon.

The issue I am facing is as follows:

I have two laptop chargers. One aftermarket charger from China and an official one from HP. Whenever I use the aftermarket charger from China, my KZ earphones (which are fairly low impedance) connected to the laptop would make electrical buzzing noises. The more power than the laptop draws, the more buzzing can be heard over the earphones. This issue does not happen with lower quality or higher impedance earphones.

As a reference, this was the sound that was heard over the earphones when the device was in use: https://drive.google.com/open?id=1OXwNHcq476hf_FxVr9n_bPVIcNwdNnzm

The Chinese charger is a Class II power supply which lacks a grounding pin. If I shunt 0V on the DC side to AC ground, I get a reading of 59.8V @ 78uA AC on my multimeter. As soon as the shunt is connected, the noise stops. Likewise, the noise stops if I laid my hand on the laptop external chassis (which I suppose had a similar grounding effect).

Thinking that the problem could be due to some kind of leaking AC current, I repeated the test on my HP power supply which was grounded. However, this time I also have lifted the ground prong using a cheater plug so that the leakage current cannot escape. There was no noise on the earphones line with this setup. That said, I was still able to measure 54.1V @ 54.1 uA AC from the ground pin to ground if a shunt was connected in that manner.

To see if the manufacturer would know something about this, I contacted them about the issue and they suggested that I remove the two Y capacitors on CY1 and CY2, or downsize them from 400V 100PF to 47PF to possibly mitigate the noise issue.

Not specifically knowing how this works, here are my questions:

a) What creates the 60V AC voltage on the isolated DC side? Is this caused by the safety capacitors used in the power supply? Why is it specifically half the line voltage?

b) What noise am I specifically hearing over the earphones? I know that it has something to do with the impressed AC voltage on the DC side since the noise is gone as soon as I shunt the 0V side to AC ground but wasn't sure specifically what I was hearing? I am guessing that the AC current is trying to find its way through the ground and parts of that current are flowing through my low impedance earphones and be interpreted as sound...

c) From what the manufacturer is suggesting, it sounds like that they may be shunting the parasitic capacitance from the transformer to the 0V output side and using it as ground. I am assuming the rationale would be that a smaller capacitor would result in smaller leakage current and hence less noise being picked up by the earphones (and perhaps less EMI compliant). Would that be a good assumption? Why did they need 2 capacitors? Is it due to polarity?

d) Would the placement of the Y capacitors matter in this setup? Would it be better to simply place them between L-N and 0V-N so the noise can be diverted from the connected device (laptop)?

e) Why did the HP charger not create this same noise even when it was ungrounded? Could it be due to the positioning of such capacitors or some kind of circuitry design difference? If so, is there a way I can confirm without cracking the charger open?

e) Without knowing anything about the schematic, would there be a better setup than using Y capacitors in reducing EMI? Would the capacitor values suggested by the Chinese manufacturer be a good value to prevent EMI emissions?

f) If I were to view this under an oscilloscope, which two points should I be using to confirm this issue and visualize my findings?

Thank you for your time!

\$\endgroup\$
1
  • \$\begingroup\$ different chargers will oscillate at different frequencies, and have different internal waveforms and different amounts of output ripple; your headphone amplifier's VDD filtering may or may not resonate at the charger's ripple frequency. \$\endgroup\$ May 15 '19 at 3:38
1
\$\begingroup\$

Yes, an unfortunately common problem.

The Y caps are indeed the source of the buzz. I have even modelled their influence in the past, using circuit analysis software. Moving them around (placement) won't help, as it's not the issue.

Short answer, you're taking a risk with third party power supplies.

Some horrible things go on in SMPS's. The electrical 'noise' generated has to be 'dumped' to neutral or ground to pass EMI regulations, This means some current flows too. This may be some of the buzz or it may simply be that your're listening to a small amount of rectified 50/60 Hz.

Put simply, the likes of Hewlett Packard design their products 'better' so as to avoid the inconvenience.

A linear supply would never have this problem as it doesn't need the Y caps!

\$\endgroup\$
3
  • \$\begingroup\$ Sadly it IS the real answer ! You can try to analyse it forever, indeed I have done so almost to death. I could almost write a treatise on earth leakage currents but no amount of navel gazing solves the problem. The awful truth is that the EMI caps in SMPS's cause leakage currents that develop a voltage when they find a home to ground through a wire that has resistance. That voltage (or current) causes the typical 'buzz' of rectified 50/60 Hz line that's a result of the ripple on the SMPS's storage cap. Actually much pro-audio equipment still uses linear supplies, hence the mention. \$\endgroup\$ Oct 6 '20 at 3:35
  • \$\begingroup\$ ...but essentially no consumer audio equipment still does. And computers haven't in generations. So actually, the engineering problem does have solutions. The user's problem on the other hand, is one of product usage, and so not really on topic here, which tends to mean it may not be answerable anyway. Irregardless of all of that, this site has rules about what constitute an answer, and mere commentary does not meet those. An unanswerable question does not justify a non-answer, it rather points to the need for closure. \$\endgroup\$ Oct 6 '20 at 3:38
  • \$\begingroup\$ Absolutely no computers do anymore but linear supplies were still in use even in personal computers into the 1980's. To this day, much low power audio equipment still uses linear supplies exactly because of the 'buzz problem' Really honestly ! FWIW I'm familiar with the issues that certain studio equipment manufacturers such as Neve, experienced when transitioning to SMPS's sometime around 2010 (possibly later), precisely because of the problem caused by leakage currents. It tends only to be a problem when you can actually hear it. \$\endgroup\$ Oct 6 '20 at 3:47
0
\$\begingroup\$

This is a very common issue.

To simulate your hand connect 1 nF from 0V to Earth ground such as on VGA jack screws or USB shield. COnnecting VGA cable to monitor will make a direct connection to earth ground. The cap just shunts the SMPS leakage current to earth ground better than the 0.25mA X cap filter.

\$\endgroup\$
6
  • \$\begingroup\$ Thank you... although what I am more curious is why this specific ungrounded charger is having this issue whereas no other ungrounded chargers I currently have is experiencing this problem. Just curious whether this was a design flaw versus some other issue... \$\endgroup\$
    – user235909
    May 15 '19 at 3:06
  • \$\begingroup\$ Often discontinuous mode DCM chargers require more expensive chokes. \$\endgroup\$ May 15 '19 at 3:15
  • \$\begingroup\$ When SMPS have high CM noise from feedthru capacitance of flyback transformer to floating DC output, it can result in differential noise from unbalanced impedances. Shunting this capacitively coupled noise to external low impedance reduces the intermodulation results of RF noise. It is a design flaw from insufficient CM noise suppression. \$\endgroup\$ May 17 '19 at 4:39
  • \$\begingroup\$ From a design perspective, what can the manufacturer do to correct this issue given an ungrounded charger? Would it be better to shunt this to neutral instead? \$\endgroup\$
    – user235909
    May 17 '19 at 5:47
  • \$\begingroup\$ Raise the switcher frequency and CM ferrite choke inductance to raise impedance and reduce current. Use lower leakage capacitance transformer \$\endgroup\$ May 17 '19 at 5:48
-1
\$\begingroup\$

Have you checked with a multimeter the groundings on a socket?

In my previous house, someone who had no idea and much less license to do the work, had caused the metals in the socket for grounding to produce variable about 50 V AC voltage 50 Hz.

I checked the casing for that, and got 180-200 V variation from the case. I was sure, there was some device in the house, which constantly is grounding somewhat. Like the lamp, which has one phase divided on to 1/230 x 3.

When I fixed those dangerous connections, the background hiss from speakers went almost away, mouse and other problems got fixed too. That is unlikely, but who knows?

Almost same kind of problem has been occurred again in newer house, and somehow it is related to sound card and which cables are connected. I almost lost hope then, but optical connections seemed to remove it.. Of course, it is not always possible. Clad, if I could help anyhow or woke some ideas.

\$\endgroup\$
1
  • \$\begingroup\$ Hi, Thanks for trying to help. However this answer does not address the original question. You ask whether the OP's household mains sockets are grounded. But the test results show that they must be grounded, as the test result changes when the (noisy) Class II PSU DC output is connected to the AC socket ground. It seems you are giving an anecdote regarding the use of PSUs which do need an AC mains ground, but one isn't connected. That cannot apply to the Class II PSU in this question! For the future, please make sure any answer you write is more than an anecdote. Also read the tour. Thanks. \$\endgroup\$
    – SamGibson
    Aug 16 at 14:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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