Author Topic: Ground loop problems  (Read 2101 times)

0 Members and 1 Guest are viewing this topic.

Offline ratataxTopic starter

  • Regular Contributor
  • *
  • Posts: 134
  • Country: fr
Ground loop problems
« on: October 12, 2019, 11:24:41 am »
Hi,
So i'm designing an audio device with the chassis connected to GND, it runs on DC 12V via a wall adapter. I run into a quite annoying problem when it's connected to other grounded audio equipment, an amplifier for example :



Using a ground-less power supply solves the ground loop problem but adds a new one : the chassis of the device now as AC current in it, because of the Y safetey capacitor in the power supply connecting the high/low side together. All certified psu's have it and those caps have a leakage current of a few uA (and a hundred volts), if the PSU isn't earthed this AC current flows directly into the device's ground.

I feel the proper way of doing this is to keep the earthed PSU, but modify/add something in the design to prevent the ground loop from taking place... But how would you do that ?
I'm searching for a cheap and easy solution, because I haven't anticipated that in my initial BOM cost :/
« Last Edit: October 12, 2019, 11:27:06 am by ratatax »
 

Offline ratataxTopic starter

  • Regular Contributor
  • *
  • Posts: 134
  • Country: fr
Re: Ground loop problems
« Reply #1 on: October 12, 2019, 02:08:33 pm »
I tested several switching power supplies and I have one that isn't earthed, and has no AC current in its output GND. It would be perfect for my use, but I don't understand why it behaves differently compared to my other "leaky" SMPS. It also has an Y safety capacitor between the high and low voltage sides.

The capacitor, the transformer and optocoupler are the only things connecting the high/low sides on all my power supplies... Maybe it has something to do with the transformer ? The capacitor itself ? How I can select a power supply that will have no AC leakage ?
 

Offline jhpadjustable

  • Frequent Contributor
  • **
  • Posts: 295
  • Country: us
  • Salt 'n' pepper beard
Re: Ground loop problems
« Reply #2 on: October 12, 2019, 02:09:35 pm »
You're making obsolete assumptions about what's inside the green box. Let me update it for you:

]

Wolfson app note WAN0131 (attached) explains the rationale and illustrates the problem, but offers little in the way of solutions.

Moral of the story:
  • Reconfigure your DAC for cap-coupled output and add output capacitors, or choose a more suitable ADC.
  • Don't directly connect headphone outputs to anything but headphones and expect it to "just work."
  • Always expect Wun Hung Lo breakout boards to be designed for the cheapest, most common case, with no regard to uncommon cases.
"There are more things in heaven and earth, Arduino, than are dreamt of in your philosophy."
 

Offline RandallMcRee

  • Frequent Contributor
  • **
  • Posts: 544
  • Country: us
Re: Ground loop problems
« Reply #3 on: October 12, 2019, 04:11:46 pm »
See this thread on diyaudio, Nelson Pass reply #8:
https://www.diyaudio.com/forums/pass-labs/1819-grounding-schemes.html

Quote
I have two favorite "ground loop breakers", a power
Thermistor and a rectifier bridge.

The Thermistor works by being an ordinary 10 ohms
or so, but in the event that a lot of current goes through
the ground circuit, it reduces to a much smaller value. A
tip of the hat to Frank De Luca for that invention.

The rectifier bridge will conduct a lot of current when you
go over .7 volts, enough of a barrier for noise, but a good
conductor in event of failure. One advantage to the
rectifier bridge approach is that it can isolate to channels
of ground both from each other and earth at once.

You will see the use of the Thermistor in most of my
DIY amp articles, and the rectifier bridge on the
Zen lite article.
A google search will show a plethora of great articles and illustrate the grounding methods. Copyrighted so I can't post here.
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: Ground loop problems
« Reply #4 on: October 12, 2019, 04:49:27 pm »
I tested several switching power supplies and I have one that isn't earthed, and has no AC current in its output GND. It would be perfect for my use, but I don't understand why it behaves differently compared to my other "leaky" SMPS. It also has an Y safety capacitor between the high and low voltage sides.
Perhaps a lower value capacitor, they aren't all the same. Of course you can take those bricks apart and investigate.
A class 2 linear PSU will have none of those problems, but price and bulk instead.

I'm not sure about the 5V supplies in phone chargers. They use primary-side regulation and the secondary is 100% passive, just a diode and capacitor. I think they may omit the Y capacitor. I would have no idea how to order similar supplies made for 12V.
 

Offline Someone

  • Super Contributor
  • ***
  • Posts: 5155
  • Country: au
    • send complaints here
Re: Ground loop problems
« Reply #5 on: October 13, 2019, 03:59:59 am »
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: Ground loop problems
« Reply #6 on: October 13, 2019, 08:16:46 am »
From the first PDF it seems like only common mode choke could possibly be acceptable and I imagine a rather large one would be required to be effective at 60Hz. I haven't skimmed the others but they probably offer similar advice. Maybe really just get a better PSU in the first place? ;)
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Ground loop problems
« Reply #7 on: October 13, 2019, 02:20:33 pm »
If your problem is solved with a PS which is not earthed, ideally use a medical-grade PS then. It will both NOT be connected to earth AND be safe to use.
They are not cheap but not overly expensive either. For instance: https://www.xppower.com/Products/AC-DC-Power-Supplies/Medical
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf