Ground Loops

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I don't know if you've heard of the UKs Compliance Club..

Its a journal of EMC/EMI related material ,with many contributions over the years by Keith Armstrong and other such luminarys. A must have resource.

There are over 100 originally printed issues that you can download in pdf form , and earthing,grounding and shielding of PCBs and ancilliaries are its bread and butter.

EMC Information Centre

You may have to register to gain access , but its free and they do not spam given email addresses.

Rob
 
www.hifisonix.com
Joined 2003
Paid Member
Firstly, have you checked that your input sockets are not shorting to the chassis (disconnect the input wires and measure).

Both the 0V and the PE must connect at one place only on the chassis

For the small signal stuff you have two options

1. Place a 10 ohm resistor between the signal ground and the power ground on the PCB. The input signal ground connects onto the PCB ground only.

2. The second option it to leave the 10 ohm resistor in place, but run a connection from the input socket ground to the star ground on your PCB.

These techniques usually solve most problems. However, not all consumer equipment is faultless.
 
Last edited:
I don't know if you've heard of the UKs Compliance Club..

Its a journal of EMC/EMI related material ,with many contributions over the years by Keith Armstrong and other such luminarys. A must have resource.

There are over 100 originally printed issues that you can download in pdf form , and earthing,grounding and shielding of PCBs and ancilliaries are its bread and butter.

EMC Information Centre

You may have to register to gain access , but its free and they do not spam given email addresses.

Rob

Wow, LOTS of good stuff here. They suggest connecting a CM choke just like I described. Reading through the articles will help to give you a sense of what the dangers of EMI are, what is susceptible and how to immunize your circuits/appliances.
 
Source ground is the ground at your source's output. Signal ground is the ground at your amp's input. This is for unbalanced connections. Balanced connections would need a 3-winding CM choke, but they are expensive, exotic and it's arguable whether they would improve a balanced connection.
 
a ground loop is formed when two points that are supposed to be zero volts actually have a voltage difference between them or that there is an opportunity for loop current to flow due to a complete circuit....

so look at your construction and see it this is possible in your build and then act accordingly....
 
Could a ferrite toroid do this duty (post20) as effectively?

Plenty of CM chokes are a ferrite toroid with two windings. If you mean the bare toroids intended for blocking EMI on cables, then probably, but I don't think it would be easy to get a high enough inductance. The 60Hz inductance needs to be high enough at 60Hz to break fault currents through the coil so that the magnetizing current isn't large enough to saturate the ferrite.
 
I'm baffled.

If you build a solid amplifier with a star 0V it will work flawlessly.

I've also got a pre-amp with a solid star 0V and it works flawlessly.

Connect the two together and you invariably end up with a loop.

I've got a commercial CD player which I'm not going to modify.

I thought breaking the loops in the pre might do the trick but it hasn't.

The question is where to break the loops.
It's not the star "0v" that's the solution or the problem. Needless to say, it's more complicated than a star ground.

It's other things (capacitive coupling between power transformers, and actual capacitors across the line, neutral and ground connections where the power cord comes into a unit, for two things) that make "ground" on Component A not the same as "ground" on Component B.

This guy (Bill Whitlock of Jensen Transformers) explains it really well, after taking several pages to explain why you do NOT want to use a "ground lift" thing on anything. He also sells audio isolation transformers which indeed solve the problem, though they're not the only (legal) way:
http://www.bennettprescott.com/downloads/grounding_tutorial.pdf
There was a link to a "blue wire" connection earlier in the thread, and there's also circuits where that "blue wire" is replaced with a 10 ohm resistor. This resistor drops most of the voltage difference between two components' "grounds" and allows the input to effectively function as a balanced input which cancels out the difference between the two chassis.

Here it is, post #57 of that same thread:
http://www.diyaudio.com/forums/soli...-lin-topology-nfb-tappings-2.html#post1629287

Here's your homework: Look over all this and tell us why that 10 ohm resistor fixes the problem.
 
That's a good PDF, although I think the OP solved their problem long ago.

Using a CMC as a signal isolator like I showed has the advantage that it doesn't cause harmonic distortion if the transformer isn't highly engineered and expensive. This option is conspicuously missing in that presentation.
 
When we design gear, we build a product the circuit and connectors are 100% isolated from the case. No ground currents are to flow through the case at all. If you are making your star ground connection to the chassis, this is incorrect. Then between the circuit's star ground point and the chassis ground point we putting in a 100 ohm 1/2 resistor. This will force the ground current to only flow along your audio shields and not through AC ground. Sometimes we put a 0.01uF 50V ceramic capacitor across the resistor for improved RF shielding.

The only connections to the case are to be the AC safety ground and the one single connection to your circuit's star ground. If you are using connectors that ground to your case anywhere, these must be isolated.
 
I use a CL-60 thermistor from the circuit ground to case. Or a 10 ohm 10 watt resistor.
cheers!
Are you recommending a tried and TESTED Safety device?
Or just guessing from other non tested sch you see scattered around the web?

I raised this recently and someone came back something along the lines of: "why don't I (AndrewT) trust what N.Pass puts into his sch?"

I don't care what PASS does, I want anyone using my equipment to remain safe. That is what is important to me and my visitors/relations.
 
I am not recommending anything. I am posting what I use, in the event that someone might comment.

In fact, I made an error in the description of the ground loop breaker -- it does use a bridge as Nelson Pass or a pair of diodes.

I would love to know how to test said device. I do not want to suggest anything or guess. I would like to have a safe way to do it.
 
I asked for and eventually did the TEST some years ago.
I reported the results of the TEST and have recently referred back to that TEST.
I have also pointed out that I am unaware of any other Member reporting the testing of these Disconnecting Networks.

I don't call them "ground loop breakers". To me that means something else and adds to confusion.
 
testing a Disconnecting Network

http://www.diyaudio.com/forums/chip-amps/92673-yet-another-3886-gain-clone-3.html#post1357794

Now more than 7years since I did that and reported.
I am still here.

How many other concoctions have been proposed and not tested.
Are the Builders still here?
Just one missing in action Member, out of our 85 thousand Members is too many.

Note, a few posts earlier that neither Ian, nor myself, were ready to step up to the plate and be counted. I was that "wimp"
 
Last edited:
http://www.diyaudio.com/forums/chip-amps/92673-yet-another-3886-gain-clone-3.html#post1357794

Now more than 7years since I did that and reported.
I am still here.

How many other concoctions have been proposed and not tested.
Are the Builders still here?
Just one missing in action Member, out of our 85 thousand Members is too many.

Note, a few posts earlier that neither Ian, nor myself, were ready to step up to the plate and be counted. I was that "wimp"

I'm not the only one who physically tests things then. I built an amp and altered it to oscillate and blow up to see what would happen to my solid state speaker relays. Safety parts aren't much good if they aren't tested.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.