Please Help with ground loop problems - Pearl & MC stages
Firstly I am no expert in audio.
I have built the original Pearl Phono stage and the MC stages from the boards in the group buy thread (Promthieus). I realise this is not the thread but it seems to me that much brain power has been concentrated into the discussions of grounding, so I am hoping that you can help me diagnose where the problem is.
I have a terrible bad hum on both channels without a source connected.
My layout is a separate PSU with XLR terminated cable. I have one toroidal Trans feeding both psu boards.
From there they are connected by the cable to the Phono Stage. I have one common ground in the cable connected to the IEC connector Earth point, and the other end to the Chassis in the phono stage.
All RCA's in the phono stage are insulated from the chassis and meet at a star ground point.
This is where I might be doing something stupid
It seems to me the RCA's will all be grounded to the chassis anyway at the star point? Or should the star ground point also be insulated from the chassis
Anyway the hum is so bad I can barely listen to the stage, which is disappointing because when the sound level is high the detail and imaging is excellent.
Grateful of any help as I have spent many hours building and don't want to abandon the project.
Firstly I am no expert in audio.
I have built the original Pearl Phono stage and the MC stages from the boards in the group buy thread (Promthieus). I realise this is not the thread but it seems to me that much brain power has been concentrated into the discussions of grounding, so I am hoping that you can help me diagnose where the problem is.
I have a terrible bad hum on both channels without a source connected.
My layout is a separate PSU with XLR terminated cable. I have one toroidal Trans feeding both psu boards.
From there they are connected by the cable to the Phono Stage. I have one common ground in the cable connected to the IEC connector Earth point, and the other end to the Chassis in the phono stage.
All RCA's in the phono stage are insulated from the chassis and meet at a star ground point.
This is where I might be doing something stupid
It seems to me the RCA's will all be grounded to the chassis anyway at the star point? Or should the star ground point also be insulated from the chassis
Anyway the hum is so bad I can barely listen to the stage, which is disappointing because when the sound level is high the detail and imaging is excellent.
Grateful of any help as I have spent many hours building and don't want to abandon the project.
To specifically answer your question: The star ground should be connected to the chassis.
You have a choice where the star ground is located. The chassis must be connected to the mains safety ground close to the IEC connector. This point may be the system star ground. Alternately, if the star ground is located elsewhere, perhaps on a circuit board, then a wire should be installed between the star ground and the point where the safety ground connects to the chassis.
Grounding is important between your two chassis.
First, there must be a dedicated wire from the point where the safety ground connects to the power supply chassis through the connecting DC power cable to the chassis containing the phono circuitry. The connector(s) on the cable should be such that this safety ground makes first when the connector is inserted and breaks last when the connector is removed.
Second, each power common (ground) should have a separate wire in the cable, as each voltage should have a separate wire. That is, voltages and grounds for the two channels should be separately wired through the cable.
In the power supply design, be careful about circuit layout and where you pick up the voltage and common.
Be as careful on routing power and ground currents as you are for signal currents.
As this is an MC preamp, shielding and layout are very important as are cable and wire routing.
This is a simple overview, there is a lot more information in my article at http://www.raleighaudio.com/Audio Component Grounding and Interconnection.pdf
Dave Davenport
You have a choice where the star ground is located. The chassis must be connected to the mains safety ground close to the IEC connector. This point may be the system star ground. Alternately, if the star ground is located elsewhere, perhaps on a circuit board, then a wire should be installed between the star ground and the point where the safety ground connects to the chassis.
Grounding is important between your two chassis.
First, there must be a dedicated wire from the point where the safety ground connects to the power supply chassis through the connecting DC power cable to the chassis containing the phono circuitry. The connector(s) on the cable should be such that this safety ground makes first when the connector is inserted and breaks last when the connector is removed.
Second, each power common (ground) should have a separate wire in the cable, as each voltage should have a separate wire. That is, voltages and grounds for the two channels should be separately wired through the cable.
In the power supply design, be careful about circuit layout and where you pick up the voltage and common.
Be as careful on routing power and ground currents as you are for signal currents.
As this is an MC preamp, shielding and layout are very important as are cable and wire routing.
This is a simple overview, there is a lot more information in my article at http://www.raleighaudio.com/Audio Component Grounding and Interconnection.pdf
Dave Davenport
Thanks for the detailed reply.
I am going to start at the beginning, always a good place to start
I will check all my wiring in the PSU and then look at the power cable to ensure I have the correct safety and ground connections. Then I will recheck every connection in the Phono Stage.
You may not hear back from me for a while due to work commitments, but I will report back when I have checked it all.
Many thanks for making the requirements easy to understand.
I am going to start at the beginning, always a good place to start
I will check all my wiring in the PSU and then look at the power cable to ensure I have the correct safety and ground connections. Then I will recheck every connection in the Phono Stage.
You may not hear back from me for a while due to work commitments, but I will report back when I have checked it all.
Many thanks for making the requirements easy to understand.
When connecting a multi-channel amp to a computer, if you use the 1/8" stereo out, the grounds of the two signal wires are connected.
Does it automatically create a ground loop, no matter how your amplifier is wired?
Or does the noise that shows up when the cable is plugged in come from the computer itself?
I have been experimenting with different grounding schemes for over a week, and can't get rid of the noise that shows up when I plug my amp into my computer.
Does it automatically create a ground loop, no matter how your amplifier is wired?
Or does the noise that shows up when the cable is plugged in come from the computer itself?
I have been experimenting with different grounding schemes for over a week, and can't get rid of the noise that shows up when I plug my amp into my computer.
No. In the '70s and '80s audio gear from Germany usually came with DIN connectors that also shared the ground for both stereo channels. All manufacturers got such a setup working without audible ground-loops.
Maybe. Sometimes the computer's power supply is the culprit. You can check that with headphones. If there is no noise, recheck by connecting a different amplifier.
Try not to see your amplifier as an isolated device, but to include the PC into the grounding scheme.
Use the same outlet for PC and amplifier. That will make the loop through the PE leads smaller.
Check if the noise is reduced, when you turn either of the mains plugs by 180°.
Did you already try to insert a resistor (< 10 Ohm) between signal and power ground?
I made ground schematic like in an attachment, posted by someone in this or other grounding related thread. I can't remember who and where it was.
But this schematic is for one channel only and shows one PSU. I use four channels from sound card for 4ch amp and four PSUs. As I observed, if you connect multichannel chipamp with single PSU in parallel, as advised, you shouldn't have any ground related problems. However when you use two or more PSUs with independent transformers this could be a little bit an issue.
At first I had a big oscillations (I guess) with four PSUs, as my light bulb testers were shining when I power on the amp. When I tried to power the amp on for one second without light bulb testers, the amplifier was producing a strong enough buzz. This only applies when connect the amp to source.
After that I found out that Ground from sound card is not a GND as it should be (perhaps, because of I don't have Earth connection in my wall sockets). If you look at PC PSU you could see a PTC thermistors connected from lines to middle point of AC line socket. In my PC PSU has happened that these thermistors conduct to PC Ground 112VAC potential between PC Ground and Earth. Of course, I claimed about this to manufacturer and the answer was that this is a normal operation of PSU.
So, when you connect multiple independent PSUs to the amplifier you have separate ground layouts for every PSU. But source always has common Ground for every channel. Thus, your connection point for every separate ground layouts will be a source ground.
My solution was to connect between each other in series star ground points of every PSU. I have four PSUs with full bridge rectifiers and two transformers with double secondaries. Also to connect like that and have no hum from speakers you should have identical secondary voltage and regulation for all transformers and maybe matched rectifiers. If this rule neglected, you should expect a voltage difference between every transformer (PSU) middle points which will produce a hum. Also to minimize a little secondary variations I used a thick copper bar (on caps) and wire from bridges for PSU Ground and voltage traces, speaker returns, little thinner wire from signal ground points and thin wire for other signal ground layouts. After all of that I don't have audible buzz and hum from speakers, only a little bit when stay close to speakers, caused by variations in secondaries of my transformers. I have 0.5V transformers difference of rectified and smoothed voltage.
Regards.
Attachments
Last edited:
(Sorry to ask a stupid question - the answer may already be somewhere in the previous 40 pages!)
The mains earth connection is there to blow the fuse if the chassis becomes live, but cannot be extended to every conductor that emanates from the box. Why is there a presumption that your audio 0V (as opposed to all the other signal conductors) should be connected to mains earth at all, either directly or through some sort of decoupling device? Is it because you want the enclosure to constitute a Faraday shield and as such it cannot be isolated from the audio 0V?
Another stupid question: could someone come up with a neat inline safety-approved RCD that is moulded into the permanent mains cable, purchasable by builders of DIY equipment? Would that then allow much more freedom in the grounding schemes people wanted to use?
The mains earth connection is there to blow the fuse if the chassis becomes live, but cannot be extended to every conductor that emanates from the box. Why is there a presumption that your audio 0V (as opposed to all the other signal conductors) should be connected to mains earth at all, either directly or through some sort of decoupling device? Is it because you want the enclosure to constitute a Faraday shield and as such it cannot be isolated from the audio 0V?
Another stupid question: could someone come up with a neat inline safety-approved RCD that is moulded into the permanent mains cable, purchasable by builders of DIY equipment? Would that then allow much more freedom in the grounding schemes people wanted to use?
That is the effect of what you describe, but it's not the intent.
The PE is a low resistance path back to the distribution board.
With chassis connected to PE there is low resistance from chassis to distribution board.
Now connect the Live to chassis.
The voltage on the chassis must be below a specified voltage. This voltage is assumed to be not life threatening. I can't recall, but 50Vac springs to mind.
The resistance of the Live feed and the resistance of the PE return act as a resistive ladder and the respective resistances determine the voltage that the chassis ends up at.
The primary purpose of the PE is to limit the "Live" chassis to a safer maximum voltage.
While the fault is ongoing the fuse is passing Fault current. The resistances in that Fault loop must be low enough to allow the fuse to rupture quickly. I can't recall how this is specified in Electrical codes.
But one can see that the lower the resistances the quicker the fuse will blow. That is part of the reason we as operators are not allowed to interfere with the PE nor the connection from Chassis to PE.
And why we are NOT ALLOWED to insert anything between Chassis and PE that could increase the resistance of that route.
Andrew, thanks for that. OK, my mistake; the PE's primary purpose is not to blow the fuse but to keep the chassis below a certain potential (but it usually blows the fuse as a result).
But my question is, why the compulsion to connect our transformer-isolated audio equipment to the earthed metal box to earth at all? By all means earth the box to make it safe, but why connect the audio 0V to it if it's only going to cause problems with hum loops etc.?
If I was building some non-audio piece of equipment in an earthed metal box, say a control unit for a CCTV camera, I wouldn't feel it necessary to connect any of the emerging wires to earth would I? What's the difference between this and an audio power amplifier?
But my question is, why the compulsion to connect our transformer-isolated audio equipment to the earthed metal box to earth at all? By all means earth the box to make it safe, but why connect the audio 0V to it if it's only going to cause problems with hum loops etc.?
If I was building some non-audio piece of equipment in an earthed metal box, say a control unit for a CCTV camera, I wouldn't feel it necessary to connect any of the emerging wires to earth would I? What's the difference between this and an audio power amplifier?
you are onto the problem.
The grounded exposed parts are all connected either directly or indirectly to the Audio Ground.
Connecting Audio Ground to Chassis satisfies the grounded exposed parts. This connection must survive long enough to allow the fuse to rupture and the arc to extinguish.
The ungrounded exposed parts are a different animal.
The hot of an RCA is not exposed. The hot of a shrouded speaker terminal is not exposed. But when you connect cables to these the ungrounded parts can become exposed.
The exposed ungrounded terminals are potentially a hazard.
What do we do about this hazard?
We can just rely on good fortune and hope that any or some of three scenario prevail:
1.) that the mains Live never touches/conducts to any ungrounded internal part.
2.) that there is sufficient resistance between the mains Live and the exposed conductive part.
3.) that the ungrounded circuit/part shorts to Chassis when exposed to 115Vac/230Vac.
I don't know how to design and test to eliminate this hazard. I cannot recall discussion of this in the Forum.
What I can do is my best to eliminate the avoidable that I am in control of.
The grounded exposed parts are all connected either directly or indirectly to the Audio Ground.
Connecting Audio Ground to Chassis satisfies the grounded exposed parts. This connection must survive long enough to allow the fuse to rupture and the arc to extinguish.
The ungrounded exposed parts are a different animal.
The hot of an RCA is not exposed. The hot of a shrouded speaker terminal is not exposed. But when you connect cables to these the ungrounded parts can become exposed.
The exposed ungrounded terminals are potentially a hazard.
What do we do about this hazard?
We can just rely on good fortune and hope that any or some of three scenario prevail:
1.) that the mains Live never touches/conducts to any ungrounded internal part.
2.) that there is sufficient resistance between the mains Live and the exposed conductive part.
3.) that the ungrounded circuit/part shorts to Chassis when exposed to 115Vac/230Vac.
I don't know how to design and test to eliminate this hazard. I cannot recall discussion of this in the Forum.
What I can do is my best to eliminate the avoidable that I am in control of.
Thanks Andrew.
So the answer to my question is that commonly used audio connectors such as phono sockets have exposed metal shells, so that is why they must be grounded robustly. If so, then all that is needed to banish ground loops forever, is to use fully-shrouded connectors with no exposed metal parts. Then the audio circuits can remain fully isolated from the mains. Could it really be so simple?
"Connecting Audio Ground to Chassis satisfies the grounded exposed parts. This connection must survive long enough to allow the fuse to rupture and the arc to extinguish."
Another question is raised in my mind. You seem to be suggesting that any exposed signal ground connection should be as stout as the mains earth connection. I don't believe many constructors ever do that; they always fit a hefty great tag for the main earth, but cable screens and bonds to phono sockets etc. are of arbitrarily skinny cable with soldered joints of doubtful mechanical strength. I don't think current rating is usually one of the headline specs people look for in a screened cable. Maybe they should.
"I don't know how to design and test to eliminate this hazard. I cannot recall discussion of this in the Forum.
What I can do is my best to eliminate the avoidable that I am in control of. "
As long as there are electrical inter-connections (exposed or not - you can always drop a shrouded connector in a pool of water by accident) you can't eliminate the hazard, but I'm not going to lose sleep over it - these so-called double-insulated pieces of equipment are nothing special inside (if I get the chance I'll take a picture of a Japanese amplifier I own with a row of switched outlets on the back panel with live mains pins connected with a 'bus bar' of exposed thin copper wire just below the ventilation grille!)
It seems to me that as long as you earth the box, physically keep the mains section away from the transformer-isolated electronics and, possibly most important of all, ensure that cables are physically restrained from touching anything they shouldn't even if the electrical joint breaks, you will be fine. (We do place a lot of trust in isolating transformers don't we..?)
What are your views on RCDs? What if they were built into the cable?
The internals of electrical equipment are not the only places that a stray mains wire could touch something it shouldn't. If you think about it, every large metal object has the potential to make a very small area of contact with "live" (caused by a damaged cable, or a pool of water around a trailing mains block etc.) and to then to present a very large, low impedance connection to the next human who touches it. Kitchen sinks and bathroom taps are already protected against unpleasant potentials, but maybe large unearthed metal objects in general should be banned from the home!
So the answer to my question is that commonly used audio connectors such as phono sockets have exposed metal shells, so that is why they must be grounded robustly. If so, then all that is needed to banish ground loops forever, is to use fully-shrouded connectors with no exposed metal parts. Then the audio circuits can remain fully isolated from the mains. Could it really be so simple?
"Connecting Audio Ground to Chassis satisfies the grounded exposed parts. This connection must survive long enough to allow the fuse to rupture and the arc to extinguish."
Another question is raised in my mind. You seem to be suggesting that any exposed signal ground connection should be as stout as the mains earth connection. I don't believe many constructors ever do that; they always fit a hefty great tag for the main earth, but cable screens and bonds to phono sockets etc. are of arbitrarily skinny cable with soldered joints of doubtful mechanical strength. I don't think current rating is usually one of the headline specs people look for in a screened cable. Maybe they should.
"I don't know how to design and test to eliminate this hazard. I cannot recall discussion of this in the Forum.
What I can do is my best to eliminate the avoidable that I am in control of. "
As long as there are electrical inter-connections (exposed or not - you can always drop a shrouded connector in a pool of water by accident) you can't eliminate the hazard, but I'm not going to lose sleep over it - these so-called double-insulated pieces of equipment are nothing special inside (if I get the chance I'll take a picture of a Japanese amplifier I own with a row of switched outlets on the back panel with live mains pins connected with a 'bus bar' of exposed thin copper wire just below the ventilation grille!)
It seems to me that as long as you earth the box, physically keep the mains section away from the transformer-isolated electronics and, possibly most important of all, ensure that cables are physically restrained from touching anything they shouldn't even if the electrical joint breaks, you will be fine. (We do place a lot of trust in isolating transformers don't we..?)
What are your views on RCDs? What if they were built into the cable?
The internals of electrical equipment are not the only places that a stray mains wire could touch something it shouldn't. If you think about it, every large metal object has the potential to make a very small area of contact with "live" (caused by a damaged cable, or a pool of water around a trailing mains block etc.) and to then to present a very large, low impedance connection to the next human who touches it. Kitchen sinks and bathroom taps are already protected against unpleasant potentials, but maybe large unearthed metal objects in general should be banned from the home!