connecting audio ground to Safety Earth

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,
direct connection of audio ground to Safety Earth can give rise to a grounding loop causing hum.

Connect the safety Earth to the audio ground through a disconnecting network to break that loop.

But is it safe?

The main purpose of the Safety Earth is to ensure the chassis and any exposed conductive parts can never electrocute anyone if the Live side comes into contact with any internal parts.

I advocate a disconnecting network between the Safety Earth and Audio Ground that consists of a 25A or 35A bridge rectifier with optional 10r power resistor, power thermistor and/or ceramic cap all working in parallel.http://sound.westhost.com/earthing.htm fig3
The bridge rectifier must be able to pass fault current (>=1kA) to earth until the mains fuse ruptures.

But does it work?
http://www.diyaudio.com/forums/showthread.php?postid=1357794#post1357794

Comments?

Will 6A diodes and T8A fusing pass this test?
Who wants to join the testing panel for 110/120Vac?
 
Earthing system

Hi Andrew,
AndrewT said:
...direct connection of audio ground to Safety Earth can give rise to a grounding loop causing hum...
I agree with you, this is a common idea.

One year ago I made this try. Examining my Marantz CD 94 pcb, I saw that designer allow more than one ground connexion to chassis. Common product, and Marantz CD 94 was, use only one connexion between electronic ground and chassis. I try to connect this second ground to chassis. Listening test were very good. Sound quality increase, more bass, less fog, more precision... First good point.

Marantz CD 94 use a two pole main plug (without heart). I extent my test and link chassis to earth. Listening test give better sound too. Second good point.

I renew test to plug CD player on earth on other systems (friend's systems) and obtain always same conclusion. Earthing devices increase sound quality. Examining system connexion, with multiple earth, you create multiple ground loop. Exactly the opposite you are testing!

In fact, each time you connect a source to earth, you win, even if this source in not powered on (just link to the overs).

Bill Whitlock's Student match with this test. He gives good description about unwanted current flowing across audio cable. I don't try to use his isolator transformer. Earthing material reduce parasitics current flowing across audio cable.

Eric
 
Re: Earthing system

Eric Juaneda said:

Marantz CD 94 use a two pole main plug (without heart). I extent my test and link chassis to earth. Listening test give better sound too. Second good point.
I think you are saying that you modified your Double Insulated equipment by making a non standard connection to chassis.

If I have interpreted your statement correctly, then you would be breaking the law and exposing yourself and other users to danger.
I suspect your EC consumer law mirrors ours.

If you are breaking the law then posting this information here also contravenes the Forum rules.
 
Re: Re: Earthing system

Hi Andrew,

AndrewT said:
I think you are saying that you modified your Double Insulated equipment by making a non standard connection to chassis.

If I have interpreted your statement correctly, then you would be breaking the law and exposing yourself and other users to danger.
I suspect your EC consumer law mirrors ours.

If you are breaking the law then posting this information here also contravenes the Forum rules.

I modified my double insulated equipment by adding a connection to Safety Earth. I add a double security!

Eric
 
AndrewT said:
Hi,
direct connection of audio ground to Safety Earth can give rise to a grounding loop causing hum.

Connect the safety Earth to the audio ground through a disconnecting network to break that loop.

But is it safe?

The main purpose of the Safety Earth is to ensure the chassis and any exposed conductive parts can never electrocute anyone if the Live side comes into contact with any internal parts.

I advocate a disconnecting network between the Safety Earth and Audio Ground that consists of a 25A or 35A bridge rectifier with optional 10r power resistor, power thermistor and/or ceramic cap all working in parallel.http://sound.westhost.com/earthing.htm fig3
The bridge rectifier must be able to pass fault current (>=1kA) to earth until the mains fuse ruptures.

But does it work?
http://www.diyaudio.com/forums/showthread.php?postid=1357794#post1357794

Comments?

Will 6A diodes and T8A fusing pass this test?
Who wants to join the testing panel for 110/120Vac?


As I understand you tested the safety aspect of this connection and it works.

Let me add two remarks that could improve the otherwise good approach.

First from an EMI point of view:
The incoming shield should be connected to the chassis at one end AND should be connected to the chassis via hf capacitors AT the entrance of the other end. A classical way is to provide a so called clean ground which is a conductive strip connected to the chassis ground and running at the entrance of incoming lines. The shields or the decoupling and filtering caps should be connected to this clean ground.
In this case, the noisy currents picked up by the shield will be diverted to earth via the chassis without flowing in the audio signal ground. This is what will unfortunately happen in the example shown by your reference.

If the hum voltage is higher than a few volts the diodes will operate and the loop is still existing. Then you cannot avoid the use of balancing techniques.

In any case, it is better to use shielded twisted pair even with unbalanced connection and connect the shield as described.


Second, inrush current

The use of ground loop breaker would require fast acting primary AC fuses which is in contradiction with high surge current in case of toroidal transformer.
A good solution is the use of a special AC breaker that controls the magnetisation of the transformer by sending pulses at switch on avoiding the high surge. see http://www.emeko.de/index.php?id=29&L=1

With this device fast fuses are OK.


Cheers

JPV
 
Earth Leakage & RCD/RCCB

Hi,
The RCCB trigger value is usually available in 10mA, 30mA and 100mA.
The 30mA is normally adopted for domestic use.
It is also usual to limit the normal leakage current to less than 25% of the RCCB rating.
This imposes a 7.5mA total leakage current for all appliances on an RCCB protected circuit.
This is equivalent to a total Line to Safety Earth capacitance of about 90nF for 220/240Vac 50Hz supplies.
 
Andrew - nice test, thanks for trying that.

As for the last point you make - 7.5mA is still a lot of leakage current and represents a serious shock hazard. Which is why Y-caps (phase to earth) are usually limited to 4n7 in input filters to ensure that, if the earthbond is lost on a given piece of equipment, the leakage current is held to a relatively safe value.

4n7 at 240vac gives about 0.5ma - a good 'sting' but you'll live. 1ma and above becomes dangerous IIRC.
 
I have both channels of my power amp connected to earthed chassis with absolutely no hum. In fact, my subwoofer and woofer cones are absolutely still. I cannot feel any motion at all.

I do use input transformers to break the ground loop via my preamp, so that aspect of hum generation is not there. However, it does show that the mere fact of earthing both channels of the amp to safety ground is not a problem in and of itself.

I have earthed it in the same way Self describes in the fourth edition of his handbook, Figure 16.1, page 438. In it, he shows the input socket connected to chassis/mains earth. He does not have the mains earth connected to star ground. It is basically a star-on-star type of grounding scheme.

If I recall correctly, for maximum shielding effectiveness, the reference of the circuit should be the same as the shield.

Therefore, if you are getting hum in this scheme when input transformers are not used, then the ground loop is more likely caused from the loop back through the preamp, and not within the amp itself. I haven't tried this scheme without input transformers, but in theory it keeps the loop current out of the ground path of the amplifiers, per se, and is similar to Leach's "fix" for hum problems.
 
Therefore, if you are getting hum in this scheme when input transformers are not used, then the ground loop is more likely caused from the loop back through the preamp, and not within the amp itself.

That is the usual problem when using grounded equipment - the ground loops through the power cords of connected equipment. When using transformers and/or good balanced inputs this problem disappears. And as you say, it is good from a shielding viewpoint to have the chassis at circuit gnd potential. But it will usually be good enough when using a isolation circuit.

My own experience is that grounded equipment without ground lift circuits can be made to work even with unbalanced signal interconnections as long as all plugs are connected to the same outlet/power strip and loop area is minimized. Trying to use different outlets for mixing desk and power amps for live sound will usually not give acceptable hum when the interconnects are unbalanced. At one place this seemed to be working, but when the lights were dimmed later there was an evil buzz. I usually run an extension cord alongside the snake for the power to the mixing desk to fix these kinds of problems. This minimizes the loop area and injected hum.

This should be applicable to HiFi-equipment too. Interconnection ground loops can be made insignificant through minimizing loop area and making sure the chassis is grounded to a clean ground - not the transformer center tap or speaker return. The pdf Eric Juaneda posted shows this, both the right and wrong way to do it.
 
megajocke said:

That is the usual problem when using grounded equipment - the ground loops through the power cords of connected equipment. When using transformers and/or good balanced inputs this problem disappears. And as you say, it is good from a shielding viewpoint to have the chassis at circuit gnd potential. But it will usually be good enough when using a isolation circuit.

This is a different scenario than the one I referred to. You are talking about both amps connected to mains ground. There is definitely a loop there. However, source equipment is usually not grounded to the mains ground. If it is, it is important, as in all loops, to minimize its amplification by keeping the loop current directed away from the amplifier. But Self's arrangement may help in this scenario, also

What I was referring to was the ground loop around the stereo pair of interconnects. The connection arrangement circulates the ground loop current in the interconnect shields, and not through the amplifier to star ground point of amplifier and star ground point of PS caps.
 
Yes, but the loop around the interconnects can be made small with proper design of amplifier and source. The jacks should be connected to each other where they enter the amplifier and this used as the ground reference. Other internal ground configurations are prone to pick up transformer stray fields. These loops don't depend on mains ground though and can exist even in non-grounded reinforced insulation equipment. I have never had it cause a problem in production equipment. What's your experience?

I have a home-built amplifier built from kits that someone else has built though, and it does suffer from the loop you are talking about. The input jacks are insulated from the case and connect with screened lead to their respective amplifier board on each side of the case. The grounds meet at a bolt on the case with the capacitor common and transformer center tap. If anything is connected to both inputs it starts humming due to the internal loop picking up stray field from the transformer and voltage drop from capacitor charging current over the star ground bolt.
 
I should qualify the scenario I mentioned in some posts, above. I use dual power supplies, i.e., one for each channel. Therefore, connection of input grounds to chassis/safety ground is at one point, only. If using only one PS for both channels, things would be a lot different. In that case, the power supply grounds for each channel would both be connected together at the common power supply. This would create a ground loop internal to the amp. Therefore, if only one common PS is used, I think the star should be at the PS cap common connection. The safety ground should connect to this star, and the input jacks should float from the chassis and the input ground should connect to the star, assuming it is separated from the board's dirty ground. If hum is created when the preamp is hooked up, one of the input grounds should be disconnected from the star ground, and the grounds of the input jacks connected together. Bottom line, it is easier with dual power supplies.
 
I've been following this thread, and have also done some searching re. grounding.

So, using my Hafler DH-500 as a specific example, my inputs are currently floating, the ground tabs are connected to each other, and a common ground lead runs from this common connection to the star ground at the filter capacitors. That is what I have now.

If I want to add a safety earth ground to chassis with a three pronged plug, then are you saying that I would simply run the safety earth connection solely to the same star ground, with no other safety ground connection to the chassis? If so, then how is the chassis grounded to safety earth? Transformer center tap?

Whew. This is tough stuff.

Thanks.

Pete
 
All exposed conductive parts must be permanently connected to Safety Earth.
In all situations the third wire (safety wire) in the supply cable is permanently connected, either welded or bolted, to the metal chassis.

The difficulty comes when you have input or output terminals that have exposed metal parts. These too must be connected to Safety Earth. That's what post1 is trying to evaluate.
 
Double insulated items are where there is no earth in the supply cable to a piece of equipment. Thus there is no earth on the primarry side of a transformer. The secondary side of the transformer will be centre-tapped providing a plus/minus voltage swing around earth. The theory then goes that if there is a fault in the equipment then the applied voltage under fault condition is halved, i.e if the secondary supplies 24 volt centre-tapped then the voltage received during fault would be 12 volt. The circuit should then contain the fault and send it to the centre-tap earth. As there is no electrical connection in the transformer then no fault current needs to flow to the supply earth (your plug/socket on the wall).

Primary Secondary
------ ----- +12 volt
) (
) (
240 v ) -------- C.Tapped earth
) (
) (
------ ----- -12volt



The diagram hasn't come out properly but I hope this helps.

Gareth
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.