understanding star grounding

pacificblue said:

Connectors can be rated for the average current. They must not withstand the peak current. A 30 A rated connector would be extremely big, because it would have to accomodate 6 mm² wires.

An elegant, but expensive solution is the N6R series from Hirschmann (250 V, 10 A). They are industrial connectors that are used e. g. for servo motors, and are available from other manufacturers as well. If you don't have crimping tools, look for solderable versions.

Nice and small are sensor connectors M8 (30 V, 3 A) or M12 (250 V, 4 A). You can see them on the same site, and they are also available from other manufacturers.

ASR uses these (400-1000 V, 22-100 A ratings) in their Emitter. They are IP65 rated and normally used in the toughest industrial environments. Somewhat exaggerated for audio applications with highly debatable aesthetic value.


Yes.
How about this instead?

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CP-1036-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CP-1236-ND

The non-locking ones are rated at only 1 A (I'm assuming per pin) for some reason, while this one does twice as much. Amphenol makes one rated at 5 A but I don't want to pay 3x as much. The shield I can use as the earth, while I can use two pins each for +, 0, and - for 4 A each. Is this OK?

Or should I use the 8-pos version, with two pins each for +/- and 4 pins for 0?
 
Those DIN connectors were used instead of RCA connectors for line level signals in the past. Ratings go from 30 to 100 V and 1 to 4 A depending on manufacturer and type. The plastic-housed have poor mechanical stability, while the metal-housed demand good soldering skills to not result in shorts.

Some British manufacturers still use them for signals and Mission uses the 5-pole versions for their PSX-R power supply, so they should work.
 
pacificblue said:
You must provide an overload fuse for the transformer's secondary for safety reasons rated at the nominal output current slow-blow. You should provide fuses for the electronic components rated according to the datasheets specifications and fast-blow. Having separate fuses for each amplifier makes fault detection easier, but is not mandatory. It could however become mandatory to use them for wire protection, e. g. if the transformer secondary is fused with 20 A, but the wires to each amplifier channel are only capable of 10 A.
I'd like to use this:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PB1044-ND

One on the +, one on the -, none on the two 0, right?

Terrible drawing (forgot the earth on the connector):
 

Attachments

  • ps.png
    ps.png
    15.8 KB · Views: 1,359
Hi 454,

First, You probably don't need a Safety Loop Breaker Circuit at all - it is a remedial fix if you are connecting to a device that has a Pin 1 Problem.

However, if you do want to have a SLB circuit, its purpose is to isolate the safety ground (and the chassis) from the audio grounds. Therefore it should go between the system star ground and the safety ground. If you have a separate power supply chassis, and the system star ground is in the amplifier chassis (most probably) then the SLB circuit should be there and connected to the extension of the safety ground in that chassis.

Dave
 
David Davenport said:
Hi 454,

First, You probably don't need a Safety Loop Breaker Circuit at all - it is a remedial fix if you are connecting to a device that has a Pin 1 Problem.

However, if you do want to have a SLB circuit, its purpose is to isolate the safety ground (and the chassis) from the audio grounds. Therefore it should go between the system star ground and the safety ground. If you have a separate power supply chassis, and the system star ground is in the amplifier chassis (most probably) then the SLB circuit should be there and connected to the extension of the safety ground in that chassis.

Dave
Thanks, Dave. What's the practical difference between putting it in the PS chassis as my (and gni's) diagram shows and putting it in the amplifier chassis? I don't see much of a system/audio star ground in gni's diagram, so how would I implement that?

Anyway, is the fusing OK?
 
Hi 454,

It is important to carefully design your grounding so that you do not have any loops or any ground current flowing where it doesn't need to flow. The best way to do this is to have a star ground somewhere in the device. The rest of the grounding can be busses or stars-of-stars, but it should all come back to this single system star ground at some point.

Having a separate power chassis does not change things. Design your grounding and make a map of the grounding and power on a piece of paper. Then draw a line around what you want in the power supply chassis (or vice verse, what you want in the amplifier chassis). Anything that crosses the line is in the cable.

Your system star ground should have going to it:
- All signal references
- All power commons
- All shields of single-ended interfaces
- Safety ground and chassis (Think of these as a single entity)

If you have a separate power supply chassis, there MUST be a separate wire from the safety ground lug on the power supply chassis to a lug on the amplifier chassis.

If you want a Safety Loop Breaker Circuit, disconnect the safety ground and chassis from the star ground and insert the SLB circuit between the star ground and the safety ground and chassis.

Your system star ground could go in either chassis but it makes sense to put it in the amplifier chassis to cut down on the number of wires in the cable. Your SLB could go in either chassis. If you put it in the amplifier chassis then both things you want it to go to (safety ground and star ground) are in that chassis. If you put it in the power supply chassis then you would need to add a wire to the cable to go from the star ground in the amplifier chassis to the SLB in the power supply chassis.

In a very simple system, assuming that the power common wire goes to the star ground, you might be able to cheat a little and use the power common wire in the cable as also the wire between the star and the SLB (with the SLB in the power supply chassis). However, using a wire for double-duty in grounding is a sloppy design practice and at some point will bite you. Also, there is no need to do this because you can put the SLB in the amplifier chassis and save the extra wire.

Dave
 
Hi 454,

Is the fusing OK?

I don't know. A fuse is there to protect something - ask yourself, what is it that you want to protect?

The main fuse in the IEC power receptacle will blow if something gets shorted to the chassis. It is sized to support the current draw of the whole device.

Additional fusing is needed for very sensitive things to be protected. For example, if your whole unit draws 5 Amps and you have a very sensitive device inside that draws 10mA that you want to protect, you would want to separately fuse the 10mA circuit because without that additional fuse the sensitive device would fry before the 5A fuse blew.

Remember, the chassis is connected to safety ground so you have protected yourself from a shock.

Dave
 
David Davenport said:
Hi 454,

Is the fusing OK?

I don't know. A fuse is there to protect something - ask yourself, what is it that you want to protect?

The main fuse in the IEC power receptacle will blow if something gets shorted to the chassis. It is sized to support the current draw of the whole device.

Additional fusing is needed for very sensitive things to be protected. For example, if your whole unit draws 5 Amps and you have a very sensitive device inside that draws 10mA that you want to protect, you would want to separately fuse the 10mA circuit because without that additional fuse the sensitive device would fry before the 5A fuse blew.

Remember, the chassis is connected to safety ground so you have protected yourself from a shock.

Dave
So I DON'T need to fuse the transformer secondary, even as a good practice?

David Davenport said:
Hi 454,

Your system star ground could go in either chassis but it makes sense to put it in the amplifier chassis to cut down on the number of wires in the cable. Your SLB could go in either chassis. If you put it in the amplifier chassis then both things you want it to go to (safety ground and star ground) are in that chassis. If you put it in the power supply chassis then you would need to add a wire to the cable to go from the star ground in the amplifier chassis to the SLB in the power supply chassis.
If I put the breaking circuit in the PS chassis, I need 4 wires (+, 0, -, gr/PE). I need 3 minimum, so if I reduce the number of wires, that means I'm going to be shorting 0 and PE together inside the PS chassis?
 
Hi 454,

You will need a minimum of four wires in the cable: safety ground, + voltage, - voltage and power common (0 volts) no matter if you have an SLB or not, of if you have an SLB, where it goes.

Good grounding practice would have five wires if the SLB were in the power supply chassis, and four wires if the SLB were in the amplifier chassis.

If you want to use an SLB, you can not double duty the power common and safety ground because these are what you want to isolate with the SLB. Also, as I said before, it is not a good idea to double duty any ground wiring.

If the secondary of the power transformer is shorted, that short will be reflected onto the primary and the main fuse will blow. There is no need for an additional fuse to protect the transformer.

Dave
 
454Casull said:

So I DON'T need to fuse the transformer secondary, even as a good practice?
David Davenport said:
If the secondary of the power transformer is shorted, that short will be reflected onto the primary and the main fuse will blow. There is no need for an additional fuse to protect the transformer.
That is, if you answer the question
David Davenport said:
what is it that you want to protect?
with: You want to protect the transformer from shorts, but not from overload.

The trouble with transformers is, they cannot be protected against overload on the primary side, because their high start-up current forces you to use a fuse well above their nominal rating.

Here is, where your own assessment comes into play. Analyze the circuit you feed from the transformer. Is it possible to overload the transformer? E. g. if an output transistor shorts, there will be rail voltage across the speaker -> I = U/R -> I = rail voltage/Rspeaker. Is that current within the transformer rating?
If yes, skip the secondary fuse, but consider the use of a speaker protection, which could be a different fuse at a different place.
If no, will the current be big enough to trip the primary fuse?
If it is, skip the secondary fuse, and again consider the use of a speaker protection.
If it is not, decide, whether it is worth to protect the transformer or if you rather run the risk of damage in exchange for saving on the fuses.

You may come to the conclusion that you don't need the overload protection, because everything before the output transistors can only short and will always trip the primary fuse, and the output transistors get their own protection. Or you may come to the conclusion that Murphy's Law applies and you will get overload through some fault that you did not foresee and use an overload protection just to be on the safe side. Or you can come to the conclusion that you don't care about the transformer, and if it overheats and burns your house down, so what? I. e. if you never liked the house anyway.
 
David Davenport said:
A fuse is there to protect something - ask yourself, what is it that you want to protect?.................................

........................Additional fusing is needed for very sensitive things to be protected. For example, if your whole unit draws 5 Amps and you have a very sensitive device inside that draws 10mA that you want to protect, you would want to separately fuse the 10mA circuit because without that additional fuse the sensitive device would fry before the 5A fuse blew.
Balderdash!
A fuse is there to stop the house burning down.
It's purpose is not to protect something downstream.

A fuse disconnects a circuit that has already been abused and/or damaged.

A fuse must be rated less than the minimum current rating of any cable downstream that does not have separate fusing.
 
pacificblue said:



The trouble with transformers is, they cannot be protected against overload on the primary side, because their high start-up current forces you to use a fuse well above their nominal rating.




Hi pacificblue,

My preference would be to size the fuse for overload protection and use a soft-start circuit to lessen the inrush current.

Dave
 
Hi Andrew,

AndrewT said:

A fuse is there to stop the house burning down.


There are a lot of commercial appliances that do not have a fuse and will not burn the house down. If a fuse was needed for safety in an audio amplifier, the standards organizations would produce a standard specifying such fusing.


AndrewT said:

It's purpose is not to protect something downstream.


Pacificblue made an excellent case that you might want to protect a speaker. I know from unfortunate experience that a speaker coil will open before there is a fire. However, this has happened to me only once in my fifty plus years in audio, so I don't deem it necessary to have a speaker fuse in my amplifiers.


AndrewT said:

A fuse disconnects a circuit that has already been abused and/or damaged.



Yes, it protects the circuit from further damage.


AndrewT said:

A fuse must be rated less than the minimum current rating of any cable downstream that does not have separate fusing.


So, you are protecting the cable?

The cable should be designed such that it will withstand an overload long enough for the main component fuse to blow.

Dave
 
Hi 454,

454Casull said:
What is the 5th wire for?

The SLB should go between the system star ground (with all of the audio grounds) and the safety ground.

If the system star ground is in the amplifier and the SLB is in the power supply chassis, you would need a wire to connect them. This is the fifth wire.

If both the system star ground and the SLB were in the amplifier then the fifth wire in the cable would not be needed.

Dave