Simple power supply grounding questions

[hodad]

I've got a linear Power One +/-24V supply I'm using for a project.  I noticed that the "Common" connection on the PSU's circuit board is not connected to the aluminum frame the supply is mounted on.  So does the Common pin at some point need to join up with mains earth?  It's been a while since I've used one of these supplies, so I'm a little hazy on best practice in this situation.

 

[ruffrecords]

Go to the DIY tab of my web site:

https://www.customtubeconsoles.com/diy

Scroll down and click on the Power folder. In there you will find grounding101v2.pdf

Enjoy.

Cheers

Ian

 

[Bo Deadly]

I've got a linear Power One +/-24V supply I'm using for a project.  I noticed that the "Common" connection on the PSU's circuit board is not connected to the aluminum frame the supply is mounted on.  So does the Common pin at some point need to join up with mains earth?  It's been a while since I've used one of these supplies, so I'm a little hazy on best practice in this situation.

Typically all grounds converge at the filter cap ground including the chassis connection which is what the frame should be connected to. If it's all in one frame with regulators and filter caps and such, then the Common terminal is where things should converge. So bolt the supply frame to the chassis of the device such that it's electrically bound. Then, next to mains in, usually there's a tamper resistant bolt that mains earth wire connects to. Then run a second wire from that to the Common of the PS. Then three wires for +24, -24 and Common to your downstream circuits. So the PS frame is connected to Common through the chassis of the device > chassis earth bolt > Common of PS.

 

[john12ax7]

The earth safety mains connection should be its own dedicated connection.  This physical connection should not be shared with anything else,  including PS common.

 

[Bo Deadly]

The earth safety mains connection should be its own dedicated connection.  This physical connection should not be shared with anything else,  including PS common.

I'm not sure I understand. How would the PS common be connected to the chassis then?

 

[john12ax7]

I'm not sure I understand. How would the PS common be connected to the chassis then?

If PS common needs to go to chassis use a separate bolt / physical connection.

 

[hodad]

Maybe not as simple as I thought.....plus it's an external power supply, which complicates things a little.  I think some reading and a little experimenting may be in order.

 

[Bo Deadly]

If PS common needs to go to chassis use a separate bolt / physical connection.

And what is the rationale for that? Is it something that we really need to follow for DIY projects? I don't see how it could make a difference to the performance of the device.

 

[Bo Deadly]

Maybe not as simple as I thought.....plus it's an external power supply, which complicates things a little.  I think some reading and a little experimenting may be in order.

For an external supply, I would recommend that you use a 4 conductor cable. Use 3 for +24, -24 and 0V. Same explanation as before but "chassis" is now the external power supply chassis (assuming it's metal). Use the fourth wire to connect the external PS chassis from aforementioned "bolt" to the remote device chassis directly at the entrance using another bolt. The fourth wire can be a shield as long as it's substantial and not just some foil. I would not recommend that you connect the remote chassis to the local common (meaning the bolt on the remote device just has the one chassis connection to the external supply chassis) but there are folks here that have disagreed with me on that specific point.

If you want to follow john12ax7's advice, just add a second bolt to the external supply chassis and use that to connect the chassis to the supply Common.

 

[hodad]

For an external supply, I would recommend that you use a 4 conductor cable. Use 3 for +24, -24 and 0V. Same explanation as before but "chassis" is now the external power supply chassis (assuming it's metal). Use the fourth wire to connect the external PS chassis from aforementioned "bolt" to the remote device chassis directly at the entrance using another bolt. The fourth wire can be a shield as long as it's substantial and not just some foil. I would not recommend that you connect the remote chassis to the local common (meaning the bolt on the remote device just has the one chassis connection to the external supply chassis) but there are folks here that have disagreed with me on that specific point.

If you want to follow john12ax7's advice, just add a second bolt to the external supply chassis and use that to connect the chassis to the supply Common.

Thanks--this is one of the directions I was considering, and the cable I already have will do the trick but will require some rewiring.  I'll give it a shot.

 

[john12ax7]

And what is the rationale for that? Is it something that we really need to follow for DIY projects? I don't see how it could make a difference to the performance of the device.

The mains earth is a safety issue (and also required to meet code). Anything else connected to it increases the chance of it getting loose.

In contrast where and how you connect PS common is a performance issue. So you have two different requirements.

 

[abbey road d enfer]

The mains earth is a safety issue (and also required to meet code). Anything else connected to it increases the chance of it getting loose.

This is not how it's analyzed here on this side. The rationale is that the most likely cause of problems is leakage from the PSU, so connecting zero volt as close as possible to the earth connection is recommended.

 

[abbey road d enfer]

Typically all grounds converge at the filter cap ground including the chassis connection which is what the frame should be connected to.

Actually, for best audio performance, it's the Ov reference point of the PSU that should be earthed. Typically the regulators reference point, if any. The filter cap ground is often very dirty and is better left floating.
Indeed, for most power amps, which don't have regulators, it's the filter caps ground, but it must be kept free from the diodes' peaks.

 

[Khron]

For what it's worth:

https://sound-au.com/psu-wiring.htm

 

[Bo Deadly]

Actually, for best audio performance, it's the Ov reference point of the PSU that should be earthed. Typically the regulators reference point, if any. The filter cap ground is often very dirty and is better left floating.
Indeed, for most power amps, which don't have regulators, it's the filter caps ground, but it must be kept free from the diodes' peaks.

I don't mean the filter caps of rectified AC. I mean the filter caps on the output of the PS after regulators.

 

[ruffrecords]

The mains earth is a safety issue (and also required to meet code). Anything else connected to it increases the chance of it getting loose.

This was not the view of UL last time I submitted  a product to them. The mains safety earth had to be the first thing attached to the bolt and and fastened with a shake proof nut. Other connections to the chassis could then be add to the same bolt.

Cheers

Ian

 

[abbey road d enfer]

I don't mean the filter caps of rectified AC. I mean the filter caps on the output of the PS after regulators.

Right. That point needed to be stressed. Many don't realize that the secondary CT and the Ov reference are not interchangeable, though being represented by the same line on a schemo.

 

[john12ax7]

This was not the view of UL last time I submitted  a product to them. The mains safety earth had to be the first thing attached to the bolt and and fastened with a shake proof nut. Other connections to the chassis could then be add to the same bolt.

Cheers

Ian

Interesting.  I haven't seen it done that way in UL approved equipment so was unaware that was acceptable. Thanks for clarifying.

 

[Matador]

For those who need to see it in action:  here are 4 simulations showing different 'reference' points.  The largest current spikes happen between the filter cap and the diodes (between nodes G1 through G3), so if you reference your load with respect to node G1, all of the current spikes will superimpose on the measurement of your load voltage (current develops voltage across R5 and R2).  This is the green trace on the simulation.  The 'best' connection is where the SPICE 0V reference is placed (pretend C2 is the regulator reference, and R8 /R9 represent the wiring resistance from the PSU to the circuit being powered).  So that should be your reference point for the 'downstream' circuits.

 

[Newmarket]

This was not the view of UL last time I submitted  a product to them. The mains safety earth had to be the first thing attached to the bolt and and fastened with a shake proof nut. Other connections to the chassis could then be add to the same bolt.

That sounds right to me and aligns with relevant BS/EN LVD standards I believe.