SMPS and grounding

[JanRSmit]

I freed up an Netgear GS108E from my business network, and decided to measure the effect of grounding as described in this topic.

As you can see, the SMPS is connected to the electronics ground plane (EGP) and the EGP via small cap's connected to the chassis, via each of the four mounting screws.

I measure with my Fluke 8010 multimeter, the first generation FLUKE that measures flat in the audio band of 20Hz - 20KHz.

When i mention V-ac it is therefore the ac voltage measured over the entire frequency band the Fluke covers.

 

For grounding i use a cord of 1m connected at the safety earth of the AC-outlet where the SMPS is connected to only the N and L (the normal 2-pin connector).

The other end is used to connect to the chassis, the EGP or both. When this grounding is used, the  minus of the multimeter is also connected to this end of the grounding cable.

 

The SMPS when connected measures 12V DC and 0.2MV (AC)

 

I measured at 4 points:

p1: Chassis,

p2: EGP,

p3: Shield of connected SFTP patch cable, not connected on the other end.

p4: any of the 8 connections in the RS45 plug of the SFTPpatch cable (1m) on the open end.

 

Case 1: SMPS connected , no grounding whatsoever: p1: 87.5 V-ac ; p2: 89.1 V-ac ; p3: 86.7V-ac; p4: 78.9V-ac

Conclusion: a significant potential above 0 on each of the measured points. Cannot be  good 😉

 

Case 2: Chassis grounded: p1: 0.2mV ; p2: 26.0V ; p3: 0.2mV ; p4: 4.3-4.4V

Conclusion, better than Case 1, but still serious potentials in the electronics.

 

Case 3: The minus pole of the DC plug grounded (thus also the EGP): p1: 0.2mV ; p2: 0.1mV ; p3: 0.2mV ; p4: 0.2mV

Conclusion: a lot better. 😉 Note that to get to a stable low   low voltage of 0.1-0.2 mV took a few seconds, which can be the multimeter.

 

Case 4: As Case 3 plus chassis grounded as well: p1: 0.1mV; p2: 0.1mV ; p3:0.1mV ; p4: 0.1mV

Conclusion: measurement wise the best.

 

Wish i had a oscilloscope with spectrum analyzer, i then could have examined the frequnecy spectrum.

 

But overall conlusion: grounding , like described , and on this switch,  does matter. 

Next step: factor this switch into my IP chain connecting my NUC as one-node ROON server/endpoint.  And find out the audibility of these cases.

 

Note: in current IP chain, the root-switch (Ubiquity) has its chassis grounded, and that made a huge difference already. The SFTP cable from the root-switch to the IP45 outlet is of course earthed on the switch side, the patch cable connecting the NUC is an utp type, to prevent a ground loop.

 

 

 

 

 

 

[JanRSmit]

On 7/29/2019 at 11:22 PM, Speedskater said:

 

But having the DC supply common or audio circuit common connected to the AC outlet is bad. These grounds must be kept inside the chassis(s).

Hi Speed skater, can you elaborate on this? I am not sure i onderstand what you mean with "common"

[JanRSmit]

14 hours ago, Speedskater said:

Because the word 'ground' has so many different meaning, circuit common is specific.

The DC supply 0V can be connected to the inside of the chassis. The chassis is probably connected to the Safety Ground. But you should not run an insulated wire from the DC 0V out to the Safety Ground.

Just because there is continuity from one Ground to another, does not make them the same Ground.

I understandy youruse of common.

Whether the chassis is connected to safety ground is a dangerous assumption.

With electronics with a 2-prong mains plug it is not.

Whether it is connected to the common as you defined it and if so how is another thing you should not assume.

 

Connecting the chassis (case) if it is metal to safety ground (the thirth prong on the mains plug) is a must. This also implies the requirement that the safety ground on the outlet is proper functioning.

In case of a Netgear switch i have it is not. (2 prong plug) Thus i need to create a connection to safety ground.

In other words you have to check.

 

 

 

[JanRSmit]

18 hours ago, Speedskater said:

On attaching Safety Grounds, commons and shields

Neil Muncy(RIP) wrote his AES paper on these almost 25 years ago, but there is still much confusion.

 

a] Safety Ground: is attached to the chassis near where the AC cord enters the chassis.

b] audio circuit common: is attached to the chassis at a single point near the input connectors.

c] DC supply common: is attached to the chassis at the same single point as the audio circuit common.

d] cable shield: is attached to the chassis at it's chassis connector. Experts have different view-points has to whether both ends or one end of the shield should be attached.

Indeed, correct. Especially forclow freq noise,

[JanRSmit]

3 hours ago, Speedskater said:

Not just at low frequencies, at all frequencies.

Jim Brown, EMI/RFI guru writes:

Antenna Action The most fundamental cause of interference to other systems is the fact that the wiring for those systems, both inside and outside the box, are antennas. We may call them "audio cables" or "speaker cables" or "DC supply cables" or "AC cords" or printed circuit traces,but Mother Nature knows that they are antennas! And Mother Nature always wins the argument.

In a paper from Bill Whitlock on grounding of 2012, and another paper i have not yet found back, a clear distinction is made between low and high frequency behaviour. 

Bill does pay a tribune to Neil Muncy. https://centralindianaaes.files.wordpress.com/2012/09/indy-aes-2012-seminar-w-notes-v1-0.pdf

 

[JanRSmit]

3 hours ago, marce said:

Double insulated don't connect to PE.

Double insulated with a metal case? Please explain hoe.