3
\$\begingroup\$

I am working on a grounding concept for a vehicle (car, aircraft, you name it) where the vehicle's metallic structure is used as DC current return path for the various loads.

Problematic Ground Configuration

More specifically, I am having trouble with the situation depicted in the above figure . The load device on the right side has its DC power return and chassis ground connected together. The intended return path for the DC current (shown in red) runs via a cable (> 5 mOhms) to a well-defined bonding point. This bonding point is connected to the ground reference point (negative terminal of the battery) via the vehicle's structure (< 1mOhm).

In addition, there is a second undesirable DC current return path (shown in green) via the device's mounting points at the vehicle's structure, which has a resistance of less than 1 mOhm. Obviously, this second DC current return path carries the major part of the device's return current.

(EDIT: The undesirable DC current return path is not guaranteed to exhibit a low resistance, thus it is undesirable in terms of reliability; in addition, the affected structural components are not designed to permanently carry currents.)

What is a technically sound approach to deal with this issue? Both the input configuration of the device as well as the requirement to use the well-defined bonding point are beyond my control and thus cannot be changed.

One solution I've come up with is to use an isolating DC/DC converter as is indicated in the image below. From my understanding the current loops need to be closed as is illustrated in the figure and thus, the entire DC current has to travel via the intended return path.

Suppression of undesired return paths using an isolation DC/DC converter

My questions are:

  1. Is this a valid approach for suppressing the undesired DC return current indicated in the first figure in green?

  2. What is the name of this concept? I've done some googling on this subject but seem to lack the appropriate keywords.

(EDIT: I am aware of power isolation and ground loops. From my viewpoint, this application of the DC/DC converter is all about directing the DC current to the intended return path...)

  1. Can anybody point me to some literature where such issues are discussed? I've done some reading in Lock / Joffe's Grounds for Grounding, however, I cannot find the topology corresponding to my problem in their text and a discussion in terms in currents as carried out in the first figure seems not to be available.

I am aware of the fact that cable shields etc. may provide additional current paths; I've omitted additional interconnect for the sake of brevity.

I am also aware of the additional problems caused by the use of DC/DC converters in terms of EMC.

\$\endgroup\$
9
  • \$\begingroup\$ Google ground loops, because this is what you are breaking with the isolated PS. It's a very common solution in the industry, with just two issues (in my opinion): cost- they are relatively expensive, not very good for high volume production, and relative complexity. Well, they actually simplify the system, but sometimes it's really hard to explain it to people, and to take care not to reconnect some grounds. Upd. Literature will be easiest to find on vendors' websites. Search for Traco, Recom, Murata, and a bunch of Taiwanian companies. \$\endgroup\$ – Gregory Kornblum Oct 1 '20 at 22:53
  • \$\begingroup\$ @Gregory thanks for the reply. I am aware that in terms of EMC this is about breaking ground loops to suppress undesired pickup of interference and to avoid emissions. However, my viewpoint is more about "directing the current in the proper direction". Maybe this viewpoint is somewhat odd, as I could not find it in literature... \$\endgroup\$ – EMC Oct 1 '20 at 23:14
  • \$\begingroup\$ It's the same issue. Your nickname is EMC, mine would be HIGHCURRENT :) Trust me, control of where the currents flow is just as important, and essentially it's the same thing: we always want the current to close in a specific trajectory, not for example under sensitive circuits, etc. For example (from my tough life) motor current is not allowed to flow through thin encoder power wires, so it sometimes makes sense to isolate the encoder. \$\endgroup\$ – Gregory Kornblum Oct 1 '20 at 23:28
  • \$\begingroup\$ So let's say high frequency generator shorts to plane hull. What happens? \$\endgroup\$ – StainlessSteelRat Oct 2 '20 at 1:33
  • 1
    \$\begingroup\$ it is unclear how the second diagram prevents the "green" path ... it is also unclear why the "green" path is actually a problem \$\endgroup\$ – jsotola Oct 2 '20 at 3:09
1
\$\begingroup\$

You have stated that the ground loop is a problem, but you haven't stated why it is a problem. Typically a ground path like this will cause two problems:

  1. The ground will bounce from common mode noise.
  2. The loop of the two grounds will pick up magnetic noise.

The biggest issue is currents should not return on the chassis because it creates saftey hazards during faults and hot plugging (if power connects first the current can go through the chassis, at 12 volts this is not dangerous to a person but it could be to your electronics).

What is the name of this concept? I've done some googling on this subject but seem to lack the appropriate keywords.

Power isolation

Can anybody point me to some literature where such issues are discussed? I've done some reading in Lock / Joffe's Grounds for Grounding, however, I cannot find the topology corresponding to my problem in their text and a discussion in terms in currents as carried out in the first figure seems not to be available.

You probably won't find that specific diagram, but if you want to read up on how to eliminate noise sources and ground loops or other challenging noise situations Electromagnetic Compatibility Engineering is for you.

enter image description here Source: http://www.hottconsultants.com/EMCE_book_files/emce_book.html

Is this a valid approach for suppressing the undesired return current indicated in the first figure in green?

Isolation is a good way to stop common mode noise from return currents, a DC to DC isolator wont really care about the a changing ground voltage, and it will regulate the voltage to a device accordingly as long as the input voltage to the isolate is within it's spec.

The ground then carries very little current (currents that might be capacitivly coupled), and no changing current from the load in the device, because the current from the load returns to the DC DC converter.

There are many of off the shelf isolators available from distributors or power supply manufacturers that have been tested and meet FCC requirements (although these do not apply to many vehicles)

\$\endgroup\$
4
  • \$\begingroup\$ Thanks for the reply! I'm well aware of the problems related to ground loops and power isolation as an approach for mitigating this. My question focusses on ensuring that the DC(!) return current actually takes the intended return path. As discussed above, I assume that this may be achieved using a DC/DC converter, however, I cannot find confirmation for this neither in Mr. Ott's book (which you cited) nor in Grounds for Grounding. Those texts seem to focus on the EMC aspects which you nicely elaborated on:-) \$\endgroup\$ – EMC Oct 1 '20 at 23:25
  • \$\begingroup\$ P.S.: See edit of original post regarding the reasons why the second current return path is undesirable... \$\endgroup\$ – EMC Oct 1 '20 at 23:33
  • \$\begingroup\$ If you use the isolator to separate chassis and power grounds then it will separate the grounds and the current can only flow back on the power ground. However, if the chassis ground is still tied to the power ground then your setup will not seperate current grounds and the current from the isolator will still flow back on chassis and power grounds. Make sure that if and isolator is used, the chassis and power grounds are also separated. Keep in mind this also references your load to chassis ground, which won't be a problem for most designs . \$\endgroup\$ – Voltage Spike Oct 1 '20 at 23:55
  • \$\begingroup\$ Could you please elaborate on how a current flows via chassis ground when using a converter and keeping chassis and power ground tied together? From my understanding, the secondary side of the converter's transformer must close the current loop, although this might be an oversimplification? \$\endgroup\$ – EMC Oct 2 '20 at 0:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.