automotive grounding

[buzzp]

I would like to bring up a discussion of automotive electrical systems. I am an EE and have recently started a career in the automotive industry for a newer company so I don't have any ideas to bounce off a more experienced automotive electrical guru. My question/concern is over using the chassis as the ground for electronics. This is to say, the main battery ground ties directly to the chassis. All grounds simply look for a good clean metal surface and the ground is tied there. This would not likely cause any problems. However, the industrial type vehicles we mfg have more than one power source with some sources having multiple batteries. I was drawing up a complete schematic and discovered that one source uses a chassis type grounding system and the other uses wires fed back to a common bus that connects to the other battery source ground. I would like to know thoughts/concerns/comments about any problems this may cause from an EMC standpoint. We have some systems that use microcontrollers but do not appear to have any problems. If anyone could comment on this or point to a source of information I would appreciate it. Thank you in advance.

 

[MMAshwin]

Hi,

I wonder why they ever called it chassis ground. (But thats another question). Surely its the return to the battery using the chassis as the return conductor.

As the car trundles along on its rubber wheels 12" or so
off the ground, who knows what the chassis voltage is.

As far as the electrics/electronics in the car it doesn't matter because the battery is strapped to the chassis so the
+ve connector is always going to be +12 (or so) with respect
to the chassis.

Now maybe, with two or more batteries they are do not share a common 'grounding' point. This may give rise to a difference in potential if the two devices used the same chassis as a return to different batteries.

Or then again maybe the devices attached to the second battery are more sensitive to noise, because (and I don't know if there are any studdies on this) how much does your car chassis ground change in voltage as you drive under power distribution lines.

Or then again it could be the old signal ground / chassis ground debate.

Anyhow just my 2¢. Any help ?, yes no let me know.

Regards

 

[melone]

One interesting phenom is ground potential differences between 2 points on a vehicle. There can be up to a couple of volts difference from the front of the vehicle (close to the battery negative terminal) versus the back of the vehicle (which has some distance of metal to travel). Using a bus bar scheme will definitly INCREASE any potential differences (less metal to get back to battery negative). This can be very detrimental if you are performing any type in input voltage reading ("ground" is approaching battery +) at the back of the car.

However, the standard chassis ground isn't all that fantastic either if done incorrectly. I have seen schemes that have several taps to the chassis at different points in the vehicle. This creates sub-ground circuits. If ground lug that your circuit is connected to draws signifigant current, then the voltage potential of your lug ("ground&quot increases. However, if you have several loads that will operate exclusively (i.e. only 1 out of 8 can be on at any one time), then this can contain your "ground" offset to these high-current loads.

To sum it up, one method isn't neccessarily better than the other. There are advantages to both methods, but a standard schematic will not allow you to chose which one is best suited to your application. Knowing where everything is physically located in the vehicle, coupled with constant attention to return current (most people forget that current flows through ground), will allow you to optimally design your grounding scheme.

Hope I haven't insulted anyone's intellegence!

Good luck and keep us posted!

 

[buzzp]

Thanks for the posts. I realize the potential difference could be an issue with multiple grounding system. I believe, in our case, that ground loops are really the problem. Given we have electrical boxes (metal) screwed to the chassis with conductive conduit running to the electrical box causes a ground loop. It just so happens we have some electronics in the 'box' that are making measurements. When a large inductive load is made or broke (likely breaking) this generates a spike that radiates through the chassis ground to the battery and also through the conduit to the battery, via the electronics. This 'ground bounce' is what I believe is causing the problems with the measurements. As all of the measurements use ground as the reference point. The electricians tell me if they insulate the conduit then the measurement straightens out. This leads me to believe that it is a combination of ground loops, improper grounding, and an inductive load. My solution is to not use a chassis ground at all, at least not for inductive loads. The ground of such devices should be ran directly to the battery ground. I may put in some arc suppression across these coils as well if proper grounding does not solve the problem. If anyone has any further comments or questions please fell free to express them. Again, thanks.

 

I have worked in the medical field for 24 years and have seen ground loop problems cause havoc for even the most experienced engineers. I worked with an old NASA engineer early in my career and he always preached using shielded cable or conduit to isolate circuits from external noise. However, as you have discovered, this is a good way to cause ground loops. He said the proper method of grounding is to ground the transmitting end of the shield or conduit but isolate the receiving end. Otherwise, if you ground the receiving end, the shield or conduit can act as an antenna and actually pick up more noise than leaving it ungrounded. In other words, since you have a device which is taking measurements and sending the data to a processor, that device is the transmitting end and the processor is the receiving end. So far I have never seen this method fail. But I have also seen some manufacturers use an equipotential grounding scheme where a ground wire is run from each and every metallic surface to one central point. They even attempt to make each ground wire the same length. But this seems to be a large waste of copper. I'd stick with the first method. Just a side note...Apparently NASA had problems with equipment in the mile long vehicle assembly building back in the early 197o's. Upon investigation they found a 400 volt difference in potential between the grounding rods that were driven into the ground on each end of the building. They ended up having to run a 4 inch copper bus bar the entire length of the building.