Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
05/05/06 03:34
Read: times


 
#115621 - I don't know much about your application
Responding to: ???'s previous message
Mahmood said:
If we have sensitive equipment and high precision sensors on board, very high current motor drives then making the chassis connected to gnd is looking for trouble, this way my big sheild will be acting as gnd noise pool affecting all other equipment

Mahmood, it's very difficult to make a proper analysis without knowing any details of your set-up. So, everything I can do is to talk in global dimensions.

Assume you have two separate locations at your car which have to be connected electronically, where sensors or actuators or something like that are located, named A and B. Further assume, that there's ground noise between A and B, due to currents flowing along the chassis. Then the rick is the following:

Reference the circuitry at A and B to these local grounds, means connect the ground of circuitry A to chassis ground at A and ground of circuitry B to chassis ground at B. As long as there's no connection between circuitry at A and B, nobody is experiencing the ground noise between A and B.

Now connect both circuitries by a signal routing methode, which provides proper common mode rejection, RS485 for instance. Then, the ground noise, which seems to shift chassis ground at A relative to B or vice versa, depending on point-of-view, will vanish to a huge amount.

Maybe the use of symmetrical signal routing looks uncommon to you? Then let me tell you, that I do nearly always use symmetrical signal routing in all of my analog applications. Every time when two separated PCBs have to be connected, even when they are sitting in the same enclousre, I use symmetrical signal routing! Symmetrical signal routing is the standard in every recording or broadcast studio. There's no better methode to suppresss common mode noise than using symmetrical signal routing.

It's impossible to arrange all your high current stuff in such a way, that no common mode noise or ground noise can develop, anyway. So it's better to expect from the beginning to have common mode noise in your application and to use a signal routing methode which can suppress it. And the solid ground plane performance of your chassis will help you to keep the inductance of all your ground return currents little.

With the scheme above you can even use cables with the shield connected at both ends to chassis ground (A and B). You don't need to fear the additional common mode noise eventually resulting from additional ground loops, because you use symmetrical signal routing. And, you reliably avoid lousy shielding performance and other unsane inductivity effects at the same time.

Another advantage of symmetrical signal routing is, that the cable shield is nearly free of any ground return currents! Means, the cable shield does not transmit any signal current, but does only shunt interference and injected noise to ground. So, this methode intrinsically avoids the develop of additional common mode noise between A and B, just by design!

Common mode rejection of symmetrical signal routing methode can even be enhanced by the use of common mode filtering, where, for instance, both lines are wound on the same ferrite bead, allowing the signal to propagate but suppressing common mode noise.

Kai

List of 50 messages in thread
TopicAuthorDate
Ground noise question            01/01/70 00:00      
   Chassis ground            01/01/70 00:00      
      measure what?            01/01/70 00:00      
   Shielding            01/01/70 00:00      
   I think mutual ground will be better            01/01/70 00:00      
   Ground noise            01/01/70 00:00      
      I'm with you.            01/01/70 00:00      
         the situation is like this            01/01/70 00:00      
            Try this            01/01/70 00:00      
               Disconnecting of battery while welding            01/01/70 00:00      
      Cures...            01/01/70 00:00      
         Shielding            01/01/70 00:00      
            Old tradition            01/01/70 00:00      
               New traditionalists            01/01/70 00:00      
                  What I want to tell...            01/01/70 00:00      
   Chassis connection makes sense            01/01/70 00:00      
      Not convineced!            01/01/70 00:00      
         Consultants ?            01/01/70 00:00      
            mmm            01/01/70 00:00      
         I don't know much about your application            01/01/70 00:00      
            I think this one has been missed            01/01/70 00:00      
               Differential, symmetrical...            01/01/70 00:00      
               Brilliant!            01/01/70 00:00      
               What about current loop?            01/01/70 00:00      
                  Concept            01/01/70 00:00      
                     Injecting signal into gnd            01/01/70 00:00      
                        Insufficient common mode rejection.            01/01/70 00:00      
                           Go differential            01/01/70 00:00      
                        That's why...            01/01/70 00:00      
                           in other words            01/01/70 00:00      
                              Split ground            01/01/70 00:00      
                           So, how do you maintain a "clean" GND?            01/01/70 00:00      
                              There's no golden rule!            01/01/70 00:00      
                                 back to the future            01/01/70 00:00      
                                    make you happy!            01/01/70 00:00      
                                       write a c code in keil 2 operate e timer            01/01/70 00:00      
                                          Err ?            01/01/70 00:00      
                                             so sorry...            01/01/70 00:00      
                                          what is Keil 2?            01/01/70 00:00      
                                             Trying SMS style?            01/01/70 00:00      
                                                naah, just early morning, fingers stiff            01/01/70 00:00      
                                 More thought            01/01/70 00:00      
                                    Crazy            01/01/70 00:00      
                                       Yes thank you            01/01/70 00:00      
                                    Eh??            01/01/70 00:00      
                                    Common mode noise of DC/DC-converters            01/01/70 00:00      
                                       why use isolated converters?            01/01/70 00:00      
                                          Exactly!            01/01/70 00:00      
                                             air craft chassis            01/01/70 00:00      
                                                Faraday cage ?            01/01/70 00:00      

Back to Subject List