Ground Loop Problem: Solving DC Offset in Video

  • Thread starter Thread starter Stefan Merfu
  • Start date Start date
  • Tags Tags
    Ground Loop
Click For Summary

Discussion Overview

The discussion revolves around addressing a ground loop problem that causes a DC offset in the video signal transmitted from a robot's camera. Participants explore various methods to mitigate the issue, including the use of isolation transformers, differential signaling, and capacitors, while considering the implications of cable resistance and electrical noise.

Discussion Character

  • Debate/contested
  • Technical explanation
  • Exploratory

Main Points Raised

  • One participant identifies the issue as a ground loop causing a voltage on the coax cable shield, leading to video signal brightness changes when motors are activated.
  • Another participant questions whether the coax is being used intentionally as a power cord or if this is an unintended effect.
  • A suggestion is made to use an isolation transformer or a differential transmitter to address the ground loop, with concerns about the effectiveness of these solutions.
  • One participant proposes adding a capacitor across the motor leads to reduce noise, while another mentions that decoupling ground with a capacitor affects the DC bias and synchronization of the video signal.
  • A recommendation is made to use differential signals to isolate the shield ground from the power loop, along with a DC-restoration circuit to recover sync levels.
  • Participants discuss the necessity of using shielded twisted pair cables for long runs in noisy environments, noting the importance of correct termination to minimize signal reflections.
  • One participant expresses a desire to understand how galvanic isolation works in the absence of AC signals, prompting an explanation of the principles behind ground loops and isolation techniques.
  • A question is raised about the effectiveness of adding a series resistor to the coax shield to isolate current, with a participant reflecting on the behavior of resistors in parallel.

Areas of Agreement / Disagreement

Participants express various viewpoints on the best methods to resolve the ground loop issue, with no clear consensus on a single solution. Multiple competing approaches are discussed, indicating ongoing uncertainty and exploration of the topic.

Contextual Notes

Participants acknowledge limitations related to the specific configurations of their systems, the resistance of the cables, and the potential impact of electrical noise on the video signal. The discussion remains open-ended regarding the effectiveness of proposed solutions.

Stefan Merfu
Messages
6
Reaction score
0
Hello, i have a really annoiyng problem , i have a robot which is equipped with a video camera(PAL) so nothing too fancy, the information is transmitted through a coaxial cable about 200 m of cables,the attenuation of high frequency is not to big to cause any problems,also the robot has some motors ,this motors take about 3 amps and the cable has a total resistance of about (2 ohms/200m),the problem is that when i turn on the motor the video starts to become brighter until reachs a saturation level.I know that the problem is ground loop which produces a voltage on shield of coax cable , but i don't know how to remove this dc offset in ground of video.I tried a common-mode choke but without success.
 
Last edited:
Engineering news on Phys.org
Are you intentionally using the coax as a power cord? Or is this some sort of unintentional effect?
 
The robot is powered up by another two conductors which have their own resistance, the idea is that the ground coax is going in the same general ground so i can't break the ground loop because will still exist some resistance in conductor that will compromise my composite signal.I'm thinking about a isoltation transformer or a differential transmitter.I don't know how good is that type of transformer and where i can find one at a international distributor like Farnel or TME.
 
Best way/place to eliminate noise is at the source. Have you tried, say a 0.1μf to 1.0μf capacitor across the motor leads? Noise could be causing an offset.
 
I will try, i forget about EMI caused by motor.i tried to decouple ground using a 10nF it's ok but the dc bias of composite signal is gone so i have some issues with syncronization of vertical an horizontal.
 
The best solution is to use differential signals so the shield ground on the (new) twisted pair video transmission cable can be galvanically isolated from the power loop. You can also try a DC-restoration circuit to recover the proper sync levels. Depending on the signal level sometimes only a simple signal diode (in a clamp circuit) after the isolation capacitors on the input and ground will restore sync-levels to something usable.
https://www.maximintegrated.com/en/app-notes/index.mvp/id/3303
 
  • Like
Likes   Reactions: Stefan Merfu
Thank you for proper answear.
Do i need a shielded twisted pair ?
 
Stefan Merfu said:
Thank you for proper answear.
Do i need a shielded twisted pair ?

Yes, for a 200M run with lots of electrical noise. It's common to use shielded/screened Cat5E cable for CCTV installations for long runs. The impedance is usually close to 100 ohms instead of the standard 75 ohm for coax so use the correct termination resistor on the differential receiver to reduce signal reflections.
http://www.farnell.com/datasheets/313294.pdf
 
I understood, thank you for help.I am not too good about transformers and stuff like this, i want to know how this galvanic isolation works if i don't have any type of AC signal on cable like DC voltage which supplies the power for device.If you would like to explaine me will be helpful.
 
  • #10
It's normal to see this problem when you have the power and signal grounds for a remote device with common connections on both ends of a long cable with a small wire resistance that develops a small voltage across the ends as current flows in every parallel ground wire back to the voltage source. It might be possible to totally isolate (if it's all low DC voltage) the camera power/signal ground and supply in the robot if they are don't share a common wires in the robot tether and only connect the camera ground to the common power ground point at the controller. Galvanic isolation works by removing the direct electron path in a circuit. The signal energy is then isolated from the total end to end circuit loop current by space into two loops and is coupled between those loops by the magnetic field of the isolation transformer or the electric field of the capacitor for any AC signal.
 
  • Like
Likes   Reactions: Stefan Merfu
  • #11
Would you like to explain me how ground loop works ? From my knowlegde i know that the ground loop is produced when are more paths to ground with different resistances and this two(three) resistances behave like resistor in parallel.But some people advice me to add a series resistor on shield of coax with a value between 10 to 100 ohms to "isolate" the current.But i know that the resistors in parallel have the same voltage across them.Thank you in advance !
 

Similar threads

  • · Replies 20 ·
Replies
20
Views
6K
Replies
2
Views
3K
Replies
6
Views
3K