Menu

Topic-icon Ground problem with AX2850HE

  • 's Avatar Topic Author
14 years 8 months ago #4842442 by
<P style="MARGIN: 0px">Cosma,
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">We experienced two related, interesting problems with a 2850HE recently that we were able to work around, but we were wondering if you had seen anything similar to this.
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">It seemed to be a faulty ground connection in the DB-15 serial port on the controller. Whenever ground on the serial line was connected, we would lose a byte from the controller every once in a while, which would cause a communications problem. Furthermore, one channel of the controller, when operating in closed loop mode with encoders, would not correctly hold its position while operating. Instead of trying to apply just enough resistance to keep the wheel steady, the controller would apply a large amount of force in the opposite direction you attempted to turn it.
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">Both of these problems were eliminated and the controller operated as normal when we eliminated the ground connection from the serial cable, i.e. just connected TX and RX to the controller from the PC. As an aside, we also eliminated our cabling as the culprit, as the setup worked just fine on a near-identical robotic platform (it to using the 2850HE).
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">Any thoughts on what this odd ground problem could be? Just a faulty controller?
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">Thanks,
<P style="MARGIN: 0px">Barry O'Brien

Please Log in or Create an account to join the conversation.

  • cosma
  • cosma's Avatar
14 years 8 months ago #4929889 by cosma
Replied by cosma on topic Ground problem with AX2850HE
<P style="MARGIN: 0px">I see no logical explanation for your problem. Certainly, the ground is a required signal for RS232 and it is suprising that it would work at all without it. This seems to indicate the ground is finding another return path. The best in these cases is to use an oscilloscope to verify the shape of the signals.
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">However, we've had a couple of users reporting problems when using cables longer than 3ft to the encoders. The cable would act like an antenna and induce electrical noise to the controller's power supply. The result was erratic behavior for the controller, in one case causing serious damage to the unit.
<P style="MARGIN: 0px">
<P style="MARGIN: 0px">

Please Log in or Create an account to join the conversation.

Moderators: tonysantoni
Time to create page: 0.077 seconds
Go to top