zappedagain
Electrical
- Jul 19, 2005
- 1,074
So this is a continuation of my thread from over 5 years ago - thread248-266906
So we designed a LVDS receiver channel into a controller along with several other communication channels. An opportunity has come up where I can run 5 channels (cables) for 30m from the sensor to the controller in a harsh environment (I expect arc welders operating nearby), or I can combine the 5 channels of data into a single LVDS channel and then separate the data in an Altera FPGA with an integrated LVDS receiver.
Concern has been voiced by colleagues that LVDS doesn't recover well from EMC events, usually due to the resynchronization time. I've seen Youtube videos with people apparently claiming sub-second recovery when they plug a cable back in. I assume it has to be better than the 5 seconds it takes my TV to resync when I switch from HDMI1 to HDMI2. If I can resync in less than 1 millisecond then the recovery time will be insignificant (as compared to the data lost during the event).
A few caveats already:
- Initial research indicates the recovery time is firmware dependent, as you have to first detect the corrupted data and then resynchronize the data.
- Well shielded cables will help this but aren't an absolute fix.
- The current design will expose a conductive path from my FPGA LVDS pins to the connector on the exterior of the case, so I may have other issues.
Does anyone have any experience, good or bad, with using LVDS over long distances?
Thanka,
Z
So we designed a LVDS receiver channel into a controller along with several other communication channels. An opportunity has come up where I can run 5 channels (cables) for 30m from the sensor to the controller in a harsh environment (I expect arc welders operating nearby), or I can combine the 5 channels of data into a single LVDS channel and then separate the data in an Altera FPGA with an integrated LVDS receiver.
Concern has been voiced by colleagues that LVDS doesn't recover well from EMC events, usually due to the resynchronization time. I've seen Youtube videos with people apparently claiming sub-second recovery when they plug a cable back in. I assume it has to be better than the 5 seconds it takes my TV to resync when I switch from HDMI1 to HDMI2. If I can resync in less than 1 millisecond then the recovery time will be insignificant (as compared to the data lost during the event).
A few caveats already:
- Initial research indicates the recovery time is firmware dependent, as you have to first detect the corrupted data and then resynchronize the data.
- Well shielded cables will help this but aren't an absolute fix.
- The current design will expose a conductive path from my FPGA LVDS pins to the connector on the exterior of the case, so I may have other issues.
Does anyone have any experience, good or bad, with using LVDS over long distances?
Thanka,
Z