I've got a network of 7 CPI Devices (mix of A108/A116's) being driven by a CR6. The CR6 is also controlling other devices and so the cr6 is in sequential mode. The system will operate fine for about 30-60 days and then the CPI bus will show 1 or 2 FrameErr's in the CPIStatus and the entire network goes down. The only way I've found to bring the network back up is to re-compile the program (either with a power cycle or program send). Can anyone tell me if there's a programmatic way to reset the just the CPI bus based on the FrameErr value? Would also be interested to know if there's a bug in the CPI protocol that it won't heal itself after a single framing error?
One of the devices dropped off the network and would not appear back in the network. We swapped this device for a new one and sent it in for repair and the repair department can't duplicate the error. Now a device in a different position is doing the exact same thing.
We are not aware of an inherent failure when Frame Errors occur. We have done significant testing to verify recovery from this type of condition.
In the first post you stated that the entire network goes down and in the second it sounds as if a single device is lost from the bus. Are you seeing two different modes of failure?
As for your question, there is not a way to programtically reset the CPI bus other than forcing a restart of the program with the RunProgram instruction. Obviously, it would be better to find root cause of the failure and address that. If you could contact our support team and pass along more details about your system, we would try to duplicate what you have observed so we can understand and correct the problem.
Thanks for the reply. I am seeing 2 symptoms.
The first is that a single device (happened with multiple devices in different installed locations) will disappear from the network entirely and everything will continue on like that device was never there.
The second is that a single divice will disappear or go inactive. The CPIStatus table will show 1 or 2 framing errors and the entire network will stop measuring. All measurements freeze at the last successful transmission. I will show skipped scans but the number is not high enough to assume it is skipping every scan. This is in a slow sequence and so the main scan continues to operate as intended and the devices in my slow sequence that are not part of the CPI Network will still measure as intended. So it's like only the CPI network is locking up but the logger does not set the previous measurements to NaN with a failed measurement.
I'm currently working with Mark Wallace who is relaying information back and forth between Tim and I. So far, we have not found a cause or solution.