cRIO Synchronization

[edjones93]

As mentioned, they appear to be synchronised and no further changes need to be made. I located this document:

http://www.ni.com/newsletter/50130/en/

However do note that Table 1 provides best case scenario, best case is synchronisation of roughly 100 nanoseconds. How is your network setup?

Why do you need this level of synchronisation? What is in fact the application and system?

[Mmehrabin]

There is a DC motor which rotates in a 20ms period. There is a sensor that detects its rotation. There is a dialed disk on the motor that has degrees. The sensor detects the 0 degrees of each rotation. I will raise signal A, after raising A, the first 0 degrees should be detected by the sensor. Then after 1/3 of the period signal B should be raised by the controller. Signa A is in 6097 and B is in 9035 (the system is distributed). So, the synchronization is the matter. As you can see in the attached file, the time difference between two controllers is 500 ms that is too much. Since this difference is about 0.5s, I am not sure even they are synchronized correctly.

[edjones93]

The other problem I see here is how you are confirming synchronisation, your network communication to look at the front panel has latency and delays and inherently is not a good way of confirming synchronisation.

What you need to confirm synchronisation is a more precise time base. I would wire a digital signal to both controllers on a digital input, have this read on the FPGA and when it goes to true, fire an interrupt. That interrupt on the RT level and FPGA, then record time. That is a good way to confirm synchronisation, what you are currently doing is actually showing the times on the controllers some time ago when the network packet was sent out.

Since network communication is in-deterministic, this is not a good way of confirming synchronisation.