Synchronizing Servo command outputs to of NI Motion Controller to an analog output card (i.e.. 6733)

[wade75]

Hello,

          I need to somehow synchronize the NI Motion Analog outputs to the analog outputs of a PCI Analog output Card, for instance a PCI 6733.  I have a motion control application that applies a +- torque on a component.  This part is working fine.  However for an upgrade I need to now apply a cross axial load on the component that would basically be a squarewave that would be positive and negative depending on the polarity of the the torque command.  Both commands have to cross zero at the same time though.  The intention is to use dual action air cylinders to provide the compression/tension loading.  But I have to somehow trigger these cylinders at exactly the right time.  I don't think software timing and digital signal won't be sufficient or trustworthy.  Any Ideas would be appreciated.

Regards,

             Wade

[Jochen]

Wade,

the internal timing of the NI Motion boards typically runs with 250 µs or faster, so I agree with your statement, that there is no way to synchronize an external AO board to such a quick signal through software (neither host nor onboard programming will work for you).

I could think of two different approaches:

1. Voltage comparator

If you just need to switch the polarity of a voltage depending on the polarity of the DAC signal of the NI-Motion board, you could design an external comparator circuitry that switches signals in hardware.

2. R-Series

For the case that building an external circuitry by your own is not an option, you might consider using an R-Series device like the PCI-7830R with LabVIEW FPGA to read the motion board's output voltage with an AI channel and output a voltage according to your requirements. You can program this board in LabVIEW and the code is compiled to the onboard FPGA of this board, so the decision making is extremely fast and should meet your timing requirements.

I don't know if solution #2 is completely oversized, but especially if you need more flexibility in terms of defining the AO voltage levels or if there is potential for more real-time requirements, this might be a considerable option.

Kind regards,

Jochen Klier

National Instruments

[wade75]

Thanks Jochen for the suggestions.  I agree option 2 is a little overboard for the application. However option 1 is definantly an option that I will have to investigate and it should be fairly inexpensive.  However I think I would use the feeback from the torque sensor in that application so the transition happens exactly when the actual feedback transition happens.