The Problem about Monitoring Motion using PCI-7358 on LabVIEW Real Time Module

[ggenc]

Hello everyone,

I have a problem about monitoring the position of an axis. First let me give some details about the motion controller system I’m using.

I’m using PCI-7358 as controller and MID-7654 as servo driver, and I’m controlling a Maxon DC Brushed motor. I want to check the dynamic performance of the actuator system in a real time environment so I created a LabVIEW function and implemented it on LabVIEW Real Time module.

My function loads a target position (Load Target Position.vi) and starts the motion. (Start.vi) then in a timed loop I read the instantaneous position using Read Position.vi. When I graphed the data taken from the Read Position.vi, I saw that same values are taken for 5 sequential loops. I checked the total time required by Read Position.vi to complete its task and it’s 0.1ms. I arranged the loop that acquires the data as to complete its one run in 1ms. But the data shows that 5 sequential loops read the same data?

[Jochen]

Hi,

this is expected behavior. When you read the position from the board, it is read from a register that is updated every 5 ms on the board.
Could you please explain, what you want to accomplish? At which rate do you need to capture position data? Do you need these data instantly (within each loop iteration) or would it be ok for you to collect bigger chunks of position data and analyse them at a lower rate?

Best regards,

Jochen Klier
National Instruments

[ggenc]

Hi,
I want to obtain a graph that shall give a precise idea about the dynamic performance of the axis when it's driven using LabVIEW Real Time Module and when commanded by a target position. 5ms sampling rate for my axis performance is not enough as I am trying to complete my motion in less than 10ms. I don't need these data in my calculations or in my program instantly (they are not input parameters to rest of the program), so I can store them in an array and draw the graph after the motion is complete if Waveform subvi's are taking long time to finish their task.

If the ReadPosition can run in .1ms then there shall be a way to read that datas in that rate? Is there a way or not? Using on board programming can be a solution?

Thanks for your interest...    

[Jochen]

Read Position.flx can execute much faster than 5 ms but as it reads a register that is updated every 5 ms on the board, it reads the same value multiple times.
To get around this problem there are two methods:

1. Buffered High-Speed-Capturing (HSC)
   With buffered HSC the board stores a position value in it's onboard buffer each time a trigger occurrs on the axis' trigger input. HSC allows a trigger rate of about 2 kHz. That means, you can store a position value every 500 µs. Please refer to the HSC examples. You may have to look into the buffered breakpoint examples to learn how to use a buffer, as there doesn't seem to be a buffered HSC example available. Please note that you need an external trigger-signal (e. g. from a counter of a DAQ board). Please note that the amount of position data that you can acquire in a single shot is limited to about 16.000 values.
   
   
2. Buffered position measurement with additional plugin-board
   If you don't have a device that allows you to generate a repetitive trigger signal as required in method 1.), you will have to use an additional board, e. g. a PCI-6601. This board provides four counter/timers. You could either use this board to generate the trigger signal or you could use it to do the position capturing itself. A PCI-6601 (or an M-Series board) provides can run a buffered position acquisition with a rate of several hundred kHz and with virtually no limitation to the amount of data to be stored. You could even route the encoder signals from your 7350 to the PCI-6601 by using an internal RTSI cable (no external wiring required).

I hope this helps,

Jochen Klier
National Instruments