6738 and -50103 error

[cfgmgr]

I am using a new PCIe-6738 analog out card to replace two older PCI-6703 cards, but now I'm running into -50103 errors on code that previously worked fine (at least it worked fine for the last 15 years! 🙂 ) 

My sequence of operations is as follows:
1) Call "DaqMxCreateTask()" to create a new task

2) Call "DAQmxCreateAOVoltageChan()" using the handle from step 1  (one channel per task)

3) Repeat steps 1 and 2 above for each channel on the AO device.

The basic idea was one AO channel per task. 

After the above was finished, I would then start each task that I'd created (using DaqMxStartTask()) and when it was time to update each channel, I would call DaqMxWriteAnalogScalarF64() passing the appropriate task handle for each channel. 

This all used to work fine with the 6703 cards, but now with the 6738, I get a -50103 error after I attempt to start the task associated with the 2nd analog channel. (Starting the task for the 1st channel succeeds, but the 2nd channel fails with -50103)

So does the PCIe-6738 not support one task per channel?    Do I have to assign all channels to one task?  Can I no longer use DaqMxWriteAnalogScalarF64() to update the channels?

What is the best workaround here?   Any help would be appreciated!

Thanks!

[rolfk · Accepted solution]

These two devices are not at all the same type of devices. The 6703 is a 16 channel static Analog Output. With it there is no on-board hardware timing circuitry that will clock the analog output converters with a preconfigured sampling rate. The data gets output whenever the software writes a value to a specific channel.

The 6738 is a 32-channel 1Ms/s hardware timed analog output device. But there is only one timing circuitry available for the analog output functionality so once you configure a task with one or more channels you can't configure another task with more channels.

I'm not sure there is any way to use the 6738 in purely software controlled (static) analog mode. Those DAC converters get triggered by the timing circuitry which there is only one of them.

[cfgmgr]

Thanks for the reply. 

Through trial and error, I discovered exactly what you are describing.  To solve the problem, I ended up just creating a single task and updating all the outputs at the same time.  It's kind of a pity that NI did not maintain software compatibility between the two boards, but after 15 years, I suppose that's understandable.

Thanks again for taking the time to reply - much appreciated.