06-07-2010 12:18 PM
Hi,
I have a PCI 6221 (M series mulitifunction DAQ board). These boards are specified to generate analog output at a max of 833 kS/s for one AO output channel. When I look at transitions from 0V to 1V at the analog output pin (AO0) using an oscilloscope, the transitions take about 1-2 microsecond to settle! Do other users have a similar experience? I find this to be a long time because if I were to produce analog output at the max rate every transition would take ~1.2 microsecond, which is essentially the measured settling time.
Thanks
06-08-2010 02:18 AM
How did you create that pulse? How did you measure it (input/output impedances, cables)? What are you measuring (10%-90%) or settle to 1% or????
(1-2ms sounds like software timing ....)
Can you try to measured the transition time with a generated square wave and hardware timing?
To measure a 1MHz square wave within 1% your scope (and cable/probe) need at least a 30time higher bandwith (OK, 10MHz scopes are historical nowadays 😉 )
06-08-2010 10:44 AM
Hi,
Thanks for the tips. I used NI Measurement and Automation Explorer. under the Test Panels tab, I used the analog output to generate a jump from 0V to 0.7V (or 1V or 10V). I used my own application (in C using NIDAQmx) and had similar results. See the attached .tiff. The output I examined using a Tektronix scope ad probe. I conencted the AO0 output directly to a 1MOhm input channel sampling >=100MHz. It takes >1microsecond to get to about 90% of target voltage.
With regards to hardware vrsus software timed output, I'm quite certin I'm using hardware timed. Here's the lines of code:
DAQmxErrChk (DAQmxCreateTask("",&AOHandle));
DAQmxErrChk (DAQmxCreateAOVoltageChan(AOHandle,"/Dev1/ao0","",StimMinAmp,StimMaxAmp,DAQmx_Val_Volts,NULL));
DAQmxErrChk (DAQmxCfgSampClkTiming(AOHandle,"",StimSampRate,DAQmx_Val_Rising,DAQmx_Val_ContSamps,SampPerChan_StimBufSpace));
Thanks in advance for any further tips and suggestions,
Unconjugated && untransposed,
Shiva
06-08-2010 10:57 AM
Hi Shiva,
The settling time for this device is actually specified to be 6 microseconds, as shown in the device manual. You are seeing well below this value. If this is unacceptable, then you may want to look toward a function or arbitrary waveform generator.
06-08-2010 11:09 AM
Hi,
Thanks for the information. When I first observed the long transition (IMO) I checked the device manual and specifications sheet and saw the 6 microseconds quoted. I guess I'm not certain why the card is rated at outputting at 833 kS/s (~=1.2 microsecond per sample) if it takes 6 microseconds to settle, or ~1.5 microseconds to get to 90% of target voltage. I was checking with users to see if I perhaps missed something.
Thanks,
Shiva