pci6602 pulse train triggering inconsistency

[zhenmeister]

Is it possible that a triggered finite pulse train in niDaqmx generates different results depending on the number of pulses in the train? 

Here is why I am asking.  I have a vb6 app using niDaqmx 8.0 driver that generates a finite pulse train which usually consists of a small number pulses.  When the number of pulses is larger than 1 (ie 3) my pulse train looks like -_-_-_ , in other words High comes first.  When the number of pulses is exactly 1, my pulse train looks like _- , or Low first.  I am aware of the knowledgebase article that describes the initial delay discrepancy for the first trigger (http://digital.ni.com/public.nsf/allkb/8f070c8cfdfec0fa86256f7a00141425), and I am not talking about the first trigger.  Any ideas? 

Thank you

[Sandra_T]

Hello Zhenmeister,

One thing I would look at is the idle state that you have set for your counter task. If you haven't already set it, do so. This specifies the resting state of the output terminal. This ensures when the counter is resting it will be in a set state. Please give this a try.

Regards,
  Sandra T.

Applications Engineer | National Instruments

[zhenmeister]

Sandra, thank you for your reply.  I am pretty sure that my idle state is explicitly set to "low", but I will check it again.  One thing to note is that these results are produced by the same code with the only exception of the number of pulses in the train.

Thank you,

-ZH

[Sandra_T]

Hello zhenmeister,

Let me know if setting the idle state make a difference. Are you also setting the intial delay? How are you measuring the pulses?

I'd be interested in seeing the code and some screenshots of the waveform if you don't mind posting it here on the discussion forum.

Regards,
  Sandra

[Sandra_T]

Hello zhenmeister,

Let me know if setting the idle state make a difference. Are you also setting the intial delay? How are you measuring the pulses?

I'd be interested in seeing the code and some screenshots of the waveform if you don't mind posting it here on the discussion forum.

Regards,
  Sandra

[zhenmeister]

Hello Sandra,
Sorry for the delayed response - I have been swamped.  I don't mind posting my code, but I don't have the waveforms readily available.  I'll post the waveforms as soon as I have them. Below you will find the function that I am using to set up the pulse.  One thing to note is that NumReads is the parameter that I have been referring to, which varies from 1 to 10.  Please let me know if you would like more details.

Thank you,
ZH




Private Function initLampGate(device As String, _
                              delay As Double, _
                              Low As Double, _
                              High As Double) As Boolean
   
  On Error GoTo LampGateErrHandler
   
  DAQmxErrChk DAQmxCreateTask(" ", LampGate)
  DAQmxErrChk DAQmxCreateCOPulseChanTime(LampGate, "/" & device & "/Ctr2", "lampgate", DAQmx_Val_TimeUnits2_Seconds, DAQmx_Val_Level1_Low, delay, Low, High)
  DAQmxErrChk DAQmxCfgImplicitTiming(LampGate, DAQmx_Val_AcquisitionType_FiniteSamps, NumReads)
  DAQmxErrChk DAQmxCfgDigEdgeStartTrig(LampGate, "/" & device & "/PFI30", DAQmx_Val_Edge1_Falling)
  DAQmxErrChk DAQmxSetStartTrigType(LampGate, DAQmx_Val_TriggerType8_DigEdge)
  DAQmxErrChk DAQmxSetStartTrigRetriggerable(LampGate, True)
  DAQmxErrChk DAQmxStartTask(LampGate)
  initLampGate = True
  Exit Function

LampGateErrHandler:
  MsgBox "LampGate: " & Err.description, vbCritical, "Operation was not successful"
  initLampGate = False

End Function

[Kevin_Price]

Can't help on driver call syntax as I only use LabVIEW.  However, I *think* I remember something that might be relevant.  I'm not sure I trust my memory though...

I *think* I remember observing that when I tried to make a Finite Pulse Train with exactly 1 pulse, that the DAQmx driver was smart enough to perform a single pulse generation instead.  Finite Pulse Trains require use of 2 counters working together, 1 for output signals, 1 to regulate the duration that output is allowed.  A single pulse only uses 1 counter.  DAQmx may be optimizing for you when you ask for exactly 1 pulse, and perhaps this is a clue for why the reaction timing looks different?

-Kevin P.

[zhenmeister]

Kevin,
If my findings are indeed correct then your explanation seems very logical.  Plus, I can easily check it by trying to utilize the second counter of the corresponding pair for the  'NumReads=1' case.   Thanks so much!
-ZH

[zhenmeister]

Kevin,
Your suggestion is absolutely correct.  While testing it, I was able to use the second counter of the corresponding counter pair ONLY if the number of pulses in the train was set to 1.
Thank you again,
ZH