How to recognize s-series daq card in DAQmx Base 2.0

[planet]

I installed the "DAQmx BASE 2.0", and then installed 3 PCI DAQ boards.  It can recognize PCI-6024E and PCI-6713, but not PCI-6133.

I followed exactly the same procedure as in the

document called "How do I use NI-DAQmx Base with PCI/PXI Devices on Windows?"
http://digital.ni.com/public.nsf/websearch/9C4E9FD0C7469BC286256F100074E ADC?OpenDocument

However, there is a web page (the link below) having labview example programs using DAQmx Base for "NI 6110 S Series Multifunction DAQ". So I think probably there is a way to recognize the S-series devices in DAQmx base.

Does anybody know how to recognize S-series device in DAQmx Base 2.0?

Thanks

[AlanArmstead]

NI-DAQmx Base does not support the 6133.  It does support other S Series devices.

Why are you trying to use NI-DAQmx Base in Windows anyway?  I would strongly suggest using NI-DAQmx if you are going to be on Windows.

-Alan A.

[planet]

Alan,

Thank you very much for the quick response.

The reason I explored DAQmx: I want to do RLP and DAQmx Base has a simpler interface and more examples for RLP.

The reason I want to go RLP instead of DAQmx is a long story.

Summary of my applications:
I need to read 4 input voltage (using PCI-6133) and then generate 12 voltages (using two PCI-6713 each has 8 outputs). Since the 12 generated voltages are related to the 4 input voltages, single sample /N channel AD/DA is required. The speed for the overall control loop has to be >100 kHz.

Summary of my configurations:
PIII, 700 MHz CPU, 512 MB RAM
PCI-6133 and PCI 6713, LabVIEW 7.0, and DAQmx 8.0
I also have PCI-6110 and PCI-6024E cards for comparison

Summary for 4 channel single point acquisition (See the table below, AD only)
1) The difference between P3/700MHz computer and P4/3GHz computer is very small
2) The S-series performed worse than E- and M- Series
     Spec (kS/s)   Tested by me(700MHz P3)   by NI engineer (3GHz P4)
6024E   200/4            15.7 kHz                    17.36 kHz
6040E   250/4                                        18.9 kHz
6251   1000/4                                        23.0 kHz
6133   2500/1            13.8 kHz

Summary for single channel single point acquisition (See the table below, AD only)
1) The S-series performed much worse than E- and M- Series (simultaneous sampling is worse?)
     Spec (kS/s)    Tested by me (700MHz P3)
6024E    200/1           54.8 kHz
6110    5000/1           19.4 kHz
6133    2500/1           14.5 kHz
Simulated device          135 kHz

Summary for single channel single point AO (See the table below, DA only)
The spec of 6024E is only 10kS/s, but the loop can run at 132 kHz (the output from the port was not correct for sure). Using a simulated device (either 6024E or 6713), the loop rate is ~ 134 kHz (probably due to my low speed computer). This means the latency from 6024E devices is virtually zero. However, the latency from 6713 is way too much.

      Spec (kS/s)   tested by me (700MHz P3)
6024E    10/1             132 kHz
6713   1000/1              38 kHz
Simulated device          134 kHz

From all the above tests, it seems that s-series device (6133) and AO device (6713) perform much poorly in single point mode compared to E-series device (6024E).  I thought the difference might be due to the driver in DAQmx.  So, I want to try RLP to see whether I can gain up control loop speed to 100 kHz for my applications.

Your suggestions and advices are appreciated.

Thanks

Jin-Xing

[AlanArmstead]

Ok, I see your problem.  However, I think you might be a little confused about NI-DAQmx Base vs. RLP Programming.  NI-DAQmx Base is a top level API, similar to NI-DAQmx.  While you might see some slightly different loop rates with NI-DAQmx Base, you certainly won't be increasing by orders of magnitude.

RLP programming, using the Measurement Hardware Driver Development Kit (MHDDK) is completely separate from NI-DAQmx Base.  The MHDDK is an extremely low level driver, which leaves most of the details to the programmer.  If you are interested in achieving the fastest possible loop rates in Windows, then the MHDDK might be your best option.  However, keep in mind that its not very user-friendly, and again, it leaves a lot to the programmer (check out the MHDDK link).

-Alan A.

[50ohmTerminator]

Hi Jin,

Though DAQmx base might have slightly easier RLP interface, the overhead in DAQmx Base is greater as DAQmx Base uses LabVIEW to make the RLP calls. The DAQmx driver is written in C so it has less overhead and hence, is faster.

Single point acquisition is quite CPU intensive, would you like to consider doing a buffered acquisition with multipe points, it might make it faster overall.

Have a good one.

Malay Duggar
Applications Engineer
National Instruments

[Stuart_G]

I also have to chime in here from a different angle.  If you want deterministic, single point loop rates of 100kHz none of these solutions are going to work for you.  The BEST you could hope for even on an RT system is maybe around 70k if fully optimized, fast controller, etc.  Let me explain deterministic.  What I mean is if you want to run a control loop where you are ensured no data is being dropped.  If you don't mind a few (or on Windows) many points being dropped in single point mode then 100k MIGHT be achievable, but your control algorithm will probably be messed up because of loss of points.  I am not sure how you are measuring the rates you are seeing, but in single point mode you would have to be missing points.  I agree with 50Ohm that if you want to achieve these types of rates you are probably going to have to step away from single point and use buffering.

StuartG

[planet]

All,

Thank you very much for answering my questions.  I am sorry for replying late since I am busy with something else these days.

Wish you all a wonderful weekend.

Jin-Xing