6015 time delay

[Fred Costello]

We are using the subject 6015 DAQ, driven by LabView. We are trying to drive a vibration table to the multi-frequency input accelerations.  We need to know if, in the range from 0.5 to 2000 Hz, the output voltage and/or phase differ from the input values.  We are particularly concerned that a time delay may be introduced that is a function of frequency such that the interplay of the various frequencies is not the same as in the input values.

[PBear]

Hi Fred,

Let me try to understand your problem a bit better. Are you wondering if their is a delay from acquiring a waveform on a single AI line and then outputting it on an AO line? Or maybe you're asking if there is a delay between analog input channels? A little more clarity could go a long way here to finding an answer to your situation. Perhaps you could elaborate on what your program does, what pins are in use, and define the inputs and outputs of your device and 6015?

The DAQPad-6015 is a multi-function device that is connected to the computer via USB. This brings up two forms of delay that you could be seeing. One is that on our E-Series and M-Series line of DAQ devices, the analog input channels are multiplexed into a single ADC. This means the channels are not simultaneously sampled and there will be a slight phase mismatch between all sampled inputs. More information about this can be found here.

Also, the USB interface does add in latency that you would not find on a PCI device. This is usually not an issue if you are setting up the device to do a certain task and letting it run with out sending interrupts. If you are constantly communicating with the device and determinism is an important aspect of your program, you could see latency between your computer and the device. While the USB standard has solid bandwidth, the latency of the interface needs to be noted. This is expected behavior with all USB devices on the market.

[Fred Costello]

PBear,

Thanks for your response.

We are using the 6015 to convert from digital to analog.  The digital data, each value being 0.25 ms apart, comes from LabView.  We want to know if digital points will fall exactly on the analog wave form if they are plotted together.  If not, what would the differences be.  We can tolerate a uniform time offset but not a time offset that would distort the wave form.

[PBear]

Hi Fred,

It sounds like you are outputting at a sampling rate of 4000Hz. Is this correct? Am I correct in assuming that you have an indicator on your front panel that shows the waveform data you are wishing to plot and that you are seeing a delay from the computer display to the output on the card? If you are writing your data to the buffer every loop iteration (and turning regeneration off), then you will inheritly have a small delay between the display and what is output. Using loopback, you can see this delay and acount for it. If you are writing to your buffer and then starting the task (regeneration on), then there will be no correlation between your display and the analog output because you're just displaying data that has already been written to the buffer and is being displayed over and over.

KnowledgeBase 385AIBQ8: Analog Output Regeneration Modes