Labview data fast data transfer

[vasylko]

Hey guys! I have a USB6363 NI card. With 4 AOs I generate 4 sinusoidal voltages which I apply to a piezo element. Simultaneously via a counter I acquire a signal at 1MHz rate from a PMT. My problem is that these two processes in parallel seem to take a longer time to perform,  than expected. What could it be? Would a PCI card speed up this process? Or should I reconsider my Soft?

[altenbach]

---

@vasylko wrote:

Or should I reconsider my Soft?

---

While I don't know what "soft" is in this context, I am guessing you are talking about your LabVIEW program. You don't give sufficient information on how you are doing things. How do you measure the difference between "actual and expected time", you don't even say what exactly you expect.

Can you show us your code? It is all express VIs or tightly coded low-level stuff? Is there execution dependency between AO and the counter?

[Gregory]

What is a PMT? I looked up your device and it is capable of 2.86 MS/s, which is 715kS/s if all 4 channels are being used simultaneously. I'll assume you want to sample at least 10 points per cycle of your sinusoidal signal, so you can output sinusoids at frequencies of about 70kHz. Where is the 1MHz coming from?

[altenbach]

---

@Gregory wrote:

What is a PMT?

---

Photomultiplier tube. The output are pulses counted by a counter..

[Kevin_Price]

I'll venture that you can probably improve the behavior via programming, but I'd also caution that you're more limited with a USB device than you would be with a similar PCIe card.  I'd very much expect your app to be feasible on a PCIe-6363.  I'd be much less confident with USB.

I agree with altenbach that you should post your code to get specific pointers. 

-Kevin P

[vasylko]

Sorry for being imprecise. 

Here I added the labview program which I didn't make myself, I just made some modifications.

What we do, we generate a sinusoidal voltage with 4 AO.  We apply these four waves on a piezotube which performs resonant oscillation of a fiber inserted in the piezotube. So, we first generate an expanding wave form during 100ms, then we stop the fiber in 15ms and keep the fiber free, Voltage is zero here, for 10ms. Then we repeat the same  thing. So we expect to be able to generate 8 oscillations per second((100ms+15ms+25ms)*8).

On the other hand, we connect a photon counter to the same NI card. We acquire signal at the same time as we scan the fiber (at 1Msamples/sec). However, in 1 second we do around 4 iterations instead of 8. There is something which slows down this process. It could be either the Labview program or a slow data transfer from the NI card to the PC, though I plug it into a USB 3.0 port. Please tell me if you need some additional information. Thanks a lot

[Gregory]

You've only attached a picture, so we cannot see much of your code... can you attach the VIs in a zip?

[billko]

[vasylko]

here you are! 

[Gregory]

Well it seems there are a lot of features you can turn on/off as the program is running. For example, you can write as many as 10 text files per iteration, which would add quite a bit of overhead. 

If you really want to get close running your DAQ at full speed, you would have to offload your calculations and file logging to a separate loop.