Communication between FPGA and RT

[belaji93]

Hello,

Acutally, i have CRIO 9030, with 2 C  modules NI9205 for measuring the voltage. 

I'am doing the aquisition of 64 channels (32 for each modules) and i am using the method of interleaving in order to send data from FPGA to RT with the help of DMA fifo of course. 

The problem is that my module NI 9205 is limited and can aquire 250kech/s. 

So, i might lose the signal  because i'am trying to measure some signals with high frequency.

Here is my question, can i use 2 DMA in place of 1DMA to increase my loop rate of FPGA vi? or that will not help? 

Also please do you have an exemple of FPGA vi with two DMA-fifo in order to cummunicate with RT ? 

Thankyou so much for your help.

Best regards,

BELAJI  

[GerdW]

Hi belaji,

how should using 2 FIFOs instead of just 1 help to increase the modules sample rate?

It's a limitation of that module, but not of the data transfer between FGPA and RT…

[belaji93]

Hi GerdW,

Thankyou for your answer.

So basing on your experience, adding 2 Fifo will not help ? 

In this case, what is the interest of having two fifo to communicate with RT ?

Do you think the method of interleaving is a better choice to send data from many channels ? 

Beest regards,

BELAJI 

[GerdW · Accepted solution]

Hi belaji,

So basing on your experience, adding 2 Fifo will not help ? 

It will not help with your current issue!

That NI9205 module does only support sample rates of <=250kS/s…

In this case, what is the interest of having two fifo to communicate with RT ?

You can use as many FIFOs as your RT target supports.

You can use those FIFOs to share data between RT and FPGA.

Do you think the method of interleaving is a better choice to send data from many channels ? 

In which way better?

When you send samples from 2 NI9205 module you are sending 500kS/s through the FIFO: both the FPGA and the RT need to handle that. It doesn't matter if you use 2 FIFOs or just one: the RT target still has to handle those 500kS/s…

[belaji93]

Hi GerdW,

Thankyou so much for your answer.

You have absolutely enlightened my issue.

If you allow it, i still have the last question.

Ok, if we suppose that we are getting data from 64 channels (2 NI9205). 

Each module is aquiring data at the rate of 250KS/s. 

So, as you said in you last comment, we are getting data at the rate of 500KS/s for 2 modules.( That is normal because we are taking measurements from 64 channels).

Do you think Crio9030 can handle 500KS/s ? 

I'am sorry for this stupid question.

Your faithfully,

BELAJI 

[RavensFan]

You haven't actually provided a key piece of information.

You are concerned that your modules can only go up to 250 kS/sec.  But you haven't told anything about the types of signals you are trying to receive and what information you want to get out of them.   What frequency do you need to actually acquire?

[belaji93]

I want to measure analog voltage with a frequency up to  400 Hz. 

And, i'am reading informations from 64 channels ( 2 NI9205 ). 

[RavensFan]

So why are you worried about 250 kS/sec?  250 kS/sec across 32 channels allows you to have about 7800 samples/sec per channel.  To measure a 400 Hz signal, you only need to acquire at 800 Hz per Nyquist.  You may want to go higher, but you've got room to go up about 20 more times.

PS: "informations" is not a word.  It is a non-countable object.  "Information" is basically its own plural whether you are look at one piece of information or multiple pieces of information.

[belaji93]

You need to say at least hi !

I am not here to correct peoples' grammar and thankyou anyway for your answer.