Motion Control and Motor Drives

cancel
Showing results for 
Search instead for 
Did you mean: 

Difference in latency between PCI and USB data acquisition board

Hi,

I am wondering if there is a difference between the latency involves when using a PCI data acquisition card and a USB acquisition card.

I am getting analog signal from a force sensor and hesiting between the 6034E PCI card and the USB 6259 card.  It also might determine whether I'll use a labtob or a desktop computer for my application.

Is there a very well known difference for the latency between those 2 technologies?

Thanks,

JBM
0 Kudos
Message 1 of 8
(13,694 Views)
Hi,

yes, there is a significant difference in the latencies for USB and PCI. Please have a look at this picture that shows bandwidth vs. latency for current bus technologies:


I hope that helps.
By the way: It would be better to post this type of question to the Multifunction DAQ forum.


Best regards,

Jochen Klier
National Instruments Germany

Message Edited by Jochen on 01-23-2007 03:55 PM

Message 2 of 8
(13,693 Views)

Is there any way I can check it by myself by writing some example code? I couldn't find an example code to do this. I am using USB 6341.

0 Kudos
Message 3 of 8
(7,550 Views)

Hi Rafii,

 

measuring bus latency is not a simple task and I don't think you can accomplish this without low-level debugging tools.

With only the 6341 in place you could try to measure the overall system latency which also includes all associated driver layers and buffers.

 

In general bus latency is not a problem for streaming applications and it doesn't lower the data throughput. For any type of single point closed loop application (e. g. acquire a voltage value - calculate a control algorithm - output a voltage value) bus latencies are limiting factors for both, control loop frequencies and deterministic behavior.

 

Regards,

Jochen

0 Kudos
Message 4 of 8
(7,528 Views)

Hi Jochen,

The figure is very informative. Thanks for sharing. Could you also share the source of this figure? I'm curious how the measurements were done and how the latency were defined. Thank you.

0 Kudos
Message 5 of 8
(7,060 Views)

I don't know if the figure is in this video, but it does discuss and compare various busses if I recall:

 

https://ni.adobeconnect.com/_a56821929/p21882777/?launcher=false&fcsContent=true&pbMode=normal

 

-AK2DM

 

Not sure if that link will work, just noted it has a https connection. Let me know.

 

Edit- The Figure appears at time index of 5:22.

~~~~~~~~~~~~~~~~~~~~~~~~~~
"It’s the questions that drive us.”
~~~~~~~~~~~~~~~~~~~~~~~~~~
0 Kudos
Message 6 of 8
(7,051 Views)

Here's a link to the webcast in case my original one wouldn't work:

 

http://zone.ni.com/wv/app/doc/p/id/wv-183

 

And here is the white paper:

 

http://www.ni.com/white-paper/3509/en

 

Note that these do not cover the newer USB 3,0 and the x8 and x16 PCI variants.

 

-AK2DM

~~~~~~~~~~~~~~~~~~~~~~~~~~
"It’s the questions that drive us.”
~~~~~~~~~~~~~~~~~~~~~~~~~~
0 Kudos
Message 7 of 8
(7,047 Views)

Thank you! The paper and the vedio do give more details. But they don't explain how these data were acquired, whether they were based on measurements or theoratically calculated. I'm especially curious in the latency of USB 2.0. The 125 us latency of USB 2.0 seems to be related to the size of a microframe, which is also 125 us. But the latency of a bulk transfer will be restricted by the size of microframe. Could you explain the latencyof USB 2.0 in more detail? Thank you.

 

0 Kudos
Message 8 of 8
(7,040 Views)