Synchronizing cRio chassis

[bominze]

Hi,

we would like to aquire data with 25 kHz and the following setup:

-1 x cRio 9024 (1 x 9239; 4 x 9233)

-6 x sbRIO 9267 (7 x 9205)

-1 x 6221 with a lower sample rate

To synchronize the data we have the idea to run the 9024 as Master and export a 25kHz trigger signal on the Digital output from an additional 9205. On the other chassis we then import the signal through the digital inputs on the chassis. The code on the fpga we would build like in the NI example (see attachment).

About this setup I have some questions:

1. Are the Digital Inputs on the 9267 chassis and 6221 capable of importing and processing a 25 kHz signal?

2.Is this way of synchronizing my equipment the best way and will it work?

3. Another idea is to send some kind of signal (e.g. square wave signal) to each chassis and synchronize data through post processing. How you would evaluate this method in comparison to the live synchronisation (accuarcy,implementation complexity,best practice)?

[JGoebel · Accepted solution]

Hi bominze,

this is Jan Göbel from NI technical Support.

Actually your first attempt seems pretty good. I havent found a max frequency for the Digital Output in the 9205-Manual.

But 25kHz should not be a problem for any of our Digital Output products, normally the Limit is more in the MHz-Area.

Its the same with the inputs. I havent found a spec for this, but still 25kHz should for sure be not too fast.

For the last question, I'm not quite sure whats the difference. You always need to synchronice the data through post processing for both approaches. In the first one you use one of the devices as a Master-Clock, for the secont one you would use an external Master. There shouldnt be a difference, but I would go with the first one, since you don't need an additional Frequency Generator.

If you just use a 25kHz clock on each chassis, without sharing it, nothing is synchronized. If you use at least a start-trigger, but not a synchronized clock, the clocks might drift which will cause a non synchronized measurement as well.

Hope this helps, if you have any further questions, just answer to this post.

Best Regards,

Jan Göbel

Staff Application Engineer

National Instruments

[bominze]

thanks for your quick reply. It's good to know that the digital in- and ouputs can handle the 25kHz.

You mentioned that I need to synchronize the data through processing in both approaches. My intention with the first approach was to eliminate the post processing. Isn't it possible to send 25.000 "trigger signals" per second from the master chassis to the slave chassis and not just one trigger signal to start the data aquisiton? In this case every chassis would read the channels at the same time right? Therefore the data should be synchronized.

[JGoebel]

Hi bominze,

oh, you are right with that. Thats what I recommendet. By postprocessing I just ment sending the Data back to the master and align it to the datapoints there. So exactly what your plan was.

Best,
Jan Göbel,

Staff Applications Engineer

National Instruments