synchronization of two PCI 6602 using RTSI cable and traditional ni-daq

[JeffL]

As we've discussed on the phone, Paul was under the understanding that you were simply counting pulses, in which case exact synchronization is not required (only start trigger). In the case of buffered event counting where you want to count the time between pulses, that does need to be synchronized, both trigger and timebase. In this case, you can use the 20 Mhz timebase over RTSI and some sort of digital trigger.

[Paul_C.]

Hello pchemjjang,

Jeff is correct that you can share the timebase between the two devices with the RTSI cable.  However, all you really need to do is share the sample clock (the one used to clock in the edge count tasks).  Can you physically wire this to both devices?  If the sample clock is coming from one of the additional counters on one of the devices, you can route this over the RTSI cable.  

Regards,

Paul C.

[pchemjjang]

Thank you, Paul.

I kind of figured out how to do it, but I have a few more questions.

When I connect the two PCI-6602 cards with a RTSI cable, Jeff was suggesting that I use RTSI7 to share the clock (20MHz) from one of the counters on one of the cards.

BTW, is the "RTSI clock" the same as RTSI7?

In Traditional DAQ, I cannot find RTSI7, but instead I have RTSI clock, Board clock, etc...

I am trying to use Route Signal Vi to share the clock through RTSI7 this but there is no device number specification option in the Route Signal Vi.

For example, if I want to specify dev1/RTSI7 as my signal source and route that to dev2/counter0, how can I specify the device number in Route Signal Vi?

Thanks for your help again.

Regards,

Taiho

[Paul_C.]

Hey Taiho,

Yes, "RTSI Clock" is the same as RTSI 7.  In traditional DAQ, you cannot specify a device to drive a line.  You can only specify how you want to drive a line.  For example, you need to specify the signal name and the signal source.  The signal name should be "RTSI Clock" and the signal source should be the name of the clock you want to share. This will drive the RTSI 7 line with the clock you specify.  Then on the other device, you will need to specify to use RTSI 7 as the clock source for your counter tasks.

I hope this helps,

Paul C.

Message Edited by Paul C. on 09-09-2009 11:52 AM

[pchemjjang]

Thanks again, Paul.

I have a couple of quick questions.

Can I use RTSI n (n=7) instead of RTSI clock?

And when I use RTSI 7as a clock source of the other PCI-6602 card, I do not have to specify the device number, because RTSI 7 can be considered as physically connected to both cards through a RTSI cable?

Thank you so much.

Regards,

pchemjjang

[Paul_C.]

Hello Pchemijang,

I'm pretty sure you must specify RTSI Clock to be able to route your clock to drive the RTSI 7 line (remember they are the same line so it really shouldn't change anything).  As far as I remember, you just need to specify the clock source on the second 6602 to be RTSI 7.  

Regards,
Paul C.

[pchemjjang]

Thanks again, Paul.

I apologize for asking so many questions, but I am a little confused.

As you mentioned last time, "...For example, you need to specify the signal name and the signal source.  The signal name should be "RTSI Clock" and the signal source should be the name of the clock you want to share...", I think that you are suggesting I should use "RTSI Clock", not "RTSI 7" in the signal name.

But what would be the name of the clock for the signal source?

Would it be timebase n (n=20,000,000)?

Or is there anything else?

Thank you so much for your help.

Regards,

pchemjjang

[Paul_C.]

Hello Pchemijang,

After a bit of confusion myself, I decided to just grab the hardware and set it up myself.  I have what I believe is a working example of synchronizing two counters on two different 6602's.  You will see that I route the "board clock" or 20MHz timebase to the "RTSI clock" or RTSI 7 line.  Then on the second device, you need to drive the "board clock" with the RTSI clock signal.  Then, on both devices you need to setup a arm start trigger to ensure that they both start at the same time.  For my testing, I simply used an external trigger from another device.  You could externally wire a digital line from one of the counters to the PFI 0 line (the default source for the arm start in the example below).  I took a another signal and wired it into the counter 0 of gate of both counters and measured the timebase ticks between edges.  This example appears to be working quite well.  I'm only off by up to 1 timebase tick every once in a while which is probably due to some propagation delay / jitter.  Please take a look at the community example below and let me know what you think.  Hopefully, it will help you get started.

The community example can be found here.

Regards,

Paul C.

[pchemjjang]

Hi Paul,

Thank you for your help.

I have purchased and received another PCI-6602 card and a RTSI cable and another breakout box and of course another 68pin connector.

BTW, I created a new customer service request number, 7258906, and talked to Jeff briefly today.

I have installed the new card and now I have two PCI-6602 cards.

I have tried your program to see if I get synchronized countings from the counter 0 of the two cards.

The thing is that I have difficulty in start triggering the clocks (board clock and RTSI clock) at the same time.

I am using one of the counters on one board to generate a trigger pulse using the "single pulse generation" scheme in Traditional DAQ.

In fact, I have used an output of the counter 3 and routed to RTSI0 and then RTSI0 to "start trigger" using Signal Route Vi for one board.

The thing is that I don't know how to trigger the second board using the same trigger mentioned above.

Could you help me how to trigger a second board using the trigger from the first board?

As a matter of fact, I have also used two different triggers (one from each board using the timebase n (n=20MHz) as clocks that are supposedly synched), and it didn't quite work.

Thanks in advance for your help.

Regards,

pchemjjang

[Paul_C.]

Hello Everyone,

Phemjjang has contact National Instruments for support.  The issue posted last was resolved by selecting the appropriate RTSI lines to be used as the start trigger.  Just wanted to end the thread for future readers.

Regards,

Paul C.