How to use PXI Star Trigger for PXIe-5663 in PXIe-1075 chassis

[serge_malo]

HI all,

I have this sytem configuration:

PXIe-8135 controller. Windows 7 64-bit, RFSA 2.7.5. NI-SYNC 3.4.1

PXIe-1075 chassis

PXIe-5663 (2x)

PXIe-6672 Timing & Sync Card (slot #10)

I want to trigger the recording of my Digitizer with an external trigger.

The External Trigger is connected to PFI0 of the PXIe-6672 Timing card.

Then, the PXIE-6672 card routes the trigger to the backplane of the PXIe-1072 (Destination "NISYNC_VAL_PXITRIG0")

The PXIe-5663 are triggered with “NIRFSA_VAL_PXI_TRIG0_STR” as the source.

The trigger fires my PXIe-5663 correctly, but the timing is not tight (> 5ns).

I would like to use the PXI Star trigger instead, I think that I should be able to acheive much better synchronization with this.

But NI-RFSA won't let me do this:

When I try to call

"niRFSA_ConfigureDigitalEdgeStartTrigger(rfsa_session, NIRFSA_VAL_PXI_STAR_STR , NIRFSA_VAL_RISING_EDGE)", I get the error:

"Specified Route Cannot Be Satisfied, Because the Hardware Does Not Support It"

I don't understand why the PXIe-5663E would not be able to use that Route.

Any idea?

Regards,

Serge

[JamesB.]

You're right that PXI_Star provides lower skew (< 250 ps) than PXI_Trig. Shot in the dark, but maybe this is just a spelling error...

The API help lists "PXI_STAR" as the correct spelling when calling ConfigureDigitalEdgeStartTrigger. Strangely, the ReferenceTrigger and AdvanceTrigger function use your spelling, "NIRFSA_VAL_PXI_STAR_STR."

Also, be sure you are exporting the trigger from the 6672 on the numbered PXI_Star lines that connect to the slots containing the 5663s. For example, PXI_Star6 connects to slot 3. (See page 1-9 in the manual for your chassis.) I am not an RF expert, but the 5663 looks like it is a compound device, and my guess is the start trigger needs to be routed to the digitizer board (the PXIe-5622).

[JamesB.]

---

@JamesB. wrote:

Shot in the dark, but maybe this is just a spelling error...

---

Actually, I checked around and it sounds like this is probably a documentation bug. You can still give my previous suggestion a try, but I wouldn't expect different results.

Is it possible that PXI_Star is already in use by some other route in the system?

[serge_malo]

Hi James,

Thaks for your suggestion, it made me go "deeper" in the specifications of STAR Trigger buses.

I think there is no solution for my case, here's why:

The PCIe-6672 Timing and Sync card is able to send Trig signal on the PXI_TrigX and PXI_StarX lines. But it cannot send Trig signals on the PXIe-DStar lines (Note that "Star" lines are on the PXI connector and "DStart" lines are on the PXIe connector).

The PCIe-5663 VSA has no connection on the PXI lines, its only PXIe. So, there is no way it can receive a Trig signal on the PXI Star lines.

Maybe it could receive a Trig signal from the DStar lines, but there is no such constant available on the NI-RFSA driver.

Moreover, in my case, my Timing card PXIe-6672 can not send Trig signal on the DStar. I tried to force it, but I got a NI-SYNC driver error.

Unless I'm missing something, or there is an undocumented backdoor, I don't think I'll be able to use a "STAR" trigger in my case.

Hope this helps anyone reading this forum...

[JamesB.]

That explanation isn't quite right. Usually, even PXIe modules have a connection to PXI_Star. The PXIe standard added the PXIe_DStar trigger buses, and it also preserved the PXI_Star bus from the PXI standard.

However, there is an additional twist in this situation. I'm assuming that your PXIe-5663 includes a PXIe-5622 as the digitizer. It turns out that a synchronization technique called NI-TClk has eliminated the need for our more recent digitizers to rely on triggering from PXI_Star. I was able to find some documentation that includes this information, here and here. Given that, I think you have two options that should result in better synchronization.

The first option is to use TClk; I found an example program that demonstrates using TClk to acheive phase coherent signal acquisition across two 5663s. The second option is to use cables of matched length to connect two PFI front panel terminals of the timing board (6672) to the PFI1 front panel terminals of the digitizers (5622). The timing board would accept the external trigger on PFI0 and then issue triggers on PFI1 and PFI2 with around 500 ps of skew (manual, page A-4) . The digitizers would use NIRFSA_VAL_PFI1_STR as the trigger. I hope one of these solutions will meet the demands of your particular application.

I will also follow up with the owners of the RFSA product documentation to see if we can include a note about why PXI_Star is not supported in some cases.

[serge_malo]

Hi James,

You are right: TCLK give a tighter synchronization than using PXI_TRIG0. However, it is in the order of 1-2ns. We would like to get below 1ns.
Still, I have never tried to use TCLK and an external trigger connected to PFI1 of one of the 5622. I hope this can work.

I have realized that the 5663's digitizer (PXIe-5622) is effectively connected to both PXIe and PXI buses. However, in the NI-RFSA documentation manual, page "NI PXIe-5622 F Digitizer Module Block Diagram", we see that only lines "PXI_CLK100" , "PXITrig (RTSI) <0..6> and "PXI_Trig (RTSI) 7" are connected to the card from the PXI bus. There is no indication that a PXI_STAR line is connected to the card. That would have been ideal for us.

I will try the "equal-length" cable solution as soon as I get SMB-SMB cables.

Thanks for your help!

Regards,
Serge