Suitable hardware configuration for 24-bit >50kHz feedback control loop?

[maoz]

I posted this question elsewhere, but this seems to be a more suitable forum.
I am looking for a hardware solution that will allow me to implement a fast (>50kHz) feedback loop, with 24-bit DAC/ADCs.
The application is feedback control of an STM (Scanning Tunneling Microscope).

For I/O, it seems that the only way to go is a NI-4461 PXI card.
I see the PXI-8196 RT controller is rated at 57kHz single PID loop in the catalog, and this may be sufficient for my purposes. However, I am worried that any extra workload will lower the loop rate, so I would like to have some "spare" CPU cycles. The PXIe-8105 seems to have a more powerful, dual-core CPU (but no loop rate numbers are specified for the Windows-based controllers). In realistic conditions, would the 8105 controller be the better choice? Is there any option to have the real-time OS run on a more powerful CPU like the one on the 8105?

Even better, can an FPGA (on one of the R-series PXI cards) be programmed to perform I/O via a NI-4461 A/D card on the same PXI chassis, without that data having to go through the PXI controller? If so, what sort of PID loop performance can be expected from such a system, compared with the RT PXI controller option stated above?

[CKap]

I posted a reply here.

Regards,

Chetan K

AE-NIC

[Jimmie_A.]

Hello!

I think that you definitely should contact sales so they can advise you on the best hardware for your specific application. 8105 dual-core PXI controllers doesn’t support RT as of yet so the PXI-8196 is the most powerful PXI controller that supports RT today.

When it comes to the number you mentioned in the catalogue I am not sure if the tests they refer to actually include PID control. LabVIEW Real-Time Module 8.20 is faster than LabVIEW Real-Time Module 7.1 and the PID VIs are greatly improved in the 8.20 release as well. Numbers on this can be found here:

http://zone.ni.com/devzone/cda/tut/p/id/3577