NI PXIe-6535/6/7 as pseudoclock

[Henry Ando]

Hello,

We are setting up a new NI PXI system with labscript. We will have some digital cards (PXIe-6536) and some analog cards (PXIe-6738). In the past, we used custom labview software which effectively used one of the digital output channels as a pseudoclock for the analog cards. I was wondering if there is support for doing something similar in labscript, or if anyone has set this up. I guess it seems a bit contrary to the labscript ethos, but it would be nice not to have to set up a separate pseudoclock. If we do need to, we will just make a prawnblaster, but the digital cards are capable of handling our full sequences with good timing resolution so we were hoping to avoid introducing additional hardware.

Thanks,

Henry

[dihm....@gmail.com]

Henry,

This is a somewhat common request, especially considering that everyone's analog cards also have Digital output lines that could presumably be used. The reason it isn't supported comes down to the original labscript development, which I wasn't around for, so I can only speculate. Ultimately, I think that arbitrary digital output streams for NI devices wasn't very performant back then (especially for devices most labs had, like the 6229 or 6733 which had 1MHz max rates). Since people would need to buy a dedicated device anyway, it was preferred to buy the pulseblaster since it was much cheaper. And since original development centered around the pulseblaster, pseudoclocks were developed around its limitations. Now that pseudoclocks form the backbone of labscript timing, porting back to a bit-streaming like platform such as the NI Digital I/O (admittedly I'm assuming higher level looping instructions don't exist here) is a little annoying and as far as I know no one has seen fit to make it happen. 

I'd certainly support including a NI-based pseudoclock as another option, especially if it supports using the admittedly low performance D I/O like using the using your PXIe-6738. Then one could buy a single DAQ and have everything they need for an experiment that isn't too demanding.

On a possible bright side, we are on the cusp of updating the prawn firmwares to increase their internal clocks to at least 200MHz, possibly 300MHz for the Pico2. Given the 5 clock cycle minimum, that means max rep rates of at least 40MHz will be possible meaning you would actually have a timing upgrade moving to the prawns.

Anyway, hopefully I'm wrong and someone else chimes in with functioning code for you.

-David

[Ian B. Spielman]

This decision predated labscript, and the reasoning was simple: a PB like device only needs to know when events change. This leads to fairly compact memory requirements, since for most outputs, most of the time nothing is changing. By contrast old NI DO devices (which were plenty fast) required their outputs to be defined every clock cycle. At the time transferring this amount of information for a 40-100 MHz clock for all of the different output and/or “pseudoclock” lines was an impossible ask.

— Ian

> --
> You received this message because you are subscribed to the Google Groups "the labscript suite" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to labscriptsuit...@googlegroups.com.
> To view this discussion, visit https://groups.google.com/d/msgid/labscriptsuite/e734b519-9841-4c50-aac0-2172994c0a73n%40googlegroups.com.

[Henry Ando]

Dear Ian and David,

Thanks for your responses and the insight, this is helpful context! I can see why nobody has bothered because the work and cost of a prawnblaster is so much less than writing that code. I think we will go the prawnblaster route too :) waiting for parts now. Cheers and thanks again!

Best,

Henry