You can find some differences from a very careful reading of section on Digital Filtering in chapter 3 of the 6602 manual and the section on PFI Filters in chapter 8 of the X-series manual. But it takes a *very* careful reading indeed. And I for one don't find the explanations to be entirely clear about the distinctions between filter clock, filter clock timebase, resolution, and delay. For example, the programmable setting in the 6602 table seems different than the others, perhaps due to the sort of ambiguous terminology used to configure these filters. The X-series table looks more consistent.
My layman's explanation of things goes like this:
- when you pick a filter width, your legitimate transitions will be delayed by at least 1 and not more than 2 widths. This is true for both boards.
- with the 6602, your measurement *resolution* is also based on this width. Resolution is set at 1/2 the width.
- with the X-series board, resolution is determined by the filter clock timebase. This is normally the 100 MHz internal clock, so you can get resolution down to 10 nanosec. Delay is still governed by the filter width you defined.
Further attempts to wrap this up: For both boards, resolution is determined by 1 period of what they call the filter clock. But they don't refer to the same kind of thing as the filter clock.
On X-series, the filter clock is defined as the internal 100 MHz timebase. Then that gets used to support filter width settings that are pretty much any integer multiple of its period. Delay is governed by filter width, resolution is one period of the 100 MHz timebase.
On the 6602, the filter clock is defined as the *derived* clock with a period equal to 1/2 the filter width. Delay is governed by filter width, resolution is one period of this *derived* filter clock, i.e., 1/2 the filter width.
Final thoughts. The 6602 (and its manual) came out before DAQmx. I cannot test, but it's conceivable that the DAQmx driver and API support methods that would make both kinds of boards behave similarly. It'd be great if the 6602 filters could be made to behave more like the X-series. But you might instead be stuck "dumbing down" the X-series board to artificially limit its resolution and make it act like the 6602.
Actually, answering this question helped me out quite a bit too. I had carried around vaguer, impression-like ideas that the 6602 filters not only delayed edges but also quantized their timing resolution, making their usage more of a compromise. Now it's more clear exactly how and why that's true. And it's also clearer how the X-series board's implementation is better, much more like I had wanted from the 6602 in the first place.
-Kevin P
CAUTION! New LabVIEW adopters -- it's too late for me, but you *can* save yourself. The new subscription policy for LabVIEW puts NI's hand in your wallet for the rest of your working life. Are you sure you're *that* dedicated to LabVIEW? (Summary of my reasons in this post, part of a voluminous thread of mostly complaints starting here).
