Inaccurate counts from DAQ boards

[Tom G.]

It seems like every time I need to count a signal with an E-series DAQ board the signal does not have a fast enough rise time to get an accurate count. The counter always seems to 1) double the frequency or 2) add or subtract a few "phantom" counts. The singals I see this happening with have rise times often as low as 5-6 microseconds. How fast does the rise time have to be to get an accurate count? Using a Schmidt trigger circuit on the front end to filter the signal always seems to solve the problem- it seems like this should probably be a common issue so why does NI not offer an SCC signal conditioning module to filter count signals?

[Filipe A.]

Tom;

Unfortunately NI doesn't have a SCC module to filter TTL pulse train.
You can instal a pair of TTL inverters (7406) in series to increase slope of the rising edge. The problem occurs because TTL has an undefined region between .8 and 2.2V. Since the DAQ STC is very sensitive, you can actually get multiple triggers within this region. An even better option is the National Instruments 6602 card with filtering on all inputs built into hardware.
Hope this helps.
Filipe

[PdB]

What you describe is a choice made in the designing of the DAQ-board, perhaps not favoring every application.
In the past I had 2 projects encountering false counts.
Thanks to the help of my collegue electronics expert extra electronics (combination of low-pass filtering and a smitt-trigger) did the job.
But just like grounding problems in data-acquisitioning, there is no solution to fix all problems.