Measure Pulse Width or Period VI gives inconsistent pulse widths

[GLV]

Using NI 6036E board, this VI gives me accurate single pulse width measurements most of the time, but not always. I suspect noise is falsely gating the measurement either on too soon or off too soon. I am measuring a TTL signal. I see some noise using an oscilloscope but it doesn't seem to be enough. Anyone know what the gate thresholds are for the counter gate inputs? Do people normally install a schmidt trigger or other signal conditioning in front of the incoming signal? Or has anyone come up with another solution?

[JB]

User manual (page A-6 Timing I/O) --> TTL compatible

I normally use a schmidt trigger with a RC lowpass filter on its output.

The PCI-6602 counter board has onboard digital filters.