Does a filter frequenz (2 or 4Hz) change the shape of a stepwise signal?

[anca]

I am measuring very fast temperature changes with mini tc (typ K) with a constant time smaller than 0.1s.  I have 5B47 Modules with a 4Hz Filter in  a SC2311 carrier and a DAQCard AI-16X50E.    Temperature changes from 250C up to 40C in less than one second.    I need to know if my temperature curves just only have fourier components under 4Hz.  If it is so, the shape is quite away of the real shape under this time intervall.  Is it right?  Is it possible to obtain a better shape using the same hardware?

[Logan_K]

The 5B47 is not designed for fast changing signals, which is why it has a 4Hz filter. The rise time for 80% of the full range is 0.2s (from the 5B user's manual http://www.ni.com/pdf/manuals/321029a.pdf). A step response function typically has a lot of high frequency components, so a lowpass filter at 4Hz will tend to filter out all those high-frequency components and leave you with a distorted version of the original signal. If you need to measure fast changing signals, then I would recommend a measuring device that has a higher cutoff frequency. The SCXI-1102B, for example, is designed for measuring thermocouples and has a cutoff frequency of 200Hz.

Logan Kunitz

National Instruments - Austin, TX

[anca]

Thanks for your quite precise answer.  It reconfirmed  unfortunately my supposition about be using the wrong hardware.