> testing in our EMC lab, we apply bursts of pulses that are only 50 nS
> +/-30% wide. Rise times are on the order of 5 nS, and the bursts last
> 15 mS. Of course, this noise is presumed to have been generated by
Antennas and systems we use have very poor (if effectively any) response to
F>20MHz signals. Any pulse is a series of sinewaves of various amplitudes
and frequencies. In this application higher frequency sinewaves can not pass
through any part of the system without rapidly increasing attenuation with
increased frequency. This rounds and filters pulses, leaving only lower
frequency energy. A 500Hz filter limits reponse time to the order of
milliseconds or longer, not nanoseconds. That's where all the "pulse
stretching" or group delay errors are.
> Let's keep the question open a little longer. Do the coherent signal
> tests fully-define preamp behavior. Could there be some value in
> tests that evaluate how preamps handle random, broadband signals? The
> impulse (or transient?) response test is my first stab at an answer.
> Suggestions anyone?
I'm not sure why we should test an amp for that, since later stages (the
receiver and our ears) would have an absolutely overwhelming effect.
A problematic amplifier system would show distortion products in the normal
tests at a level worse than the receiver front end on the same wide and
narrow spaced tests. By the time the power level is high enough to
gain-compress or overload a reasonably good preamplifier, the rest of the
system (including the S/N ratio of the desired signal and our ears) are
toast.
73 Tom
|