Re: gEDA-user: Any DIY USB Scope project on schedule? Or some recommmendation?

[Daniel Nilsson <daniel@oden.homeip.net>]

On Tue, Nov 23, 2004 at 04:51:58PM -0800, Samuel A. Falvo II wrote:
> On Tuesday 23 November 2004 04:50 pm, Daniel Nilsson wrote:
> > Regarding square waves, the frequency of the square wave doesn't
> > matter from a measurement perspective. What matters is the rise and
> > falltime of the signal that you are measuring. To calculate the BW
> > based on the risetime a good approximation is BW=0.35/tr where tr is
> > the fastest out of rise or falltimes. For example, is you signal has a
> > 10ns risetime, the bandwidth requirement would be 0.35/10e-9 =
> > 35MHz.
> 
> This quantifies exactly what I was saying.  In order to capture those 
> nice, clean, sharp edges, you need a wider scope bandwidth to grab the 
> odd harmonics.  And you are correct; the sharper/steeper the edges, the 
> wider the bandwidth required.
> 
> But if you measure a 250kHz square wave, and the analog bandwidth of the 
> input section to the scope is only 250kHz, then what you'll end up 
> finding on the screen is a 250kHz sine wave -- just the fundamental, 
> with no harmonics (or, at least, substantially attenuated harmonics).
> 
> It's a handy formula to have -- however, I'm curious though: where does 
> the factor of 0.35 come from?

Technically, it comes from taking the fourier transform to convert the
signal from the timedomain to the frequency domain and setting a cutoff
energy level from which you say that you are not interested in
frequency content that contain less then a certain amount of energy.

In reality, the value 0.35 comes from years of working with simulation
of digital signals where we look at both the timedomain and frequency
domain, the range from 0.35 to 0.5 was agreed upon as the factor
to use if you want to get the majority of the signal converted from
timedomain to frequency domain without having to actually perform the
FFT transform. "Rule of thumb" kind of thing...

-- 
Daniel Nilsson