[sdiy] Reality versus simulation?

[Tom Wiltshire]

Hi All,

I've been playing with the following circuit. It's a couple of simple passive filters to follow a digital noise generator. The LPF is set around 25KHz to smooth the digital pulses, and then the HPF acts as a DC blocker at about 15Hz.

http://electricdruid.net/images/PentaNoiseFilter.pdf

Now, the interesting thing is that in simulation, this works fine, and I have a flat frequency response between the two filters. However, in reality, there's a pronounced downward slope to the frequency response (high frequency rolloff) which amounts to about -10dB at 25KHz. This is measured by feeding the noise into my oscilloscope's FFT function. I know the rolloff isn't coming from the 'scope or the FFT, since feeding the digital noise in directly gives me a flat response out to 100KHz or so.

So what's going on with the real-life filters? Can I improve their performance?

Thanks,
Tom