Please check this filter for me

I want an anti-aliasing filter that's fairly flat from almost DC to 10MHz, with a cut of at 15MHz. The signal is about 600mV peak to peak, offset above ground with the centre at 345mV.

Please will someone take a look at my filter and check that I've got it right? Also, is the AD8502 a reasonable choice of op-amp for this? It looks it from the data sheet. I hope so because I already have a few of them in my bits box.

Dear FoxyRick
Hi
Did you simulated it ? i think if you want to be ensure with your design , you'd better to simulate it .
Best Wishes
Goldsmith

Looks like you are using a Sallen-Key Lowpass filter structure with 3 stages, so it is a 6th order filter. What kind of polynomial approximation are you using for the filter (Butterworth, etc)? If this is a 6th order Butterworth lowpass filter, I would expect the Q of the filter stages to be approximately (0.52, 0.71, 1.93). I just want to first check if this is the way you designed the filter. I can double check the values based on that.

Best regards,
v_c

v_c - that's exactly right - but I've messed it up I've just realised. No wonder it didn't behave in the simulation.

Back to the drawing board (and the maths book...)

You can start with each section normalized to 1 rad/s. This will have both resistors at 1 ohm and the capacitors at 2Q and 1/(2Q) for each section. Here, the 2Q is the capacitor feeding back from the output into the midpoint of the resistors and the 1/(2Q) capacitor is the one from the (+) terminal to ground. You then need to frequency scale the capacitors by dividing them by kf=2*pi*15MHz. This will make the 3dB frequency 15MHz. Then you can magnitude scale by km=1000, to make all resistors 1k. The capacitors also need to be further divided by km=1000 as well. Frequency scaling only changes the capacitor while magnitude scaling changes both resistors and capacitors. Therefore Cnew=Cold/(km*kf) and Rnew=km*Rold. I can send you further information if you wish.

Best regards,
v_c