Hi,

I'm trying to design a active 6 pole Bessel filter with a 3.5MHz cutoff frequency to use as a premod filter. I've come up with a design on paper using the sallen-key topography, but I don't think it will work in practice, as it uses 10pF capacitors, and im pretty sure the capacitance of the circuit board will be on the order of 10pF.

Has anyone here dealt with this issue before? Is there a way of scaling up the values on a filter to get more reasonable capacitance values?

Thanks,
-ntropi

 

Are you trying to build an active filter (op amp based)? Tough or impossible to make that work at 3.5MHz.
I would use a passive filter. Look at AADE Filter Design (free program) for help...

 

It sounds problematic to use an active filter for that high a frequency but what are the resistor values of the filter? Lower resistor values give higher capacitor values for a given frequency. Post the schematic.

 

I scaled the seed resistor value to 1K and thereby increased the capacitances. However, to convince you that this is not a practical design, look at the required Gain Bandwidth products for the opamps...

 

I scaled the seed resistor value to 1K and thereby increased the capacitances. However, to convince you that this is not a practical design, look at the required Gain Bandwidth products for the opamps...

Would a Sallen-Key filter help the op amp bandwidth requirements?

 

Would a Sallen-Key filter help the op amp bandwidth requirements?

Nope, made it worse...

 

Nope, made it worse...

Intgeresting. I did a similar Salen-Key filter using the Filter Wizard on the Analog Devices website and it uses AD8032 80MHz op amps for the same filter response. Seems like a rather large difference in requirements.

 

The filter I had come up with was pretty similar to Mike's, with the resistors in the range of 1-3k. See attached.

As for the GBP, 714MHz is pretty high, but TI and Linear Tech make several in the GHz range. I was initially planning for the LTC6253 at 720MHz. Looks like it won't matter anyway if I end up going to a passive filter...

In any case, I really appreciate your input.

 

Below is a simulation for your circuit and the response looks good. I doubled the cap values and halved the resistor values to make the circuit less sensitive to strays (but you will still likely have to tweak the values when the circuit is built). I used the LTC6254 which is the quad version of the op amp but you can use the dual LTC6253 if you like which may simplify making a good layout.

Note that I used dual supplies. If you use a single supply you will need to offset the input DC level by 1/2 the supply voltage.

Building this circuit will require careful layout with all connections as short and direct as possible, using a ground plane, and connecting 0.1μF surface-mount ceramic caps directly from the op amp power pin(s) to the ground plane. The layout is the most critical part of getting the circuit to work properly. It will not work with a solderless type breadboard.

But certainly a passive filter for this would be easier to get working properly.