differential sallen key fet/bjt

Hi, I soldered a proto board with the circuit in figure, which is a test bench for a differential sallen-key highpass, 1.4kHz. Signal is converted from single ended to differential, filtered, and then back to single ended so that I can test it without using a differential probe.
Purpose of this test was to study the effect of R3. Result is that changing R3 from 100k to 0R does not affect the circuit (I found a paper where they recommended R3>>R2&R4, but I could not verify that this is advisable).
So, first concern is: should I mount R3 or short it?
Paper it this
**broken link removed**
Now, the circuit works perfectly with a TL072, even with a 10Vpp signal. I tried to replace TL072 with OP200 (pin2pin compatible), but the circuit does not work: the signal is distorted, a sinusoid becomes a triuangular wave.
I imagine this has something to do with bias current, since the first is a JFET amplifier while the second is a bjt amplifier, but I do not understand what is going on.
Any clue?

el00,

The resistor R3 is necessary in both cases. It provides a DC ground connection of the non-inv. opamp input terminals, which is required for BJT as well as for FET inputs.
However, the value is not very critical - it influences the CMRR.

LvW said:

el00,
The resistor R3 is necessary in both cases. It provides a ground connection of the non-inv. opamp input terminals, which is required for BJT as well as for FET inputs.
However, the value is not critical.

Click to expand...

Of course it is necessary. I was asking if I could short it (0R), not remove the connection to gnd.
I still wonder why it does not work with OP200, BTW.
Thanks anyway for your answer.

el00 said:

Of course it is necessary. I was asking if I could short it (0R), not remove the connection to gnd.
I still wonder why it does not work with OP200, BTW.
Thanks anyway for your answer.

Click to expand...

Yes - I think, you also could short this resistor. However, i am not sure if this is the best solution - perhaps not. Read carefully the paper linked by you. They explain why the value should not be too small.
Question: What means "...not work with OP200"? No signal? no bias point? perhaps the slew rate is too bad?
At first, check if the dc bias point is correct.
Next compare signal input and output of this opamp.

LvW said:

Yes - I think, you also could short this resistor. However, i am not sure if this is the best solution - perhaps not. Read carefully the paper linked by you. They explain why the value should not be too small.

Click to expand...

Yes, they do explain in terms of CMRR; I was expecting to see it affecting the frequency response as well, but it is not. I will keep it large.

LvW said:

Question: What means "...not work with OP200"? No signal? no bias point? perhaps the slew rate is too bad?
At first, check if the dc bias point is correct.

Click to expand...

I think you are right: TL072 has a slew rate 100 times > than OP200. At low frequencies they both work (and dc bias is correct for both). OP200 starts distorting large signals (10Vpp) when frequency goes over 3kHz.