BUF602: Using BUF602 for a unity gain sallen key LPF

Hi Nir,

I would try to carry out a phase stability analysis first.

Kai

Hi Kai,

Thanks for the tip

Though I don't fully understand what you refer by phase stability analysis

Do you mean stability under capacitive load?

Anyway, I've simulated the circuit (there are some slight changes)...any thoughts?

Thanks!

Nir

Hi Nir,

firstly, to confirm that the simulation is also running in TINA-TI:

nir_buf602.TSC

Kai

Hi Kai,

Thanks for double checking the design

Regarding the response graph, something is weird with the frequency scale...the -3dB point should be around 30MHz

What about the phase stability analysis you've mentioned?

Thanks again,

Nir

Hi Nir,

the frequency response is correct. It's only a scaling issue:

Kai

Hi Kai,

I will try to build and test it in real life conditions,

Thanks for the simulations!

Nir

Hi Nir,

you can increase the stability a bit by inserting a small resistance between C4 and the output of BUF602. !0...20R may already help. But this will be at the expense of filter curve, of course. The curve will be less steep. You can easily simulate this.

Kai

Hi Kai,

I've tried something else...I have no problem inserting another filter stage, so I've reduced the capacitors value, and pushed the cutoff frequency around 46MHz in order not to attenuate the 1st 32MHz harmony.

Can you please run the phase stability analysis with the new design?

Thanks for helping!

Nir

Hi Nir,

the gain margin is 3.9dB now. So seems to be a bit more stable.

Have you already thought about using a passive LRC low pass filter?

Kai

Thanks Kai,

Output buffer is required because my input source cannot handle 50ohm load. a passive filter would not work for us.

Hi Nir,

why not using an arrangement like

"BUF602 + LRC low pass filter + BUF602"?

And the LRC low pass filter need not necessarily to be designed in 50R technique. 200...300 Ohm could also do.

Kai

Hi Kai,

Thanks for the tip...something like this?

Hi Nir,

yes.

You can find tons of suited passive RLC low pass filters in the www.

Kai

Hi Kai,

Great, I'll try it...thanks again for your support!

Nir

Good luck

Kai

I was looking at your RLC filter - looks like 2 2nd order stages just cascaded? 

Here is the 1st stage by itself, 

And then both together, so I guess that looks ok for a 30Mhz target, 

Here is a more classical 3rd order 0.2dB Chebychev, I prefer a small series input and much large final R - you could recover that gain loss using a 2nd op amp, like the OPA695. 

And putting this into TINA, more standard rolloff - you will notice at 100MHz this is 30dB down while the 2 2nd order stages are only 20dB down

Hi Michael,

Thanks for sharing your knowledge

This is the cheapest 10dB I have ever gained 

Can I increase the R4 500R resistor to 5k? this would help a little with the in-band loss.

I really don't want to add a new amplifier to my design just to compensate for the in-band loss...

Many thanks!

Nir 

Hi Nir,

Nir Alon said:

Can I increase the R4 500R resistor to 5k?

Run a simulation and you will see

Kai

I did

Looks ok...but I'm not sure if I didn't messed up other parameters...

Hi Nir,

(especially) in HF circuits it's usually a good idea to keep the source impedance low. This is the case for very low frequencies (because of L2) and very high frequencies (because of C3). But at the resonance frequency of L2 and C3 a peak of high impedance could be seen by the input of OPAmp which can make the circuit sensitive to noise and EMI. R4 will dampen this resonance and limit the impedance peak to 500R or 5k in your case.

Run a simulation showing the input impedance seen back by the OPAmp input.

Kai

Hi Kai,

Understood, Thanks again!

Nir

Hi Nir,

as you can see there's no relevant difference between 500R and 5k:

nir_buf602_2.TSC

Kai