High Pass Active Filter 4th - Sallen Key A=1 - Bessel approximation

[jeremieblanc]

Hello,

i have problem with simulation model (PSpice) - high pass active filter 4th, sallen key, bessel approximation. I need a graph of the phase shift and i always used to function P(V(name)). For Low Pass filter (3,4,5.....row) and High Pass (2 row) all is OK. For this model i calculate +120 phase shift for 1 kHz. FilterPro also gives this value, but pspice give -240 deg. If i write P(V(Vout))+360 or P(V(Vsr))+P(V(Vout)/V(Vsr)) (first stage + second stage) i had good graph. Maybe somebody know why PSpice give me wrong result ???
https://imgur.com/a/J4K2n

 

[MikeMl]

Welcome,

Try it first with an "ideal" opamp, instead of that 40 year old toy. To make this filter work, you simulate first using an opamp with a higher GBP... Once the circuit simulates with an "ideal" opamp, you can try some modern opamps...

And next time, post your graphics here. I resent having to look at all the b.s. advertising at Imugr...

 

[jeremieblanc]

MikeMI thanks for your answer and comment (I apologize for the photos). I try with "ideal" opamp when i leared realized a lower rows - differences only for a high frequences where "real" opamp have limiting factor (ratio - amplification*bandwidth) and you have to use vdc +15/-15 to power opamp. So that is wrog clue,contemplation.

 

[MikeMl]

Using LTSpice and ideal opamps with a GBP of 1e7:

If I set the GBP of the two opamps to 1e5, then it changes the phase shift...

 

[jeremieblanc]

Thanks for your answer and time. In PSPice OP have gain 1e6 when i change to 1e5 or 1e7 nothing changes with graph 1kHz = -240. I made 5th HP filter and the same problem. Also when i used to ideal opamp for 5th filter and function which give me good graph for real opamp, PSpice give me wrong graph - same like P(V(Vout)).

4th -
real opamp - P(V(Vout) - wrong graph
real opamp - P(V(Vsr)/V(Vin))+P(V(Vout)/V(Vsr)) - good graph
ideal opamp - P(V(Vout) - wrong graph
ideal opamp - P(V(Vsr)/V(Vin))+P(V(Vout)/V(Vsr)) - good graph (independently from GBP)
5th -
real opamp - P(V(Vout) - wrong graph
real opamp - P(V(Vind)/V(Vin))+P(V(Vind2)/V(Vind))+P(V(Vout)/V(Vind2)) - good graph
ideal opamp - P(V(Vout) - wrong graph
ideal opamp - P(V(Vind)/V(Vin))+P(V(Vind2)/V(Vind))+P(V(Vout)/V(Vind2)) - wrong graph (independently from GBP)

 

[MikeMl]

Are you plotting Gain = V(out)/V(in)?

I re-ran the LTSpice sim as a frequency sweep from 1Hz to 1e6Hz. I asked LTSpice to plot the Phase of V(out) in real degrees, since by default, V(in) is 1V@0degrees.

Note that the Ph() plot function in LTSpice takes the Phase modulo(360degrees). However, it would also be correct to say that the phase is 120.8 + n*360 degrees, where n is an signed integer.

 

[jeremieblanc]

MikeMl once more thanks for your response and time (you are good man) .I mostly think that when i use some function it will show a good result (of course if schematic is made good). I learned at the University that we should always check us graph and try as far as possible interpreted it but here i can only describe my problem and tell how we can - in this case- make good phase shift graph P(V(Vout)/V(Vin))+360 for PSpice. Unfortunately at this stage of master thesis i can't show that LTspice handles with this simulation much better because it would change the design assumptions (but I think it would be a great complement to its). Thank you for advice and learning from you.