Simulating Multiple Feedback Low Pass filter in LTSPICE!

[sean87]

Hi all, again!

I am trying to simulate the follwoing circuit in LTSpice IV. What I am looking for is overall performance in frequency domain.



I downloaded and readed a lot of LTSpice guides, but none of them were speaking about how to do for this kind of circuit. If you know how-to, I would really appriciate yuor help! Maybe it is stupid but I do not know what voltage source (AC, DC, sine or ...) I should use and I do not know which simulation type will be what I need.

I have attached the schematic design, if you want to look.

[jahonen]

Unless you want to use your own model for the opamp, I suggest you use suitable LT opamp in simulation. Or if you want just theoretical result, then substitute E (voltage controlled voltage source) component for the opamp and give it a very big gain (something like 10^12 or similar).

Another thing, delete the f's in capacitor values. LTSpice might complain about them.

Regards,
Janne

[Bored@Work]

Treat it like you want to measure a real circuit. You decide what you want to measure. From there it follows what input signal(s) you need to apply, what type of simulation you want to run, and for how long or over what range (not all applicable).

[sean87]

Thanks for caring guys

I want to see how is the frequency responce. Actually I am trying to design a filter and I amfollowing this guide. I managed to do the calculations and it seems to be like what I need it to be.

The problem now is simulating the circuit in LTSPICE! Luckly I have found this site that plots the frequency responce, and it is like this:



But when I am trying to simulate this circuit by myself using LTSPICE, I get something like this:


And here is the schematic and setting for the circuit I designed in LTSPICE as well as simulation commands:


The opamp, I right clicked and changed "Aol" to 100M, that is 100,000,000
There was another option in the opamp called GBW which is set by default to 10Meg

For voltage source I just put 1 in AC Amplitude

Can you please tell me what I am doing wrong? how can I get a similar result to the plot generated by that website?

Thanks!

[jahonen]

Try changing your simulation command to ".ac dec 500 1 1meg" and remove "F"'s in capacitance values. Further, you are missing a connection dot in crossing of R1-R2-R3-C2.

Regards,
Janne

[qno]

Hi all,

I think a junction is missing between R3 and R1 in the simulation schematic.

[sean87]

yeah guys, I was so stupid to miss that node anyway after putting that node in the junction I got the following plot, but it is not still the same as the generated file from that website. (I also used the commands janne gave). What can be the problem?

[sean87]

Well after some investigation I found out the last plot I posted here (above post) is correct.

Now can somone please tell what kind of information this plot can give to us? whats up with all these curves?

Can we see the gain and cutoff frequency here?

[ejeffrey]

It looks like the gain is 6 dB (2x gain) and the 3 dB point is around 200 Hz.  The amplifier is inverting, so the phase at DC is 180 degrees.  It is a second order filter, so the slope in the stop band is 40 dB/decade, and the phase shift in the stop-band is 0 degrees: 180 from the inverting nature, and an additional 90 degrees from each order.  The bumps in the phase response don't mean much.  The transition band (curved region of the amplitude plot) is fairly wide with gentle phase response, so this is a low Q filter.  It will be insensitive to component variation and won't screw up your time domain waveform too much, but it won't attenuate signals unless they are well above the pass band.

[Bored@Work]

Now can somone please tell what kind of information this plot can give to us? whats up with all these curves?

Guess why I wrote you should think about what to measure first? Now you have some plot and don't know what to do with it.

[Conrad Hoffman]

Here's one that's fun. To get a steeper response you can build a 4-pole filter with just a single opamp. This one is the Butterworth version and you can scale the resistors and capacitors for whatever frequency you like. Enter exactly as shown, at least the first time around.

[sean87]

It looks like the gain is 6 dB (2x gain) and the 3 dB point is around 200 Hz.  The amplifier is inverting, so the phase at DC is 180 degrees.  It is a second order filter, so the slope in the stop band is 40 dB/decade, and the phase shift in the stop-band is 0 degrees: 180 from the inverting nature, and an additional 90 degrees from each order.  The bumps in the phase response don't mean much.  The transition band (curved region of the amplitude plot) is fairly wide with gentle phase response, so this is a low Q filter.  It will be insensitive to component variation and won't screw up your time domain waveform too much, but it won't attenuate signals unless they are well above the pass band.

Thanks, but would you mind to tell how you got this info? would be really nice if you mark them in the picture and upload it again here if you feel for it

[Conrad Hoffman]

One of the fundamentals needed to read these graphs is an understanding of the dB system. Do you understand dB- if not, we need to start with that.

[sean87]

One of the fundamentals needed to read these graphs is an understanding of the dB system. Do you understand dB- if not, we need to start with that.

Thanks Conrad, I know how dB is. Just knowing the plot I got would be nice!

[Conrad Hoffman]

If you look over on the left side scale, that's the ratio of the signal source amplitude to where ever you clicked to make a measurement, hopefully the output. Note that the input and output labels serve no function here. They're used to connect to other circuit blocks, or just to help identify the nodes.

Filters are specified by the -3dB point, so you follow the gain curve from left to right until it drops by 3dB. At that point, go down to the frequency scale at the bottom and read off the frequency. Now you know the -3dB point.

The filter order determines many things, like the attenuation in dB/octave or dB/decade that ejeffrey gave. The phase plot is read in degrees on the right hand side of the graph. Though digital filters can do some strange things, analog filters are very constrained. Just knowing the order and construction of the filter tells you what the phase plot will look like, without even simulating it.

[codeboy2k]

I'm a little late in this thread, but from reading, I think Conrad has just completed someone's homework for them,
including telling them how to read the basic Bode plot.

The OP doesn't sound like he even knows the basics of filters and how to sweep the input to get a frequency response
output.. he doesn't even know what he's got, once he got it.



Thumbs down to anyone using the Internet for their homework.