Problem in filter design

[guillermorf]

Hi, first of all i must say i´m not English and excuse me if this thread doesnt fit in this theme.

My problem is that i´m designing a passband-filter, with 2 stages for a highpass filter and with 4 stages for the low-pass filter. I´m using operational amplifiers THS3001 which are supposed to work in this design.

The thing is that in every very stage there is an oscillating signal in the inversive-pin of the amplifier. I´ve tried to isolate every stage and the signal is still there, and the last thing I tried was to not introduce a sine wave in the opAmp but there is still a signal on the inversive-pin.

I hope anyone can help me to solve my problem.

Thanks in advance.

Here i leave you the 2 stages of the highpass filter. The thing is that i have the same problem when all the circuit is connected and when i have just the first stage and without introducing any signal through Vin.
**broken link removed**

 

[ericgibbs]

hi,
Download this free filter designer.
http://focus.ti.com/docs/toolsw/folders/print/filterpro.html

 

[guillermorf]

Hi ericgibbs

The filter design im showing you has been designed by that program

 

[Mikebits]

Are you doing this on a breadboard? Are your supply pins properly bypassed?

 

[Mikebits]

This from the data sheet.

DRIVING A CAPACITIVE LOAD
Driving capacitive loads with high-performance amplifiers is not a problem as long as certain precautions are
taken. The first is to realize that the THS3001 has been internally compensated to maximize its bandwidth and
slew-rate performance. When the amplifier is compensated in this manner, capacitive loading directly on the
output will decrease the device's phase margin leading to high-frequency ringing or oscillations. Therefore, for
capacitive loads of greater than 10 pF, it is recommended that a resistor be placed in series with the output of
the amplifier, as shown in Figure 52. A minimum value of 20Ω should work well for most applications. For
example, in 75-Ω transmission systems, setting the series resistor value to 75 Ω both isolates any capacitance
loading and provides the proper line impedance matching at the source end.

https://www.electro-tech-online.com/custompdfs/2011/07/ths3001.pdf

I have a feeling this op amp is not right for your application.

 

[guillermorf]

I first did this on a breadboard and then mounted it using the wrapping technique.

I have two condensator for each supply voltage, -5V and +5V, and then one pair of condensators for each opAmp, one for +5V and one for -5V.

 

[guillermorf]

This from the data sheet.

DRIVING A CAPACITIVE LOAD
Driving capacitive loads with high-performance amplifiers is not a problem as long as certain precautions are
taken. The first is to realize that the THS3001 has been internally compensated to maximize its bandwidth and
slew-rate performance. When the amplifier is compensated in this manner, capacitive loading directly on the
output will decrease the device's phase margin leading to high-frequency ringing or oscillations. Therefore, for
capacitive loads of greater than 10 pF, it is recommended that a resistor be placed in series with the output of
the amplifier, as shown in Figure 52. A minimum value of 20Ω should work well for most applications. For
example, in 75-Ω transmission systems, setting the series resistor value to 75 Ω both isolates any capacitance
loading and provides the proper line impedance matching at the source end.

https://www.electro-tech-online.com/custompdfs/2011/07/ths3001-1.pdf

I have a feeling this op amp is not right for your application.

Why you say it isnt the right op amp? I read that text, but i dont understand it pretty good as im not an english speaker.

Im not putting any capacitive loading at Vo as im just now working with one single stage. Im working with a circuit that has from Vin until before C1=5,62nF

 

[Mikebits]

The op amp is a high speed wide bandwidth amp. You will have to use a board layout or you will have problems. See user guide. https://www.electro-tech-online.com/custompdfs/2011/07/slou021a.pdf

Do you need that much bandwidth? Your filter has a 20KHz cutoff, I would pick a different part.

 

[guillermorf]

I´m using this op amp because of the gain bandwidth product required, in the highpass filter its a small GBWP, but in the lowpass part a 600MHZ GBWP is needed.

With all the stages its a bandpass filter going from 20KHZ to 2.5MHz, but i just pasted here the highpass filter so you can see the circuit.

So if I dont use a board layout it will not give problems but if i use a breadboard or a wrapping technique it will?? :S

 

[guillermorf]

I think the user guide you posted here its about the evaluation module

 

[MrAl]

Hi,

I too have to wonder if you need such a high frequency part for this.

Also, when you work with those op amps you should probably limit any of the resistor values to 1k or less.
When you test it with no input, the input should be grounded.

 

[guillermorf]

I need this because of the GBWP, which needs to be higher than 625MHz in the last stage (according to FiltroPro specifications).

Its better if i use resistors less than 1k? I didnt know that, I´ll try it, but i thought FilterPro was taking this in consideration when calculating the components values

 

[MrAl]

Hello again,

Does "FilterPro" know you are using a current feedback amplifier?

You may also have to change the design to a SK that also uses Rg as shown on the data sheet. I guess you'll want to see if lower value resistors work first.

 

[guillermorf]

No, i dont think filterPro knows that... because the design it makes has a feedback with a resistor meanwhile this current feedback amplifier says it needs a 1K resistor in feedback.

I dont have such resistor values, but trying with the Rg as you say doesnt work either... its driving me crazy!!

 

[MrAl]

Hello again,

Since you are having this much trouble then i suggest that you try to design one of the reference designs on the data sheet and get that working well. For example, you can try a standard amplifier first and get that working with no problems, then move to the low pass filter they show, etc.

If you have the same problem with a standard amp or low pass filter then the only thing left is the layout and/or the way the chip is wired up. These things really need a PC board that is etched most likely rather than a point to point wire hookup.

So see if you can get a regular 5x amplifier working first, then a low pass filter as shown on the data sheet, then take it from there. Note that a design with Rg would be different than without it.

Oh yeah one more 'little' thing...
When you say you have oscillation on the inverting input, how do you know this? You shouldnt be looking at that point with a scope, only the output. The scope probe hookup could change the circuit too much when connected to the inverting input.

 

[guillermorf]

Ok, ill try the datasheet designs and lets see if they work...but they obviously should

 

[Mikebits]

Layout is most likely a big issue here. You may get by using double clad vector board with meticulous parts placement.