# Filter design

Status
Not open for further replies.

#### Electronman

##### New Member
Hello everybody.

I am a newbie in filter design arena, I really need help to create a high pass filter.
I prefer it to be an active filter.

Ok, the fundamental frequency is 40kHz and I need a high pass filter with almost 10kHz of bandwidth.
Is there any expert in the field to help me to create that please?
Please consider that I have a single ended power source so the active filter has to work with that single source at 12 to 15V.

I really need it for a project so Any help would be much much appreciated.

#### Electronman

##### New Member
As I told you I am using a single power supply so I do not know If I have to change the values of the resistors or capacitors too?
I have this program: "FILTER SOLUTIONS V9.0.34" but do not know how to use it to create the filter which I need, specially with single ended power source?

#### dknguyen

##### Well-Known Member
It does not matter. Op-amp bipolar or unipolar power source only affects the voltage range that the circuit can accept and output. It does not change frequency response.

Last edited:

#### flat5

##### Member
Post a clear picture (gif or pdf) of the circuit you have and people here will help you to make the necessary changes.

Last edited:

#### Electronman

##### New Member
I told you before, I am a newbie in the filter designing,
But here is a start point to go:
I designed it by the above software you mentioned:
The cut off freq is at 39kHz.
I am not sure if I have done it right though?
please keep in the mind that I need it for 10kHz of the bandwidth. Is is used to create SSB with carrier out of my before modulator design.

The second one is a 6 stage one too, it is made by Filter solution software too.

I even am not sure if I have chosen the right active filter (i.e Butterworth)

#### Attachments

• 286.2 KB Views: 483
• 63.2 KB Views: 309

#### crutschow

##### Well-Known Member
Not sure I understand the 10kHz bandwidth, The high pass filter will have a bandwidth of 39kHz up to the rolloff frequency of the op amps. If you need a 10kHz bandwidth, then do you need a bandpass filter? Or do you just need a minimum 10kHz bandwidth above 40kHz.

A butterworth filter has a flat passband with moderate rolloff. Other filter types can have faster rolloff but have ripple in the passband. And certain types have well controlled phase-shift with frequency or minimum overshoot with a pulse input. So the type you chose depends upon your requirements.

Last edited:

#### Electronman

##### New Member
Not sure I understand the 10kHz bandwidth, The high pass filter will have a bandwidth of 39kHz up to the rolloff frequency of the op amps. If you need a 10kHz bandwidth, then do you need a bandpass filter? Or do you just need a minimum 10kHz bandwidth above 40kHz.

A butterworth filter has a flat passband with moderate rolloff. Other filter types can have faster rolloff but have ripple in the passband. And certain types have well controlled phase-shift with frequency or minimum overshoot with a pulse input. So the type you chose depends upon your requirements.
Again I am a newbie in the field...

the modulating signal comes from my mic and or from my mp3 player, I want to play a tone via mp3 player so I noticed I need more than 3kHz of bandwidth (the bandwidth of human speech).

I want to convert the current DSB with carrier modulator into a SSB with carrier, So I think I can do so via a high or low pass filter. I myself prefer high pass filter for that.

Besides, I have more that 10 of 40kHz crystals If it does help...

Thanks

#### Electronman

##### New Member
I am sure that my modulating signal does not go beyond 10kHz anyway. If for some reason I could not build a good filter then I will try a filter for just 3kHz of bandwidth (for human speech).

Is here any expert in the filter field please?
I read a small paper about active filter designing but it could not help me to design my own filter...

#### crutschow

##### Well-Known Member
If I understand correctly, you are DSB modulating a carrier at 40KHz and want to roll-off everything below 40KHz to suppress the lower sideband.

The slope of the filter roll-off below 40kHz is determined by the filter order, which determines how much of the lower sideband is suppressed. A 6-pole filter, as you showed, rolls off at the rate of 36db per octave of frequency.

If that roll-off suppresses the lower side-band sufficient for your purposes, then the filter you generated in the Filter-Pro attachment should work. Do you have any questions about that filter?

The modulation bandwidth of the circuit is limited only by the op amp frequency response. If you want to limit the bandwidth for any reason, then you could add a low-pass filter. For example for a bandpass of 10KHz you would add a 50KHz low pass filter. It could probably have fewer poles then the HP filter since that filter is not so critical.

To use the circuit with a single supply you will need to connect the points shown going to ground to a pseudo-ground at 1/2 the supply voltage (as shown in the previous reference I mentioned) and use a coupling capacitor at the filter output.

#### Electronman

##### New Member
Yea I want to remove one side band (the lower one). The problem is that I do not know if I have done every thing well in the above circuit (the one that I have drwon using your suggested software).

I do not know if I need to limit the bandwidth for a high pass filter that is used to remove just one side band?? Any way I made the above circuit (the one using your suggestion software) but It did not work at all? I had no output at the output pin of the op-amp??

Can you guys help me with a good circuit to do the above task correctly please?

Thanks

#### audioguru

##### Well-Known Member
What was the DC voltage at the output of your filter?
Of course it will work if it has the correct parts and is powered and biased correctly.

Post your filter's schematic showing the opamp's part number, supply voltages and biasing for the opamps.

#### Electronman

##### New Member
What was the DC voltage at the output of your filter?
Of course it will work if it has the correct parts and is powered and biased correctly.

Post your filter's schematic showing the opamp's part number, supply voltages and biasing for the opamps.
The main question I think is if a 6 pole (is that equal to 6 order?) does work for the audio I want to use (i.e 10kHz)? I am pointing to what I have heard: the more order of the filter the better result .

The second question is Can I maintain the carrier at the output of the filter(I need for it)?

The third question Is how to maintain the gain ( I do not like to lose the gain out of my current DSB with carrier modulator)? I even would like the gain more than 1.

And finally, I fear the filter changes the waveform somehow?

I am not familier with active filters so do not know how to give you a better schematic? Anyway I am using TL074, where ground needed I use two 1k resistors (these resistors have 5* of tolerance, Is that important for my opamp designs??).
Anyway if noone help me with a better schematic the I will try to make one with those softwares again, but please direct me with the above questions and let me know which kind of filters (Bessel, or Butterworth or..are the best choice? What about the type (i.e. MFB Single-Ended, Sallen -Key)?

Thanks

#### Electronman

##### New Member
Are you guys sure that the parameters are correct for a high pass filter at 39kHz (I am reffering to the above circuit designed by the suggested software)? Again please consider that I need the carrier itself?

#### audioguru

##### Well-Known Member
Two filter circuits were posted.
Which one did you make that did not work?
Both must be driven from a very low impedance like the output of an opamp.
The first circuit with only 3 opamps must be biased with a lot more than only 2 resistors.
That is why your schematic was requested.

No power supply voltage?

The 39khz filter reduces the level of the 40kHz carrier by about 0.85 times.

The TL071 barely makes a filter as high as 39kHz. An opamp is needed that performs well at 10 times 39kHz (390kHz).

#### crutschow

##### Well-Known Member
A DSB modulator normally suppresses the carrier.

As Uncle $crooge says you need to post the schematic exactly as you built it, otherwise we can not determine why it doesn't work. Saying you "had no output" doesn't tell us much. You can draw the schematic with a free computer program such as Tiny Cad. I'm quite sure the circuit, as generated by the program, will provide the given response if the op amps are properly connected and powered. For example you said you used two resistors where ground was needed. Where do these two resistors go? If they go between power and ground they need to be decoupled with a large capacitor from the junction to ground. For the differences between the various filter types and configurations do some searches for "active filters", etc. on Google and Wikipedia. Then ask us what you don't understand. Of course higher order filters give a faster rolloff but we can not determine what your requirements are for the SSB signal. That's for you to determine based upon what exactly you are trying to do (of which you have given only a vague description). For more gain just add an additional op amp circuit with gain. We are here to help you to do the design, not do the total design for you. #### Electronman ##### New Member Two filter circuits were posted. Which one did you make that did not work? Both must be driven from a very low impedance like the output of an opamp. The first circuit with only 3 opamps must be biased with a lot more than only 2 resistors. That is why your schematic was requested. No power supply voltage? The 39khz filter reduces the level of the 40kHz carrier by about 0.85 times. The TL071 barely makes a filter as high as 39kHz. An opamp is needed that performs well at 10 times 39kHz (390kHz). Ok I will draw a new filter in Pspice and put it here . The modulator is based upon 4066 chip so the input of the filter comes from the output of 4066 (can you remember the modulator circuit at the other thread?) Please let me how to biase that op-amps because I want to draw it with pspice and want to use your guide to biase it correctly? The power supply is 14V single ended. I have several LF411 and LF356 Can they be used instead of TL074? I told that I need the carrier so What is the solution??? 0.85 time is too high I think. #### Electronman ##### New Member A DSB modulator normally suppresses the carrier. As Uncle$crooge says you need to post the schematic exactly as you built it, otherwise we can not determine why it doesn't work. Saying you "had no output" doesn't tell us much. You can draw the schematic with a free computer program such as Tiny Cad.

I'm quite sure the circuit, as generated by the program, will provide the given response if the op amps are properly connected and powered.

For example you said you used two resistors where ground was needed. Where do these two resistors go? If they go between power and ground they need to be decoupled with a large capacitor from the junction to ground.

For the differences between the various filter types and configurations do some searches for "active filters", etc. on Google and Wikipedia. Then ask us what you don't understand.

Of course higher order filters give a faster rolloff but we can not determine what your requirements are for the SSB signal. That's for you to determine based upon what exactly you are trying to do (of which you have given only a vague description).

For more gain just add an additional op amp circuit with gain.

We are here to help you to do the design, not do the total design for you.
Ok here is the modulator itself, working at 40kHz and its related FFT.
please look at the FFT, It contains the carrier too.

I will draw a 6 or 7 pole high pass filter by your suggested software and redraw it in Pspice (I will connect it to the current circuit You are looking).
I will put the filter circuit here tomorrow.

I thought maybe you have seen a such filter circuit or can handle the softwares which are used to generate active filters so asked you if thats true please put it here (not wanted you to bother yourself by designing one)

#### Attachments

• 288.4 KB Views: 163
• 108.7 KB Views: 97
• 469.8 KB Views: 107

#### audioguru

##### Well-Known Member
You posted two filter schematics.
Select the one you used and use Microsoft Paint program or something to copy it and post it here showing the resistors used to bias it. Show the input voltage and frequency that you tested it with.

How did you measure the output of the filter? On an oscilloscope?
What was the output DC voltage of the filter?

Your CD4066 "modulator" does not have an output coupling capacitor and has an output impedance that is not low enough for both filter circuits.

#### Electronman

##### New Member
Ok here's my new filter design. It is designed by 'FilterPro' (the software suggested in this thread)

this both pictures are the same, I just put the first picture to show the biases in real circuit.
Are you thinking that this filter does work at 39kHz and removes or attenuates the loser sideband? (I need the carrier and the upper sideband at the output of the filter). Please keep in minde that the input bandwidth is almost 10kHz.
Any suggestion would be appreciated.

P.s did you care of the single power supply and the 1k biasing resistor?? No need to capacitors for them?

#### Attachments

• 91.3 KB Views: 112
• 116.7 KB Views: 118
Last edited:
Status
Not open for further replies.