high gain LPF

Thread Starter

raa

Joined Apr 5, 2011
18
hi all,

am looking for a LPF with very very high gain and its cut off frequency at 30KHZ i can't find any one till now, what i found is this one

http://www.simplecircuitdiagram.com/2010/02/09/a-20-khz-third-order-low-pass-filter/

but its only 20KHZ and am not sure if its gain is high enough , also i found this wonderful circuit that allow me to choose the desired gain and frequency

http://www.maxim-ic.com/app-notes/index.mvp/id/700

but the problem is the max4174 is no longer available, am still looking for it, any help would be greatly appreciated
 

Papabravo

Joined Feb 24, 2006
21,159
I would advise a different approach. There is no real reason why you have to do everything with one circuit. Arguably you will be better off if you attack filtering and gain separately. There are countless posts from people who want a power gain of 30 dB in a single stage. While such a thing is possible it may ultimately prove superior to to have three stages at 10 dB each. I think this may be the case here.
 

Thread Starter

raa

Joined Apr 5, 2011
18
thanx for the good informations, here it is the LPF circuit am going to use
its the low pass sallen-key Architecture

http://img823.imageshack.us/i/unti22tled.png/

can some one tell me plz the equation i need it in the design and the gain equation and some guidelines to follow it in my design, i tryed google but i found different equations for it and when i try it on a known circuit, it doesn't give me the right values, i trust all about circuits more :D
one last thing guys i found plenty of ciruit without R4 and R3 , what is the diffrent between the circuit with R4 and R3 and the one without it?? and thank you guys for helping i truly appreciated.
 

Audioguru

Joined Dec 20, 2007
11,248
How steep do you want the slope of the filter? The 2nd-order filter you found has a gradual slope of only 12dB/octave. Then 60kHz will have a level that is 1/4 the level at low frequencies. 120kHz will have 1/8th.

The Sallen and Key circuit you found can have equal value capacitors and equal value resistors. Then it has a Butterworth response when its gain is 1.6.
 

Thread Starter

raa

Joined Apr 5, 2011
18
sorry man, but i really did't understand any thing...
all i want to know
a)the design equation and the gain equation.
b)the different between the circuit with R3 & R4 ,and the circuit without it.
 

Audioguru

Joined Dec 20, 2007
11,248
R3 and R4 add gain which causes a small peak in the response to counteract the drop in the response from equal value resistors and capacitors. If the gain is more than 2 then there is a peak in the overall response. I think it oscillates when the gain is higher than 3. Use a separate opamp for more gain.

I made a +10dB peaked response at 3.5kHz with a lowpass filter for tele-conference systems to make muffled voices sound crisp and clear. I made peaked highpass filters with a peak of +10db at 30Hz as bass boost for my stereo.
 

almoatazbellah

Joined Nov 24, 2010
22
Dear Raa, make in your design R1=R2=R and C1=C2=C.
the reactance of a capacitor decreases as the frequency of the signal through it increases. Thus, in your circuit, C2 tends to shunt high-frequency signals towards the ground, causing them to be attenuated at the output. on the other hand, the lower the frequency of the input signal, the higher are the reactances of C1 and C2, and the closer the circuit resembles a non-inverting amplifier. the gain G of the amplifier is approximately given by G = R3/R4.

The cut-off frequency fc of the low-pass filter is approximately given by fc = 1/(2∏RC). At fc, the output is 0.707 times the maximum output level. The output level increases as the input frequency falls farther below fc, which is why this is a low-pass filter.
 

Ron H

Joined Apr 14, 2005
7,063
Dear Raa, make in your design R1=R2=R and C1=C2=C.
the reactance of a capacitor decreases as the frequency of the signal through it increases. Thus, in your circuit, C2 tends to shunt high-frequency signals towards the ground, causing them to be attenuated at the output. on the other hand, the lower the frequency of the input signal, the higher are the reactances of C1 and C2, and the closer the circuit resembles a non-inverting amplifier. the gain G of the amplifier is approximately given by G = R3/R4.

The cut-off frequency fc of the low-pass filter is approximately given by fc = 1/(2∏RC). At fc, the output is 0.707 times the maximum output level. The output level increases as the input frequency falls farther below fc, which is why this is a low-pass filter.
No, the in-band gain is (1+R4/R3).
 

Audioguru

Joined Dec 20, 2007
11,248
You did not hear me.
With equal resistors and equal capacitors the gain must be only 1.6 for a Butterworth flat then sharp dropoff response. If the gain is higher than 1.6 then there will be a peak in the response near the cutoff frequency and if the gain is more than 3 then it will oscillate. Use a separate opamp for gain.

The opamp you selected has "less than ordinary" high frequency response to about 15kHz and will not make a good 30kHz filter.
 

Thread Starter

raa

Joined Apr 5, 2011
18
@Audioguru thanx for replying, what is opamp i need to use for 1.6 gain and 30KHZ cut-off frequency and the values for R3 and R4
 

Audioguru

Joined Dec 20, 2007
11,248
Use an OPA134 opamp that goes to 8MHz and has extremely low distortion.

Why do you want to reduce frequencies above 30kHz?
Is the gradual slope of a 2nd-order filter good enough?

The resistors for a gain of 1.59 are a 3.3k feedback resistor and a 5.6k resistor to ground.
 

Thread Starter

raa

Joined Apr 5, 2011
18
now my final design is R1=R2=16KΩ , C1=C2=330pf, R4=3.3K and R3=5.6KΩ (Fc=30Khz)

yes the gradual slope of a 2nd-order filter is good enough for me :D

thank you really for your help.
 

Thread Starter

raa

Joined Apr 5, 2011
18
@Audioguru , can you tell me plz how to connect OPA134 in the LPF circuit because its a little different from max9613 it has 8 pins and according to the datasheet pin 1 and 8 is Offset Trim i don't know what those pins for also there is a +V and -V while Max9613 was having only one Vcc , also in the datasheet it say that it is wide supply range: ±2.5V to ±18V is that mean i have to choose any supply voltage , and do i have to use a bypass capacitor. i know am asking too much questions but am really confuse and am not that good in electronics.. i need some help in the connection :D
 

Audioguru

Joined Dec 20, 2007
11,248
Have you looked at the extremely tiny size of the MAX9613? It is 2mm long and 1.25mm wide. You can't even see its 5 tiny legs that are only 0.65mm apart. How can anybody solder to it?
The Maxim schematic shows a single supply and no gain. Then its input must be biased at half the supply voltage somehow. One capacitor must have double the value of the other for a Butterworth response and no gain.

Any opamp including the OPA134 can use a single supply and will also need no gain to have its input biased at half the supply voltage somehow.
If you do not need to adjust the input offset voltage then do not connect anything to the offset adjust pins.

What feeds half the supply voltage (at a low impedance) to the filter circuit?

Maybe it will be much easier to use a dual polarity supply for the filter opamp and bias the input at 0V (ground). Then the opamp can have a gain of 1.6 and have equal value capacitors.

Your resistor and capacitor values result in a cutoff frequency of about 21kHz. I couldn't find values for exactly 30kHz.
 

Thread Starter

raa

Joined Apr 5, 2011
18
thanx for replying

i will connect the opamp with :

1-dual polarity supply

2-the input offset voltage not connected

is any thing missing in the connection ??

i have some questions :

1-what is the supply voltage value

2-what do mean by bias the input at 0V (ground)

3-what is the equation did you used to calculate the corner frequency
 

Audioguru

Joined Dec 20, 2007
11,248
1-what is the supply voltage value
The datasheet for the OPA134 shows that its minimum supply is plus and minus 2.5V and its maximum is plus and minus 18V. Are you blind??

2-what do mean by bias the input at 0V (ground)
The input must have a DC voltage of 0V at a low impedance like from the output of another opamp that has its input biased at 0V.

3-what is the equation did you used to calculate the corner frequency
I am sorry I made a mistake in my calculation. Your values are correct for a 30kHz cutoff frequency.
 
Top