5KHz simple low pass audio filter needed

hello I design a low power (500mW) audio amplifier using small transistors.
The amplifier is going to be used in a SSB modulator so I need to cut frequencies that exceed 3-4KHz
Is there any simple filter schematic I can use (RC maybe?) with low passband loss in order to cut frequencies higher than 5KHz?
Simple can mean a lot of different things. There are several ways to make a low pass filter:
1. first order filter: a series capacitor. This is by far the easiest to do, but attenuation is poor.
2. second order filter:
A. using op amps and sallen key or multiple feedback design
B. time continuous filter ICs like the UAF42 from Burr Brown
C. switched capacitor filters like Linear Devices LTC1062 and related products. Maxim is another company who makes these I think.

There are many different switched capacitor filter chips available. I mention the LTC1062 because it is DC accurate, needs only two or three external components, and implements a 5th order Butterworth filter function. But there are many, many different switched cap filters. One common problem is that they are not as low noise as analog circuits. Most are low pass, but a few are general purpose and/or have HP functions as well.

Which implementation you choose depends on your budget, attenuation needs, available DC power supply voltages (most need +5V DC only), etc. Choose carefully.

Thank you all for your replies.
I usually prefer simpler circuits and avoid using opamps where possible, although I would not expect the performance of an opamp using single transistors.
I have found this http://freecircuitdiagram.com/wp-co...e-Baxandall-Tone-Control-2-Band-Equalizer.gif which is a passive graphic equilizer which could be used to let only the voice pass through. Of course this is passive, meaning lossy and I have used an input and output transistor buffer to overcome the loss.
On the other hand I have found this http://www.rlocman.ru/shem/i/users.otenet.gr/~athsam-Circuits-EQ_3.gif which is an active equalizer but it uses more transistors, I wonder if I could omit a few bands without problem. The problem is that it uses symetric psu +-24v


2006-09-08 2:04 pm
The 500mW is the speaker output power. Part of the amplifier chain will be used for the phasing modulation combined with a passive audio polyphase network.

The thing is that this polyphase network keeps relatively accurate phases at about 300Hz-3KHz. Out of this range the phases loose their accuracy and the unwanted sideband starts to appear. Also, the greatest audio bandwidth is, the greatest the RF bandwidth is, because the phasing method of ssb generation does not have any RF bandpass filtering mechanism. So I need to limit the audio or at least cut off the higher audio frequencies. Unless my considerations are not right...

A simple audio compressor will allow for almost the same RF level for all audio variations (keeping the SSB signal power at maximum), but for bandwidth limiting I need an audio filter.

I am just wondering if a simple filter can be made using a few transistors. The buffered passive filter works ok but I just wonder if a better simple solution can be done.
I am still not clear whether you are building a receiver or a transmitter. In either case you will need good audio filtering for polyphase SSB. Cascaded Sallen-Key is how I would do it, using 2 or 3 stages (i.e. 4 or 6 poles). If you find out about filter design you can learn how to combine stages to get the response you want: Butterworth or Chebyshev. This is a non-trivial exercise, so I suggest you either copy an existing design or start reading about active filters.
So why do you need a 500mW audio amp to make a modulator? Is your polyphase network particularly low impedance? What are you using for mixers? What power output are you aiming at?

No, the amplifier is used as a speaker amplifier on the RX state and part of the amplifier (early stages, not 500mW) is used on TX state to provide the nessesary audio gain to the phasing network.
I use 4 bipolar transistors for the balanced modulator driven by a phase splitted RF signal (simple construction) and a combiner and small RF amplifier at the ouput. The power output is about 3vpp.

There is always the good old MC1496 but I prefer simple homebrown solution for this modulator
OK, it makes sense now. To get good sideband suppression on output and good selectivity on input you need to ensure good phase and amplitude matching.

I started building a polyphase network many years ago, then other things came along and I never finished it. One day I will have another go.

I know, it is simple, but the components need to be accurate, NP0 capacitors and low ppm resistors on the phasing network.