Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Filter for voice comms??

Status
Not open for further replies.

Electricman2K5

New Member
Does anyone know how to make an simple RC filter for voice comms i.e. stop anything above 4Khz?? How do /i calculate the component values and the cut off frequency

What about RLC filter is thias better?? How doyou work out the cut off here.??

Thanks
 
An LC filter has only a 12dB/octave slope, and an inductor takes up pcb space.
How sharp a cutoff do you need? What's a voice comm and why do you want it to sound worse than an AM radio or telephone? Do you test noisy jet engines at work?
A switched-cap low-pass-filter IC like the ones from Maxim have slopes up to 48dB/octave which is the equivalent to having 4 LC stages or 8 buffered RC stages. You can easily change the cutoff frequency.
**broken link removed**
 

Attachments

  • max291.gif
    max291.gif
    31.5 KB · Views: 2,433
Filter

I mean voice comms as in voice communications a simple intercom I only want to pass voice frequencies which will not be greater than 4Khz right??? Why will it sound like a noisy am radio??? - I (and this may shock you) havent made a filter before.
 
Hi Electricman,
Unlike the deaf old geezers who invented lousy-sounding telephones and AM radio, most people can hear up to 20KHz. Most of the intelligence is up there: did you say "s" or "f" and "ick" or "six" and so on. Those deaf old guys didn't care, lousy-sounding 300-3KHz was much better than Morse Code, and they didn't know how to go lower or higher anyway.
They were very worried about crosstalk. Horrors if you could hear your neighbour! Capacitance between their wires coupled high frequencies, so they got rid of the high frequencies. They didn't realise that their balanced wiring cancelled crosstalk. They were also very worried about high distortion, so cut the highs more to reduce it from their terrible-sounding transducers.

I wasn't talking about noise on AM radios, I was talking about the high frequency rolloff they have above only about 2KHz. AM radios must attenuate 10KHz extremely to avoid an annoying 10KHz interference-between-adjacent-stations whistle. When they use a simple cheap low pass filter, it begins its cutoff at about 2KHz.

Modern telephones with electret mics sound much better than the old carbon ones that cutoff above 3KHz. But the telephone network is also bad. When I worked with telephone systems I asked Bell, "Why is the response at only 2.5KHz and above, down 15dB when I call next door?" They said I was far from their equipment and the spec allows -7.5dB one-way. My location was at their limit and I had terrible sound! So I sold an equalizer circuit to many customers with my boardroom telco conferencing system that boosted 3KHz +10dB to make it sound nearly normal, and as soon as they heard it they said, "Wow, we must have that!" Don't tell Bell, it was illegal, but no complaints.

Even television announcers sound bad with their fancy little lavalier mics hidden under their turtle-neck sweaters. Since their mic is close to their chest and throat, their voice booms loudly but no high frequencies come through. Lately they have been using headset mics (due to my complaints) and they sound much better.

I worked with a sophisticated intercom system for many years. Its sound quality (and price) blew away the competition for large systems in life insurance companies (room to room), banks' head offices (hundreds of stock-traders' communication) and airports (announcements over the PA) where quality was important and price didn't matter. Its response was flat up to about 12KHz, it had very low distortion, you could also play background music through it and it sounded great! I added bass-boost to the background music to make it sound even better! I had to tweak it a little to make it pass 15KHz and meet the airport's "broadcast quality" spec. I use its stations' 3" speaker and my bass-boost circuit in my clock radio and it sounds like a hi-fi with a sub-woofer! It weighs a lot more than an ordinary little clock radio but looks the same.

Have you ever heard good wide-band speech through just a tweeter? It sounds odd like whispering, but is crystal clear. It produces only those frequencies you want to get rid of.

So if you want to sound like you are talking through a thick pillow, cutoff frequencies above 4KHz. At first I thought you wanted to reduce loud background noise or people or snakes hissing at you. :twisted:
 
Voice

ok thanks for the insight - im not too bothered with quality really telephone quality seems fine for my application. However I was concerned with noise. Have you ever used the MAX chip you specified in your earlier post?, if its easyto use ill try and get hold of one. Also have you ever used an A/D to digitise voice signal can you recommend a simple (very simple) device that I could use?

thanks again
 
Hi Electricman,
You should make the sound quality as good as you can. Where is noise going to come from?
I used National Semi's similar switch-cap filter ICs many times. They are easy to use and work very well. National doesn't make theirs anymore.
They filter so well you can feed a square-wave in and a pure sine wave without any harmonics comes out.
 
Intercomms as well as telephones are all derived through baseband, baseband multiplex, has nothing to do with Modulation in the AM frequencies. Unless you have some sort of wireless type of intercomm.

A band stop or a 4th order band stop for selectivity. It would help to know what frequency these noises are coming thru on, if the noise is a spread spectrum distortion only a analog to digital to analog (ADA) filter will clean it up.
 
Hi Juglenaut,
The 8th order Butterworth switched cap filter IC that I recommended is like a brick wall, it removes noise and distortion above its cutoff frequency very well.
The high quality intercom system I worked with used time-division multiplexing instead of relays for switching so that anyone could talk to anyone else, without blocking caused by running out of relay paths. Because of the multiplexing, all audio in it was modulated and demodulated, something like wireless.

How can ADA remove wideband random noise and distortion any better than a low pass filter? It will digitise noise like voice and play it back the same plus its own quantization noise if it doesn't have enough bits.
 
My application

I want to use the Switched cap filter to pass anything below 5khz to the audio amp. The MAX291 seems ok here -My design is for an optical intercom it is FM with about 40Khz carrier. At RX end the PLL demodulates I want to filter afterwards - its quite nosiy at the moment. Also I can send music OK for say 2m (currently with low power Ir LED's) but if I plug in a MIC at the TX I can't get anything im figuring I need a MIC amp of some sort, only if I shout into it do i get some kind of response at the RX - Any idea's??
 

Attachments

  • rxnewer.gif
    rxnewer.gif
    7.4 KB · Views: 2,342
The 1K input to the op amp is attenuating the signal, try this circuit which has 1 meg input resistance.

The formula for R-C cuttof is: F = 1/2/PI/R/C.
If you can identify the source of the noise, it may be better to kill it at the source. The 741 may be the source of the noise, better op amps are available.
 

Attachments

  • preamp_193.jpg
    preamp_193.jpg
    77.8 KB · Views: 2,186
Hi Russ,
Ouch, ouch and ouch again! What are you doing with a DC and AC gain of (gasp!) 1001? You don't need nearly so much gain unless you were trying to pickup IR reflections off the moon!

Just think about how noisy the signal will be, if the opamp even had an output. If the opamp has an input offset voltage of only 6mV, its output would be at 6V. Plus the input bias current difference beween one input with a 1K resistor and the other with 1M. I am sorry but your circuit will probably rectify the loudest sounds or its own noise and not pass the others.
Add a cap from the 1K resistor to ground and the opamp will be biased correctly and might make a pretty good FM limiter to reduce the noise, but won't have much gain with an ultrasonic carrier.
I agree with you that the old 741 is too noisy for audio.

Let's concentrate on the LM565 circuit. It has huge caps so its VCO is tuned for a 40Hz carrier! A bit too low. I think that even 40KHz is a poor choice, beware of anyone pointing a TV remote at it. The LM565 circuit also has a huge cap across its output.

With a mic input a preamp certainly is needed, and if the mic is an electret then it needs a power/load resistor.

National Semi still has an application note for their obsolete and discontinued LM565. It shows an FM system using 100KHz and 200KHz carriers for stereo communication on the AC mains. They use a multiple-transistor IC (it might also be discontinued) as an FM limiter and even show a modulator/transmitter circuit. They can easily be modified for using IR.
https://www.electro-tech-online.com/custompdfs/2005/03/AN-146.pdf
 
Hi Russ,
He, he. I'm not done yet.
I think the original circuit has a gain of only about 9. The collector of the transistor is a fairly high impedance current sink, maybe exactly 220K at times. I think the circuit will work about the same with the 1K resistor bypassed.
 
ok what does this mean guys!!!!

Russ in the circuit you drew what is the purpose of R4 is to filter the input if so why does it have a cut off of 1.5Hz? How do you guys suggest that I connect up the input- Will it help to use a switched cap filter IC??
Thanks for all the help
 
The opamp input has to have a DC path to ground, that is the only purpose of R4. C1 is a DC block, you can set the cutoff frequency wherever.
If you make R1 100K, the circuit should work, even with 741 but please try a better op amp. SSM2135 from Analog Devices is low noise.
 
Hi Electricman,
Russ admitted that his design had errors. The purpose of its R4 is to provide a 0V reference voltage for the opamp. Without R4 the DC output voltage could be anywhere and cause severe distortion.

The input capacitors of each circuit have a value much too high for a 40KHz FM system.

I bet your noise is coming from the noisy old 741 opamp. And I bet you have another one in your transmitter. Use low-noise audio opamps like TL071 instead.

Are you actually using 1uF caps in the LM565 circuit and a 47uF cap across its output as shown on your schematic? Haven't you seen the datasheet and the link to the application note I posted? The values of your caps are about 1000 times too high for a 40KHz FM system.

I just realized that the high level of the 40KHz carrier at the output of the LM565 could messup a switched cap filter and your LM386 output amp.
 
Mic problems/LM565 values/noise/hissing

Sorry my diagram was incorrect im using 3.8k 2nf cap by the oscillator
frequency formula f0 = 1/(3.7CoRo) its abut 35Khz and I matched
that to the TX output.

My issue was regarding the fact that when I input audio (a walkman etc) into the transmmiter I get an output from the RX at about say 2m at
the moment its noisy (in terms of hissing) but it works.

Now when i attach a microphone (not powerred to the transmitter I get
nothing by speaking when I shout I hear something though its
completely unrecognisable against the hissing if I turn up the gain I just get more noise + distorsion!!!

Its for an intercom. The diagram below is the TX and the previous one is the RX there will be one of each on both sides.
Im guessing mic can't supply enough power the problem is I
don't really want to use a powered mic because I want the user to be able to plug in their own headset.

Russ/Audioguru I have the TL071 op amp will this have lower noise???- I want to increase the gain from the MIC but im also increasing the noise/hissing also which i dont want.

Also Im assuming (though im not too sure why!!) that if the system/design can transmit audio, as in music, it can definately
work for voice communications without too many modifications.

Is that a reasonable assumption or have a got to go back to the drawing board.??

Thanks for the help guys
 

Attachments

  • txin-out.gif
    txin-out.gif
    46.3 KB · Views: 988
Yes, the TL071 is low noise. The circuit should work OK for voice. I would put a 75 ohm resistor in series with the LED to limit the current, the current output of the 555 is not well controlled but is at least 200 mA.
 
Hi Electricman,
Thanks for posting corrections to your schematic and its important details.
I was right! You are using a noisy old 741 opamp as a mic preamp. Of course a TL071 will be much, much better.

It looks like you are using an electret type of microphone, its terminals might be backwards and they don't have much output when backwards.
The schematic shows a two-wire mic, but the icon shows a completely different 3-wire mic. You must be certain which type of mic the users will have or maybe have a separate circuit and input jack for each type. They might not even be electret mics and need a different circuit again!

The mic preamp circuit looks odd:
1) When you turn the volume up, part of the volume control loads-down the mic, reducing its output. The other part of the volume control increases the gain and noise of the opamp. Low sound and high noise!
2) As you turn the volume up, it isn't capacitor-coupled so it messes-up the biasing of the opamp, causing horrible distortion!
3) The load/powering resistor for the mic is the variable volume control so it is correct at only one setting, where the gain and biasing won't be correct.
4) The center-supply voltage reference of the opamp at its pin 3 doesn't have a filter cap so will be noisy, especially with a 555 high-current driver on the same battery supply!

The FM modulator is a 555 high-current driver. Its supply decoupling capacitor isn't detailed and might not be big enough to filter its 400mA supply current transients it produces when switching. The transients feed directly into the mic preamp's input. I recommend having a separate resistor-capacitor supply decoupling network for the mic and preamp.
The 555 might be operating with a 10% duty-cycle which must be very confusing for the LM565 PLL. I recommend re-calculating its timing parts so its output is close to a 50-50 square-wave.

I will post the preamp circuit that I use with electret mics soon.
 
duty cycle/555/lm565

Hi guys my previous schemeatic was not quite accurate- I posted the new one. Anyway ive changed the Reciever op amp to a TL074. When the TX variable is set to 37K that corresponds to 53.75Khz by(1.44/(Ra+2Rb)*C) with an audio input I get a good quality audio (music) with little noise att the output. The PLL has (with 3.8K and 2nF cap ) has a centre frequency of 35Khz which makes me suspicious though. This is for the IR RX and TX Led's no more than 10cm apart- After that I lose signal strength, but now im pretty sure the electronic noise problem is sort of under control. The noise I get whilst moving the Tx and Rx apart I think is mainly from optical sources from the environment become more predominant as he incident flux (IR) on the RX diminshes with distance .. the hissing decreases significantly as i dim the lights!!!- I plan to use better IR phototransistors with much less suceptibility to ambient light.

Audioguru - I want to test with a 2 terminal electret mic which is what I a have not sure how to use the thing though. Anyway ill carry on testing with audio for the moment.

The problem is I need a range of at least 5m with this system. At the moment Im using vary basic TX and RX infra red LED's. Nowadays there are some great emmiters about which can take a 1A in pulsed operation they should eaasily be able to provide the power I need(I hope). but the duty cylce must be low or you will burn out the devices.

It says the 555 can sink 600mA I think this should be enough except Im not sure how to control the current output at pin 3. Do any of you guys know how i can make a high current driver to drive an array of transmitter LED's (maybe 3-4) from the output of the 555????

You mentioned about the getting a 50-50 square wave-Is this the mark/space ratio??

Thanks
 

Attachments

  • 111new.gif
    111new.gif
    10.2 KB · Views: 1,116
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top