Very simple Audio compressor for FM transmitter

[ronsimpson]

When the FET has the two resistors on the gate with a series capacitor then the distortion is reduced a lot but the attack time is very slow. The circuit from Rod Elliot that I posted (post #2) reduces distortion and is fast.

VOLTAGE CONTROLLED ATTENUATORS
The CAD cell has slow response and we know why.
All the rest of the versions we have talked about are fast. Faster than we need.
The "slow" is a function of attack time and release time and not the speed of the gain block.
Here is the MOSFET circuit I used. In this example the sine wave is 1khz. The control changes every 10mS.

Here I speed up the sine to 10khz and the control up to 1mS.

 

[audioguru]

The Mosfet appears to attenuate and release very quickly. How is its distortion?

 

[MrAl]

Hi again,

ronsimpson:
Thanks for the circuits. They look very interesting. The only thing that worries me a little is that the drain of the MOSFET appears to be biased at 0v DC, which means sometimes it goes negative. Is that device (or similar) going to be that symmetrical in it's response? I have not tried this yet but i would think offhand that it needs to be biased to some reasonable DC level like 1/2 Vcc, which would make the drain always positive even though it would still vary up and down with the AC input signal.

audioguru:
So are you saying that you think it would be ok to substitute a MOSFET for the FET in the circuit you posted in post #2 then? If so we can try that too. The whole idea now is to be able to use a MOSFET so Willen does not have to find an FET and he already has some MOSFETs that are similar to the 2N7000 type, hopefully they are still working because he did not know that they are static sensitive.

 

[Willen]

...he already has some MOSFETs that are similar to the 2N7000 type, hopefully they are still working because he did not know that they are static sensitive.

First, I shorted Gate and Source just a second then tested Source-Drain conduction, It has no conduction. Then I touched to Gate by finger just a second then it starts to conduct Source to Drain. I think it is ok.

 

[audioguru]

Touching the gate of a Mosfet can easily ZAP it. How many thousands of volts are there at the end of your finger?

If the drain of the Mosfet has a DC voltage then the gate control voltage will cause it to make a loud POP sound if it is turned on quickly.

Years ago I worked with large high quality (translation: very expensive) intercom systems that were installed in the head offices of large companies and banks.
It used pulse-amplitude-modulation to multiplex up to 12 voices over one wire. It used Signetics (later bought by Philips) NE572 compressor/limiter ICs so that all voices had the same level.

I designed a background music fade up and fade down module using a Jfet. It had very low distortion. Many were installed in the intercom systems.

 

[ronsimpson]

MOSFETs can and do run in reverse. It is OK to have a small amount of reverse voltage.
Note: It is important to keep the output audio below the nee voltage of the MOSFET internal diode. I think the peak output voltage needs to be in the range of 200mV to 400mV. The input voltage can be very high. Because this is a 'limiter' the whole idea is to have a wide range of input voltages and a limited output range.

The 1000 drain resistor can be much lower like 200 or 100 ohms. It depends on the GS-voltage to DS-resistance curve and where you want to run.

The circuit that 'measures' the audio is critical. We should use the one Audioguru likes. We need fast attack and slow release.

Does anyone have this thing running? I think we could pull out the JFET and put in the MOSFET+ resistors and see what happens. Slight biasing changes might need to be done. And the Drain resistor needs some experimenting.

 

[canadaelk]

Why not use the THAT2180, it is in a 7-pin SIP and easy to use. I have about a hundred of the 2150 from old projects (some new, some reclaimed), the precursor to the 2180. PM me and I will mail them (max. 5), free, as long as I have stock. E

 

[MrAl]

Hi,

canadaelk:
Thanks, Willen was looking to do this project as he was the one who wanted to build the limiter/compressor. I think he might appreciated it if he got some parts but you'd have to mail them to him.
I'll see what he says about this too.

 

[ronsimpson]

Why not use the THAT2180, it is in a 7-pin SIP and easy to use. I have about a hundred of the 2150 from old projects (some new, some reclaimed), the precursor to the 2180. PM me and I will mail them (max. 5), free, as long as I have stock. E

I like the THAT parts. Do you have any 4301? They will make a good compressor/limiter.
OR
Your 2180 and a 2252. Will make a good limiter.

 

[canadaelk]

Sorry Ron, only the 2150s are spares. E

 

[Willen]

Wow wow wow...

I am feeling I found more advanced limitter than old JFET limiter. THAT 2180A datasheeet says it has very low distortion!

If any one could find THAT 4301, it would be a complete package i think but if nobody has it then also finding 'THAT 2180' (don't know which one is best 2150 or 2180) made me so happy than before!!!

I guessed the datasheet has no complete and practical schematic exactly about Audio Peak Limiter for FM broadcasting purpose, can anybody help to find (Or modify) the circuit of great and advanced limiter based on 2180? Pease! Found an application **broken link removed** but it is preamp anly.

More excited! Thank you Canadaelk for the chips!!

 

[audioguru]

The THAT voltage controlled attenuator ICs are very good.
In addition to the old Signetics/Philips NE572 "compander" IC there was another manufacturer who made a good voltage controlled amplifier IC SSBxxxx (maybe) that is also obsolete.

Some good peak limiter circuits are in the datasheet of the LM3915 bar graph IC.

 

[MrAl]

Hello again,

Another interesting idea would be to use a microcontroller with either a digital pot or a multiplying DAC. Anyone try this yet? Im not talking high end uC here though that has lots of DSP features i am talking just low end uC to keep cost way down and to keep it simple yet get something that can do decent compression. I am guessing that even 256 different discrete levels would be good enough, maybe even just 128 or even less like 64, as long as we dont hear much stepped breathing in the audio.

 

[Willen]

I like the THAT parts. Do you have any 4301? They will make a good compressor/limiter.
OR
Your 2180 and a 2252. Will make a good limiter.

THAT Corp is sharing their ICs as a sample free of cost within USA, isn't it?
http://www.thatcorp.com/THAT_Sample_Request.shtml
You people are lucky being there who can buy all components and also who can request lots of components as a sample, free of cost from almost every semiconductor company.

Some good peak limiter circuits are in the datasheet of the LM3915 bar graph IC.

I downloaded a datasheet from National semiconductor, but unfortunately it has a partial schematic of Full-wave peak detector. I searched in THAT corporation's site and got some design notes, please check it out- http://www.thatcorp.com/Design_Notes.shtml
Aren't there any complete peak audio limiter circuit made by 2180 IC exist around the world? Now a complete circuit has been my main problem!

 

[audioguru]

Why do you want a compressor anyway? The extremely high amount of compression used on radio stations and TV sounds awful. The loudness of the music is modulated by the beat of the music (many people don't notice because they think it is normal in the music). Voices sound like a barking dog.

CDs already have some compression on them. Then of course copies of CDs (MP3s) also have some compression on them.

I was going to ask if you have ever heard live music but it is all amplified and compressed today. IN THE GOOD OLD DAYS live music was distortion less and had a very wide dynamic range.

 

[MrAl]

Reply to Willen's post #54:

Hi,

Didnt you check your own link out at least a little bit <chuckle> ?

http://www.thatcorp.com/Design_Notes.shtml

The first link in that link shows a schematic for a limiter compressor. Might be a little too complex but it does have a full wave rect in it too you might use.

 

[Willen]

Why do you want a compressor anyway? The extremely high amount of compression used on radio stations and TV sounds awful.....

hahahaha your each styles of examples makes me laughing actually, eg:-
-"Voices sound like a barking dog."
- about your local TV 'breathing before speaking sounds like a hurricane'
- 'first pronounce of a voice sounds like a shut gun' etc...

I just want to compress audio and want to limits its peak, that peak is causing distortion now like frying potato, in Mod-4 if I turned input volume little high, but broadcasting audio is still little smaller than my local powerfull station. So I want to broadcast louder audio by limiting its peak without distortion. I will feel realy good if I did so but I don't know more old thing and old experience... My two of local station has very loud and nice audio. I am inspired by them.

Alternatively I downloaded this 16KB zip file (jar inside) which is said to be a compressor/limiter software. I used it in PC and played a song on a media player. I adjusted faders on software (don't know how much is necessary values) but I didn't feel more than distortion. I transmitted the compressed audio from my FM tx but again feel no difference between 'compressed' and 'bypassed' audio. Can you check it once audioguru? (I used Realtek HD audio input and Realtek HD audio output in the software.) **broken link removed**

Didnt you check your own link out at least a little bit <chuckle> ?

http://www.thatcorp.com/Design_Notes.shtml

The first link in that link shows a schematic for a limiter compressor. Might be a little too complex but it does have a full wave rect in it too you might use.

I saw this circuit, it won't be complex to make for me but there is a RMS detector chip 2252, and you can simply understand that it is impossible to get. Canadaelk said that he can send 2150 or 2180 chips but may be he don't have these 2252 RMS detector chip. Then I didn't focus on this circuit which is impossible without 2252 chip. I am feeling so much tired...aren't you?

 

[MrAl]

Hi,

Oh ok, well that makes sense then, but one thing to keep in mind is that <another chuckle> that company obviously wants to sell more chips, so they stuck in an RMS converter chip too. But this isnt mandatory, you can use the average value which most other compressors use anyway. Granted it is a good idea to to use an RMS signal rather than an averaged signal, but the difference wont be enough to worry about unless you have a very demanding application. So just skip the RMS chip and connect the input to the full wave rectifier circuit and go from there. Might have to read that app note to know what level to expect out of the RMS detector so we know what resistor to replace it with, but that shouldnt be too hard and can probably even be done experimentally.

 

[canadaelk]

THAT parts are available through Mouser. E

 

[audioguru]

I just want to compress audio and want to limits its peak, that peak is causing distortion now like frying potato, in Mod-4 if I turned input volume little high, but broadcasting audio is still little smaller than my local powerfull station. So I want to broadcast louder audio by limiting its peak without distortion. I will feel realy good if I did so but I don't know more old thing and old experience... My two of local station has very loud and nice audio. I am inspired by them.

I also hate hearing distortion. So my sound systems have plenty of "headroom". Headroom is having some extra space like in my home I can jump as high as I can without hitting my head on the ceiling because there is at least 1m of headroom above my head.

In 1975 everybody took a "boom box" to the beach (there are huge lakes near where I live). A boom box was an portable AM-FM radio with a cassette player. It had a maximum output power of about 0.5W and played one-note bass. People turned them up too high so that the output was severely distorted.
But not me. I made mine with a little sub-woofer powered from my own 6W amplifier with bass boost and two satellite speakers each powered by my own 3W amplifiers. I used a rechargeable Ni-Cad battery. Very nice sound, fairly loud and NO distortion.

The stereo in my home has lots of headroom. It has never produced distortion because with it playing VERY loud there is still plenty of headroom. I don't know how loud is its maximum. Same for the sound system in my car.
I think a cheap FM radio causes distortion because its IF has a narrow bandpass. My stereo and car radio have a modern IF that has a wide bandpass with a flat top but with very steep sides. The wide bandpass allows very loud FM (high deviation) without distortion.

Alternatively I downloaded this 16KB zip file (jar inside) which is said to be a compressor/limiter software. I used it in PC and played a song on a media player. I adjusted faders on software (don't know how much is necessary values) but I didn't feel more than distortion. I transmitted the compressed audio from my FM tx but again feel no difference between 'compressed' and 'bypassed' audio. Can you check it once audioguru? (I used Realtek HD audio input and Realtek HD audio output in the software.) **broken link removed**

Thanks. I downloaded it and will try it later because I am going out (the temperature is -8 degrees C but it is nice and sunny) now.