Circuit to eliminate Dead Air

Years ago, I had a VOX circuit in the program

line.  Usually the VOX was running in the ON

position, but if there was too much dead air,

the relay would go to the OFF position.  This

would trigger a latching circuit which put a

weather radio into the line.  I also had it set

up a different way - so that a station ID would fire

off from a 20 second audio memory chip when the

program audio stopped.  When the program audio 

came back, the ID would turn off.  (This was for

the space in between songs.)

This setup was very simple and had problems.  Sometimes

it worked really well.  Other times I couldn't get it to work

to save my life.  

Your set-up looks much much better.  

Bruce, The Dig Radio Group 

The Vmid signal is a bit confusing. Are you limited to a single voltage power supply? If you are not then it would be better to connect U1 pin 3 to the power supply common and use a + and - supply.

The gain for U1 is likely too high.

There is a better op amp circuit for rectifying low level signals. Search on op amp precision rectifier. This will work down to millivolts so you won't need much gain at all.

On the 4066's you need to keep the audio signals between VDD and VSS limits, preferably midway between them. If you used a dual supply (+ and-) then the audio signal can be kept centered around ground and referenced to ground. Some CMOS switches are subject to latch-up if this isn't done. This would be another advantage to using a dual supply.

It is not obvious but trying to use a single supply for an op amp circuit can cause all sorts of problems, for example the input impedance may not be constant.

Neil

Neil, The Vmid voltage is my "Fake Ground", a point halfway between real ground and Vcc (15 VDC).

Yes, I noticed some problems with my Vmid floating around untill I lowered the the Voltage Divider Resistors (R15 & R12) to 1K.  I guess I was just being lazy and should consider a dual supply.  (All the other stuff uses single rail suply).  Guess I figured the input caps allow theaudio to float and "fake the op amps out" to make them think Vmid was actually Ground.

Should I get a voltage converter like this and get rid of the Fake Ground nonsense?

http://www.linear.com/product/LT1026

I'm checking out the Precision Op Amp Rectfier now, however I have to be careful about noise triggering if I make it too sensitive. The ROKU Box (that feeds the audio) is pretty high-level and when the audio cuts out, there's probasbly some signal there.?

Thanks for your help, Neil.  Gotta run now and boil my first run of maple sap!

The gain for U1 is likely too high.

I guess my theory was to drive OP AMp U1 into saturation with a fairly small signal , say 20 millivolts. I don't want this thing to trigger if the DJ walks away from the mike for a second or 2. 

I also want to charge the cap at the same rate whether I have a 20 mV signal or a 1 volt signal. 

Perry,

Since this is going to be a go or no-go situation then it is probably not necessary to have great precision in the circuit. I wouldn't bother with a voltage converter and if the single supply works then go with it. I am used to designing instrumentation where precision is needed so maybe I am biased a bit.

Be aware that noise on the power supply will appear in the audio signal through R5, R10, and Vmid since these couple to the audio signal. C6 could made larger if needed.

You have a clever circuit so let us know how it works.

Neil

As Neil pointed out, the problem with voltage divider "fake grounds" is obtaining a stable ground reference.

An interesting "fake ground" circuit is used in the DBX-118 compressor/expander.

A "floating" DC power supply is used.  A power transformer suppies a voltage doubler which develops its output across two filter capacitors in series.  

The doubler supplies a 723 regulator which provides the positve output voltage.  

An op-amp in closed loop mode is biased to provide 1/2 of the regulated 723 output voltage at its output.  This becomes the ground reference.

The minus supply is the other end of the voltage doubler.

In this way, you have +/- voltage compared to the op-amp output which is the circuit ground return path.  This provides a very stable and stiff voltage point midway between the +/- output.

It took a bit of thinking to figure this one out seeing the output of the op-amp grounded.  I'm not sure why they went with this design but it does work.

Be aware that noise on the power supply will appear in the audio signal through R5, R10, and Vmid since these couple to the audio signal. C6 could made larger if needed.

Actually, I'm still not sure I even need R5 and R10.  I put them in to give the Audio a DC reference point at Vmid (7.5 VDC).  Without the resistors, the audio (AC SIGNAL) may just float to the point halfwayy between Vdd and Ground anyway.  I'll find out, I got the 4066's today. 

Don't float, leave the Rs in. Some CMOS switches can latch up if the input voltage goes above VDD or below ground. I don't recall if the 4016/4066 series do this or not but I wouldn't risk it. Latch-up can cause circuit damage.

Neil

An FM station I worked at ran giant reels of music tape on their subcarrier, which served grocery stores and dentists offices. While we handled our on air shift it was easy to ignore the background reels, so the engineer hooked up a silence sensor which activated a large horn that was loud enough to be heard all the way across the ocean, and there was no way to ignore it.

I should have poked my nose inside his circuit.

Okay, she's not pretty but she's working under test conditions.  I made a few minor component changes, but otherwise, the circuit shown above is it.  I have the primary Audio supplied by the Sony Voice Recorder/mp3 player and the output goes to a set of computer speakers.  The audio sounds great running through the analog switch. (I can't notice a difference).  I may still want to increase the gain some to allow for very quiet passage nuicance triggering.  I also may want to add a Schmit trigger to the comparitors to provide some hysterisis.  I also stuck a Red LED which lights up when Backup Audio is enabled.

The right hand chip has 4 op amps and I use three of them (1 for amp and 2 for comparators).  The left had chip has the 4 analog switches.

Got the card all soldered up and mounted in the Audio Switch Box that I used for the Box.

I replace the 20K  bleeder resistor with a 50K Pot to allow for a variable delay before the Backup Audio kicks in.  I've tested the unit and the comparators are flipping properly with the test audio.  Now to plug it into the system between my Primary Audio and my Transmitter and see how tit sounds.

Here's my circuit card and the switch box.  I bought a cheap 10$ switch box complete with a bunch of RCA phono plugs.  It seemed easier than buying a box, drilling holes and mounting my own connectors; plus the ribbon cable worked nicely for connecting to my card!:

Here's my card mounted in the Box.  Fits nicely!  I'm using 7 of the RCA phono plugs (2 for primary audio, two for backup audio, two for output to the sStran and one for a  monitor.  I was able to use the ribbon cable that was in the switch box and connect it directly to a plug I installed on my card:

Hi Perry,

That is a great accomplishment.

I have been following this thread from its start. I looked carefully at your schematic and tried to find anything wrong. I really couldn't! I guess I should have posted some words of encouragement.

You obviously are quite expert in practical circuit design, simulation and fabrication.

The part 15 world needs more people like you.

Nice concept, planning, and implementation. Thanks for sharing this.

Neil

Thanks for the kind words Phil, though I did have some help from Bill, Neil, LT Spice and the internet.  Just hated to put all this effort into hooking up a Part 15 station only to come home and hear Dead Air.   And thanks to you Phil for the two Sstran's.  What a pleasure to build and to listen to!  I fooled around with the Talking House's and once I heard the sstran, I put the Talking Houses in the junk pile.