AN AUTOMATIC LEVEL CONTROL SYSTEM
By J. C. Jones
Copyright July 13, 2001

 
What I REALLY need at church is a robot that will sit at the controls of the mixer for the pa system and as different people speak or the same speaker with different emotional outburst,etc gently move the volume control up or down to compensate for the change in behavior.

In this electronic age, there are such things as automatic gain controls, compressors, limiters and expanders on the market, but these all suffer from the same malady. They are all active amplification units whose amplification is  varied depending on whether the input is  too low or too high. The undesirable side-effects of these types range from increasing  background noise when low levels exist, to un-natural sounds when a high peak comes through the system, (quick clamping,etc)

In the usual true mixer setup, when you change the level you are changing passive element (variable resistor) that changes the amount signal that goes to the next stage. All stages have  fixed amplification factors.

I have formulated a system based on the true mixer arrangement, using an opamp
conditioning input, a digitally controlled attenuator, an opamp out with a fixed gain, plus a basic stamp (or similar) micro computer unit.  Here is how it would work.
 


 

Assume an input to the conditioning opamp(input) is 24db. (unity gain)

1. digitally controlled attenuator normally would be set  to 20db attenuation
2. For an input of 24db, input to final opamp would be +4.
3. Gain of output opamp would be set for 20db
4. Output under normal conditions would equal input.

5. Scenario #1 normal operation 11vac symmetrical,
     a. Dual rectification and digital measurement of the output would be sent to mcu.
     b. Processor would do an absolute of measures (no sign) and take the larger of the two
     c. No difference it would select positive data and compare it to reference of 11v.
     d. Since it would be equal to this value, no adjustments would be made to the attenuator.

6. Scenario #2 input level less than 11vac symmetrical
     a. Dual rectification and digital measurement of the output would be sent to mcu
     b. Processor would do an absolute of measures (no sign) and take the larger of the two
     c. Data difference and larger value taken and compared to reference of 11v
     d. (Assume 10v is value, comparison is done to determine new attenuation)
     e. Since input is less, then less attenuation would increase output to acceptable level
     Note: Here is where the system varies from conventional systems. When level is low, you 
can decrease your attenuation as much as you desire but since there is a set level input to the  attenuator, there will be  a point where nothing else will happen. (no breathing, sucking up, etc)

     f. Once comparison is completed, new data to attenuator is sent, establishing new level
7. Scenario #3  Input level is more than 11vac symmetrical
     a. Dual rectification and digital measurement of the output would be sent to mcu
     b. Processor would do an absolute of measures (no sign) and take the larger of the two
     c. Data difference and larger value taken and compared to reference of 11v
     d. (Assume 12-volt input, level comparison is done to determine new attenuation)
     e. Since input is more, then more attenuation would be needed to keep output at an acceptable level.
     f. Once comparison is completed, new data to attenuator is sent, establishing new level

A new page outlining program in non-code description coming soon.

Note: No correction of less than 1 db would be made. all attenuation values would be in interger amounts, only.