[sdiy] Improving LED response when driven from half-wave rectifier

[KA4HJH]

On Sep 13, 2012, at 12:38 PM, Mattias Rickardsson wrote:

>> What I'm referring to is the response of human vision, which is non-linear. If you use exponential conversion, the LED will *appear* to be fading in and out in a more linear fashion. I've used the old reverse-biased LED pair myself in the past and I wasn't satisfied as they always seem to spend most of the time being brighter than 50%.
>> 
> CV values on a linear scale need to be exponentially converted, just as you say... but both vision/brightness and hearing/loudness is perceived exponentially-ish, so visualizing an _audio_ signal with an LED is more straightforward, I guess? :-)

A single LED may well work as you suggest. I remember a simple LDR-based compressor by Anderton or Orman that had an LED to show the level of compression and it was very intuitive. 

On a more general note, visualizing audio signal levels requires measurement and display of a greater dynamic range (and higher resolution) than control voltages in order to be useful, hence the use of analog meters and digital bar graphs. A single LED would best be used as a clipping indicator. OTOH, the main purpose of displaying the level of a control voltage is to see how it is changing, not to determine its peak or average value. A single/bi-color LED works well for this...especially if the response is exponential.

The crazy thing is that it would probably be cheaper (or at least simpler) to program a small MCU to perform the conversion than doing it with an analog circuit. The downside is you'd have to deal with RFI.


Terry Bowman, KA4HJH
"The Mac Doctor"