Peak detector for VU meter

[ljcox]

A peak detector is supposed to have a very fast attach and a slow release that cannot happen if the second opamp has a capacitor across its negative feedback resistor.

Did you look at the NS circuit I posted in post #4? The second Op Amp is configured as an Integrator in order to do the arithmetic & to filter the Full Wave rectified signal to provide a DC level. Thus the R6 & C5 filter in the circuit posted by the Op is not necessary.

 

[audioguru]

The original circuit posted by the OP has a fast charging and slow discharging peak detector, not an averaging detector.

 

[ljcox]

The original circuit posted by the OP has a fast charging and slow discharging peak detector, not an averaging detector.

Yes, I know. But I asumed that it would not matter whether it was average or peak.
However, on reflection, perhaps that assumption is wrong like a few of my other assumptions in this thread.
I seems that it does not pay to make assumptions.

 

[audioguru]

I think it is very important for a sound level measuring device to have a peak detector.
The first detector described on the datasheet for the LM3915 is simply nothing. It makes a blur of LEDs with the brightest part the average and you can barely see the peaks.

 

[ljcox]

I think it is very important for a sound level measuring device to have a peak detector.
The first detector described on the datasheet for the LM3915 is simply nothing. It makes a blur of LEDs with the brightest part the average and you can barely see the peaks.

Yes, that sounds reasonable. The peaks would be what causes overload when recording, so the average is of limited value.
I have never built a audio level meter. However, I have a vague recollection that the old style moving coil VU meters had damping in order to make them show the average.
A moving coil meter would not be fast enough to show the peaks.