please critique this level indicator

[gain]

would like to implement it twice and build it into my preamp.

thanks.

[Iain McNeill]

Well it'll give you a linear scale (rather than a logarithmic scale like we hear) loosely based on the peak level. You won't get anything until the input signal is greater than 0.6V because the 741 buffer can't bias the diode.

I'd turn the 741 into a precision rectifier and add x10 gain. If you want to get really fancy then Analog Devices has log amps/multipliers that could give you a dB scale.

edit: whoops, something seriously wrong here. -ve of cap goes to +ve rail? If the LM3916 is -15V referenced then you'll always have the lower half LEDs on and only get display on the top half.

Cap should be ground referenced as should the LM3916 and providing a discharge for the cap is also necessary or it'll be a peak hold meter.

[gain]

so by providing 10x voltage gain in the 741 we are making the 0.6 v drop across the diode insignificant? sounds good. and i thought 3916 was log and the 3914 was the linear one .. no?

[Iain McNeill]

on third thoughts:

swap power connections.

stick diode in "DIV_LO" to 0V connection to compensate for diode on 741 output

reference cap to 0V

increase gain to suit (line level = 1V right?)

[gain]

edit: whoops, something seriously wrong here. -ve of cap goes to +ve rail? If the LM3916 is -15V referenced then you'll always have the lower half LEDs on and only get display on the top half.

Cap should be ground referenced as should the LM3916 and providing a discharge for the cap is also necessary or it'll be a peak hold meter. [/B][/QUOTE]

accidentally flipped the +/- on the supply rails, sorry bout that. the cap is tied to the neg rail.

[gain]

breadboarding ... brb

[Iain McNeill]

an esoteric point but, there's no point waiting for the cap to charge from -V to 0V on start up - just reference it to 0V.

LM3916 vs 14 - hmmm, distance memories evoked. you could be right re. log scale. It's what the datasheet says....

Seeing as the 100K pot is the discharge path, the LM input impedance may reduce the discharge time at max gain (input impedance in parallel) your time constant is about 600ms so this things gonna dance pretty good.

[gain]

got it working on the breadboard. the mods i tried on this circuit were first changed the gain to 10x in the 741 as Iain McNeill suggested, and trimming down the pot/divider accordingly. this resulted in the bottom 3 led's turning on in sequence as the sound came up instead of all coming on at once like before (when the 0.6V mark was crossed).

tying the cap to the ground rail produces unexpected effects as the led's seem to 'smear' when signal is applied (all glowing dimly but in sequence with the music). the led's also remain glowing very dimly with no signal applied. retying to -15V eliminated this.

as expected, higher values for c make the display a little more sluggish while low value ones make it faster and more smear like.

the 3916 runs very warm to the touch with all 10 led's lit, although im within spec according to the datasheet. think i should reduce led current? figure it will be rare when all 10 led's are all on for an extended period of time.

[Iain McNeill]

Sounds like the ground rail isn't really ground. There's going to be a significant current through the averaging cap and if this is running through the "ref_adj" or "div_low" pins rather than straight to PSU/741 then this noise injection will make LED jitter (for want of a better word)

Can you post a pic of the BB?

edit: and noise injection into the MODE pin modulates the LED brightness.

[MrTransistorm]

Take a look at National's datasheet for the LM3916. There's a wealth of info in there.