lm3916 led vu meter implant on gssl

[dagoose]

I'm going to implement a lm3916 vu meter into my gssl and i have some questions about it.
What i want to do is make it switchable between input and output and so i'm looking for decent points to sniff the signal.
I was thinking of using the same points as the crush and blend installation by livingnote since they are unbalanced and are 'tapped' anyway.

Is this a good idea or should i use other points maybe because it will degrade the signal to much in combination with the crush and blend?
I was also thinking of just using the + and ground side on the real input and output (so on the xlr plugs that is) but i think it might give some problems with impedance then.

I'm also thinking about the summing of left and right since it will be just a 'mono' vu meter.
Will resistors of something like 47k do the job or is there another better way to do it? I know Livingnote was working on a different version but I want to keep it as simple as possible.

About the lm3916, since it is based on consumer levels (?) i think the input needs some changes to give real VU scale values. Is there a standard thing to use for the input?
http://www.datasheetcatalog.org/datasheet/nationalsemiconductor/DS007971.PDF

 

[livingnote]

Hey dagoose,

here we go - the thread of my little sojourn into wonderful dB-land

http://www.groupdiy.com/index.php?topic=28403.0

Just buffer them and you'll be alright. Grab a 072 or so and use one for input, one for output as a summer and set the gain of the buffer to something which makes setting Vref easy (I twiddled the **** out of the ref voltages to get some mV refs - not to be advised). Put the switch after the buffers. Optionally set Vlo to your dB minimum since it's a logarithmic scale and you're good to go.

Use whatever detector scheme you find useful from the datasheet, for me it was the "precision full-wave peak detector to PPM standards", I think that was the second to last one.

Matter of fact you can just use an 074 and have all 4 opamps you need for some nice meter action

 

[dagoose]

OK, going to check that out, see if i can breadboard something nice
Picked up a lm3916 and the leds today, now it's time for a really cool new frontpanel.  :

 

[nielsk]

How about using it as a gain reduction meter?

 

[dagoose]

How about using it as a gain reduction meter?

I allready have a GR led meter, the one from the pico compressor, works perfectly!

 

[nielsk]

Ahh, so Kimosabe!

 

[dagoose]

Ahh, so Kimosabe!

BTW if your'e thinking of inserting it to your gssl, these are the values for the resistors, thanks to radiance.

R1= 225K (FS deflection @ 20dB GR)
R2= 112.5K (FS deflection @ 10dB GR)
R3= 56.25K (FS deflection @ 5dB GR) (made mine with 27k for about 2db GR)
R4= 100K (Feedback resistor)

 

[dagoose]

ok... been busy thinking about implementing it now since the rest is done allready.  ;D

I need some L R summing and debalancing or the single unbalnced input of the lm3916.
I was thinking of going for a tl072 approach as in the pdf file below.
Will this do the job?

BTW... it was my first time using eagle soft for drawing something.  8)

http://www.dagoosemusic.nl/temp/lrsum.pdf

 

[livingnote]

eagle = fun  ;D

Is there any reason to use the balanced signal for metering? I just took them pre-output circuit and summed 'em...

 

[Harpo]

Hi dagoose,

feeding the LM3916 from the audio part (this is not sidechain control voltage) will need some rectification IMHO.
Your file showing the (not needed) deballancing, ending at the summing node.
Something like this rough scetchup

 

[matthias]

a good alternative is the sanyo lb1412

- 12 leds with peak hold
- internal rectification
- not much parts needed to make it work...

http://www.bader-frankfurt.de/elek/praezpegelanzeige.gif