adding gr metering to an optical comp when you can't match the opto cells

Some years ago I built a dual la3a ish compressor and made my own opto cells... they came out sounding very nice, but i neglected to think about making ones for metering as well...    :-[    So in the interest of putting a lid on this project, i'm wondering what my options are for adding some kind of gr metering, aside from a matching opto cell?  It seems like there must be a way of tracking input vs output and driving a meter.    It'd be nice to have something other than an led indicator...

Not sure if it's helpful but Ian (ruff) was talking through something in this thread .....

https://groupdiy.com/index.php?topic=52847.0

You don't really need matched optos. Using a matched opto is simple but you could also just tap into the CV with a little extra circuitry. Using the difference of the output vs input could work too but it would depend on the presence of signal which is not ideal and probably require more circuitry than just tapping off the CV.

Ask yourself: How much precision in GR monitoring do you REALLY need?

..and then use a system like in the 1178..

Jakob E.

It'd be nice to have something other than an led indicator...

Click to expand...

For precision, go LEDs.
Maybe only one LED for indicating comp action on or off?

The Chinese sell cheap VU meter boards on eBay. So if you want a needle bouncing round and about more or less meaningfully. ..

Or if precision is really not the goal, how about a magic eye tube? They also sell cheap from China. Real eye candy. Just don't expect any meaningful indication here. But could serve double function as power On indicator

Thanks for the replies, I’ve tried the one led indicator thing but, I find a meter a bit more useful.

gyraf said:

Ask yourself: How much precision in GR monitoring do you REALLY need?

..and then use a system like in the 1178..

Jakob E.

Click to expand...

gyraf said:

Thanks Jakob, this looks like a simple solution to my problem!

Click to expand...

Seeker said:

...... must be a way of tracking input vs output and driving a meter. 

Click to expand...

I was thinking the  same thing for my optos. But seeing as I know more about coding than electronics I was thinking of  using something from the arduino family, would need 2x ADCs as inputs to read before and after compression and enough outputs for the LEDs. The code should be straightforward....

Why not measure the analog in and out, and compare with a diff amp. Just my first thoughts Doesn't have to work at low levels

You would have to do it in log domain for something like that to make sense...

Jakob E.

I suppose just putting them after the full wave rectifiers would be too simple. I did play with logamps a few decades ago, but not very successfully

gyraf said:

You would have to do it in log domain for something like that to make sense...

Jakob E.

Click to expand...

+1... 

First let em get this out of the way.... "why".  :  OK on to solutions.

I have designed many log conversions using transistor arrays (the Vbe junction voltage characteristic with current is logarithmic).  Nowadays with cheap transistor arrays history, the simplest (accurate) solution is to use a pair of THAT corp analog engines that contain precision rectifiers and log conversions inside ICs (ignore that you would be discarding decent VCAs). Subtract the output log result from input for dB difference (voltage will need to be scaled.) This requires two analog engines so not cheap.
====
For a crude GR meter, over a limited dynamic range and resolution, here is an idea to chew on. Using several LM339 quad comparators calibrated to respond to something like 3 dB level steps (over modest range), with an input set of comparators responding to input level, and an output set responding to output level, the 339 comparators could be configured to either add X mA per 3dB of input level, or subtract the same X mA per 3dB of output level from a single summed current node. The net current would represent the level "difference", albeit crudely and over limited dynamic range.  This is probably impractical to realize both fine resolution a wide dynamic range but there is no reason why you can't scale the threshold width and output current contribution to provide more resolution around 0VU and less resolution at lower levels (imagine 1dB steps for top 10 dB, 3 dB steps below, 3dB steps provide 3x current per threshold, to maintain scaled results.) This could be reasonably accurate.

JR

PS: no I don't feel like designing one.

JohnRoberts said:

+1... 

First let em get this out of the way.... "why".  :  OK on to solutions.

I have designed many log conversions using transistor arrays (the Vbe junction voltage characteristic with current is logarithmic).  Nowadays with cheap transistor arrays history, the simplest (accurate) solution is to use a pair of THAT corp analog engines that contain precision rectifiers and log conversions inside ICs (ignore that you would be discarding decent VCAs). Subtract the output log result from input for dB difference (voltage will need to be scaled.) This requires two analog engines so not cheap.
====
For a crude GR meter, over a limited dynamic range and resolution, here is an idea to chew on. Using several LM339 quad comparators calibrated to respond to something like 3 dB level steps (over modest range), with an input set of comparators responding to input level, and an output set responding to output level, the 339 comparators could be configured to either add X mA per 3dB of input level, or subtract the same X mA per 3dB of output level from a single summed current node. The net current would represent the level "difference", albeit crudely and over limited dynamic range.  This is probably impractical to realize both fine resolution a wide dynamic range but there is no reason why you can't scale the threshold width and output current contribution to provide more resolution around 0VU and less resolution at lower levels (imagine 1dB steps for top 10 dB, 3 dB steps below, 3dB steps provide 3x current per threshold, to maintain scaled results.) This could be reasonably accurate.

Click to expand...

Or to get farther into the weeds, how about a typical led bar graph driver with the usual comparators, where signal hits one side of all comparators, and the other side is the usual resistor string from a reference voltage.

Instead of a reference voltage feeding the resistor ladder, use it as a second input, the uncompressed signal. This way, instead of comparing signal to a reference voltage, you are comparing output signal to input signal.

Just change AC signals to DC envelopes with similar time constants.

Would it work?

Gene

Gene Pink said:

Or to get farther into the weeds, how about a typical led bar graph driver with the usual comparators, where signal hits one side of all comparators, and the other side is the usual resistor string from a reference voltage.

Instead of a reference voltage feeding the resistor ladder, use it as a second input, the uncompressed signal. This way, instead of comparing signal to a reference voltage, you are comparing output signal to input signal.

Just change AC signals to DC envelopes with similar time constants.

Would it work?

Gene

Click to expand...

I may need to drink a few more beers to understand your concept, but as a perhaps simplification, if we ASSume only gain reduction (it is a GR meter), perform a FW rectification on both the input and output. Connect the input  rectified voltage to the top of a resistor divider string  calibrated for -1dB, -2dB, ... etc.  Compare the FW rectified versions of the output to this divider string from the input.

Presto gain reduction LED meter...

Less parts than mine...I like that.  Win goes to Pink... (if this is what he was describing... or not).  :

JR 

PS: I'm still not drawing a schematic.  8)

JohnRoberts said:

(if this is what he was describing... or not).  :

Click to expand...

That's pretty much it.

Gene

Unfortunately, comparing the input level with the output level does not really tell you much about the amount of gain reduction. Every time you alter the make up gain, the difference between input and output levels changes but the amount of compression does not.

Cheers

Ian

ruffrecords said:

Unfortunately, comparing the input level with the output level does not really tell you much about the amount of gain reduction.

Click to expand...

If comparing the level before to level immediately after the gain element, it will give us the gain reduction caused by that gain element. 

Every time you alter the make up gain, the difference between input and output levels changes but the amount of compression does not.

Cheers

Ian

Click to expand...

Not sure that I follow.  If the make up gain is added inside the measurement loop it will be included (if not, not). This is a gain reduction meter, not a compression meter.  I don't know how to make a compression meter without capturing and crunching some other dynamic characteristics.

JR

JohnRoberts said:

If comparing the level before to level immediately after the gain element, it will give us the gain reduction caused by that gain element.

JR

Click to expand...

You are, of course, right. Silly me.

Cheers

Ian

JohnRoberts said:

I may need to drink a few more beers to understand your concept, but as a perhaps simplification, if we ASSume only gain reduction (it is a GR meter), perform a FW rectification on both the input and output. Connect the input  rectified voltage to the top of a resistor divider string  calibrated for -1dB, -2dB, ... etc.  Compare the FW rectified versions of the output to this divider string from the input.

Presto gain reduction LED meter...

Less parts than mine...I like that.  Win goes to Pink... (if this is what he was describing... or not).  :

JR 

PS: I'm still not drawing a schematic.  8)

Click to expand...

Is this gain reduction metering similar to what is being talked about????

scott2000 said:

Is this gain reduction metering similar to what is being talked about????

Click to expand...

no.

Perhaps the same result, to represent gain reduction with LEDs  but not remotely the same approach.

JR

Thanks....Just trying to put a name to a face  :-[.... To understand better.....