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

[JohnRoberts]

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

When you know the gain law of the gain element it is reasonably straightforward to compute gain reduction from the control voltage.

The task is made difficult when we do not have a known gain law to use.

Comparing the voltage output to the voltage input generally gives a linear delta, but the meter can be scaled in dB thresholds to deliver useful GR display.

I would be really surprised if nobody has done this already. The ancients are always stealing our ideas.

JR

 

[ruffrecords]

When you know the gain law of the gain element it is reasonably straightforward to compute gain reduction from the control voltage.

JR

The gain law of pretty much all optos is log10(resistance) = K * log10(LED current) where K is a factor that depends on the opto and the pot divider values used.

Cheers

Ian

 

[gyraf]

..and tempco, temperature, and production tolerances - but still, that is why I suggested going the 1178-route earlier..

Jakob E.

 

[Seeker]

... 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...

So I've been trying to wrap my brain around the actual implementation of this... I feel like I am missing something, but can we use the LM3915 and connect the input (pre gain element) to the signal in, and the output (after the gain element) to the ref in? 

 

[CJ]

a meter for output level is nice,

use your ears when turning the GR knob, the meter will screw you up.

in other words, put the lid on the box and move on...

 

[JohnRoberts]

So I've been trying to wrap my brain around the actual implementation of this... I feel like I am missing something, but can we use the LM3915 and connect the input (pre gain element) to the signal in, and the output (after the gain element) to the ref in?

OK no, and probably not.

First the reference needs to be calculated from the input (pre gain cell) level. Then the output level gets compared to this input reference to determine how many dB of attenuation is occurring, based on how much lower it is.

WRT using a 3915, I don't know and I am not motivated to dig up a data sheet to see how flexible it's internal reference string is. I never used a 39xx in any commercial product, but did fire one up on my bench back in the 80s (to do a simultaneous Peak/VU meter by switching between dot and bar mode while simultaneously commutation between peak and ave level input.).

So you might be able to get a 3915 to work, but it will be harder than just rolling a simple comparator from a couple LM339s  and your own resistor string. If you really feel compelled to try the 3915 study the data sheet and application notes specifically about the internal reference string.

You will also need to rectify and smooth input and output levels separately.

Good luck. 

JR

 

[scott2000]

no.

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

JR

What about ART???Closer???

 

[Seeker]

...So you might be able to get a 3915 to work, but it will be harder than just rolling a simple comparator from a couple LM339s  and your own resistor string. If you really feel compelled to try the 3915 study the data sheet and application notes specifically about the internal reference string.

You will also need to rectify and smooth input and output levels separately.

Good luck. 

JR

Thanks for clearing up my thought process a bit John, i figured it would not be that simple. 

a meter for output level is nice,

use your ears when turning the GR knob, the meter will screw you up.

in other words, put the lid on the box and move on...

I'm going to take Jacob's advice and go with the 1178 ish metering, it should track peak well enough, and I should be able to adjust the release a bit if its an issue... I only really watch the peak anyway.

I actually have a pico comp that I use on everything I mix these days that has no meter.... but it does more of an rms type job, so hitting it too hard isn't  much of a problem. 

 

[JohnRoberts]

Thanks for clearing up my thought process a bit John, i figured it would not be that simple. 

I'm going to take Jacob's advice and go with the 1178 ish metering, it should track peak well enough, and I should be able to adjust the release a bit if its an issue... I only really watch the peak anyway.

I actually have a pico comp that I use on everything I mix these days that has no meter.... but it does more of an rms type job, so hitting it too hard isn't  much of a problem.

IIRC the Pico is based on THAT chip set so you may find advice about metering in THAT app notes.

JR

 

[Seeker]

IIRC the Pico is based on THAT chip set so you may find advice about metering in THAT app notes.

JR

Thanks John, you are correct... I actually have the schemo with metering, etc.  I  never got around to adding the metering because once i got it in service, I thought it sounded so good on just about everything, I didn't want to take the lid off again.    ;D