Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Matthias Heitzer]

Quote:

Quote: Matthias Heitzer wrote on Mon, 22 November 2010 07:33 Even in the analoge world, a fader can't boost the signal, it's just a variable resitor. +6 or (+10) printed on the surface next to the slot just means that the signal is amplified by that amount before the fader. Not true; pretty much all modern consoles use the fader as a variable voltage divider in front of a gain-of-10 dB op-amp. When the fader is at unity, it's actually attenuating by 10 dB, which is made up by the following op-amp.

In the context of the first post,
it does not really matter at wich side of the fader the op amp sits (or at both). There are different designs and all of them (except  VCA or DCA) have to somehow make up some loss since the "0dB position" is not at the end of the fader in most cases.
"variable voltage divider" is just the application the variable resistor is used for


Of course, this opamp does more than just boosting the signal a bit. No one denied that.

[Ivan Beaver]

Matthias Heitzer wrote on Mon, 22 November 2010 09:33

Even in the analoge world, a fader can't boost the signal, it's just a variable resitor. +6 or (+10) printed on the surface next to the slot just means that the signal is amplified by that amount before the fader.

I don't know if there are any modern consoles like this, but historically (Tapco comes to mind), the "volume control" was actually in the feedback loop of an op amp.

So as you turned the knob clockwise-the resistance increased-and therefore the opamp circuit had more gain.

The problem with this, is that if the wiper should lose contact with the carbon track, the opamp would go into open loop gain (BUNCHES OF LEVEL INCREASE), and you would have instant feedback.

[Matthias Heitzer]

Seems like I've opened a can of worms with my posting. I did not intend to make a detailed statement about the exact circuitry of  all consoles in the whole wide world, i just tried to explain that the actual amplification does not happen in the fader (since it's a passive component) but somewhere else. Therefore the markings on the surface are relative (well that's what our logarithmic friend dB always is). It depends on the other gain stages throughout the console wheter something is really amplified or not. In both the digital and the analouge world. Digital bords have the limitation of 0dBFS, and analoge circuits are also limited, for example by the rail voltage.

Just wanted to answer to this sentence:

Quote:

but, afaik, there's no such thing as "+" or amplification on the digital realm, only analog can do that

I did not want to bring more confusion than necessary, so i kept my first posting brief and simple. obviously this had just the opposite effect.

[Brad Weber]

Brian Ehlers wrote on Tue, 23 November 2010 13:11

It's also not true that "in- and output levels are within a narrow range by now." An easy example is a pro CD player which outputs +24 dBu for CD data at 0 dBFS.  That's WAY above the "standard" pro interface level of +4 dBu.

I agree with the general point that source levels, both signal and noise, can vary greatly.  However, in your example 0dBFS is the maximum or peak level possible, so that means +24dBu would be the maximum or peak analog output level of the CD player and not the average signal level, which is what the +4dBu represents.  A +4dBu average level for a signal with a 20dB crest factor, which is high but not unreasonable, would be a +24dBu peak level.

This does seem to highlight the importance of knowing what any level metering is showing.  Whether a meter is displaying dBu, dBFS, dBV or whatever and whether it is showing peak, VU or average levels can make a significant difference in what is represented.

[Brian Ehlers]

Very good point, Brad.  It's important to understand the "ballistics" of your console's meters.  I am assuming that most modern boards with LED (not mechanical) meters are indicating peak voltages, or at least something closer to peak than to average.  But my assumption may be wrong.

Which brings us full-circle back to the importance of understanding our gain structure.....

[Andy Peters]

Brian Ehlers wrote on Mon, 29 November 2010 12:15

I am assuming that most modern boards with LED (not mechanical) meters are indicating peak voltages, or at least something closer to peak than to average.  But my assumption may be wrong.

Your assumption is wrong. There is no reason to assume that an LED meter has any particular ballistics, so RTFM.

An LED meter is capable of displaying both peak and average values at the same time -- see the metering on the APB Dynasonics consoles.

-a

[Brian Ehlers]

RTFM?  No need to get snippy.  I have read the manual for my console, and it says absolutely nothing about the time-domain behavior of the LED meters.  However, I do have the full schematics, so I'm able to reverse engineer it.

Quote:

There is no reason to assume that an LED meter has any particular ballistics

Actually there is reason.  By far the easiest circuit to implement to drive an LED "meter" is to have multiple comparators (at multiple reference levels) monitor the voltage on a capacitor which is fed a half-wave rectified version of the signal.  The capacitor is charged through one resistor and drained through another resistor.  Those two RC time constants determine how quickly the LEDs will attack and then release when a peak comes along.  Typically, the attack is made relatively quick (so that the meter responds to peaks), and the release is made relatively slow (so that your eye can actually see the LED light up even if the peak is quick).

Case in point:  my church's Crest X-four console.  The 5-segment meter for each channel strip uses an attack time constant of 100 microseconds and release time constant of 200 ms.  While the attack time constant does cause some smoothing of the signal before it is metered, it's response will still be well under 1 ms.  These really are peak meters, not averaging, and certainly not true-RMS.

Reverse engineering the 15-segment meters on my output buses is not as easy, since an off-the-shelf IC is used.  The release time constant is outside the IC and is 100 ms.  But the attack time constant is inside the IC.  Still, based on how the meters behave and how such circuits are usually implemented, I assume the attack is well under a millisecond.  So again, these are peak meters.  Regardless, their time-domain behavior will be far different than that of an old-school mechanical VU meter.

That's cool if the APB Dynasonics consoles offer you an option for the metering ballistics.  What's interesting is that the guys who designed those consoles are the very same guys who designed the Crest X-series consoles (before Peavy bought up Crest), and the consoles are "cut of the same cloth."  Sounds like they made a nice improvement to the metering (though personally I'm much more concerned about peaks than averages).

[Andy Peters]

Brian Ehlers wrote on Wed, 01 December 2010 11:10

RTFM?  No need to get snippy.  I have read the manual for my console, and it says absolutely nothing about the time-domain behavior of the LED meters.  However, I do have the full schematics, so I'm able to reverse engineer it. Quote: There is no reason to assume that an LED meter has any particular ballistics Actually there is reason.  By far the easiest circuit to implement to drive an LED "meter" is to have multiple comparators (at multiple reference levels) monitor the voltage on a capacitor which is fed a half-wave rectified version of the signal.  The capacitor is charged through one resistor and drained through another resistor.  Those two RC time constants determine how quickly the LEDs will attack and then release when a peak comes along.  Typically, the attack is made relatively quick (so that the meter responds to peaks), and the release is made relatively slow (so that your eye can actually see the LED light up even if the peak is quick). Case in point:  my church's Crest X-four console.  The 5-segment meter for each channel strip uses an attack time constant of 100 microseconds and release time constant of 200 ms.  While the attack time constant does cause some smoothing of the signal before it is metered, it's response will still be well under 1 ms.  These really are peak meters, not averaging, and certainly not true-RMS. Reverse engineering the 15-segment meters on my output buses is not as easy, since an off-the-shelf IC is used.  The release time constant is outside the IC and is 100 ms.  But the attack time constant is inside the IC.  Still, based on how the meters behave and how such circuits are usually implemented, I assume the attack is well under a millisecond.  So again, these are peak meters.  Regardless, their time-domain behavior will be far different than that of an old-school mechanical VU meter. That's cool if the APB Dynasonics consoles offer you an option for the metering ballistics.  What's interesting is that the guys who designed those consoles are the very same guys who designed the Crest X-series consoles (before Peavy bought up Crest), and the consoles are "cut of the same cloth."  Sounds like they made a nice improvement to the metering (though personally I'm much more concerned about peaks than averages).

{sigh}

-a

[ProSoundWeb Community]