Mixer internal level indicator

[JAY X]

Hi!

I'm currently designing a two levels signal present/peak indicator for a summing mixer. This indicator will represent the output level of the summing amplifiers (unbalanced output), BEFORE a balanced line driver, that gives +6dbu gain.

The signal present will light up at -20dbu.
The peak indicator will light up at ¿-2dbu?    (-2dbu + 6dbu = +4dbu)

Or, the internal level should be ¿+4dbu?   

I'm a bit confused... ???

Attached is a simple schematic 

Jay x 

 

[ruffrecords]

If your nominal output level is meant to be +4dBu then you need to set the indicator threshold to -2dBu to indicate that +4dBu to the outside world has been achieved.

Are you sure +4dBu represents your peak output level?

Cheers

Ian

 

[JAY X]

Are you sure +4dBu represents your peak output level?

Cheers

Ian

Hi,
Of course not, I was referring to the internal level. This is why i got confused... :
Máximum peak level would be Around +26dbu. For +/-17v rails. Aprox.
So, if i was to indicate two levels, one would be -20dbu, and the other, the peak that the circuitry can handle without distorsion...right?

Jay x

 

[JohnRoberts]

We have discussed this before, I think I even posted a schematic of the one op amp bicolor LED circuit I used back last century.

When using output stages with +6dB of voltage gain (from two output drivers), I would scale nominal internal levels down the same 6dB (to -2 dBu) so the internal circuit paths would saturate at the same time as the output drivers. 

JR

 

[zamproject]

Hello Jay

The nominal internal level is up to you and the mixer design.
Mixers I know have about 25/30dB at each mixer section "nominal" level and max level margin,
max is given among others for a specific THD...also up to the designer to pick a number... 0.1...0.5...1%

Now the output line level is up to the studio/operator (within the capable output range of device, but "pro" console have no issue about this)

I'll put output nominal peak level 10dB below max studio level which in the case of a PPM mixer can be your "max peak"

So let say your mixer output +26dBu max but your recorder/ADC is +22dBu, you better set your studio level at +12dBu (peak) !
To translate this if you pick the signal before line driver your level trig for the meter is +6dBu (12-6)

But
Ideally you can decorrelate internal mixer level and I/O level in a mixer, with input and  output trim for line level alignment.

Now...
If you say max out is +26dBu without any trim gain/attenuation, it mean internal max level (unbal) is +20dBu,
I'll consider internal level to be about -5dBu (for 25dB internal headroom) which "should" be your internal peak level
And now backward...
say you still want your output at +12dBu line level, you need 17dB boost to go from -5dBu to +12dBu, you already have 6 with balancing out, and you need 11dB more at line level.... but in this case your peak (trigged at -5dBu pre line out) will show internal nominal level AND line out level, everything correlate.

Not sure I reply to your question... but my meaning is that before any visual level monitoring/setting, proper nominal level alignment at each step should be (eventually) considered.

In your case maybe I overthink it as it's a "simple" summing,
what coming in is probably already handled dynamically/spectrally and you don't need 25/30dB internal headroom, so just set your peak 6dB below line out level (to just handle the 6dB line out boost)

Best
Zam

 

[zamproject]

side note

your block diagram is little confusing
For line I/O (insert and line out) do you mean 6dB/0dB gain, or 6dBu/0dBu level ?

You might have unity gain issue between insert line I/O and line out if you want them same level (balanced line level)
I'll set the fader buffer at +4dB gain (you already have +6dB gain at line out)
But maybe you want 16dB attenuation at nominal fader position ?

A level diagram should help 

Best
Zam

 

[ruffrecords]

Hi,
Of course not, I was referring to the internal level. This is why i got confused... :
Máximum peak level would be Around +26dbu. For +/-17v rails. Aprox.
So, if i was to indicate two levels, one would be -20dbu, and the other, the peak that the circuitry can handle without distorsion...right?

Jay x

To be useful, a peak indicator should activate a few dB below clipping so that you have time to grab the relevant fader before things do reach clipping. Anything from 1dB to 3dB below clipping is typical.

Cheers

Ian

 

[Newmarket]

To be useful, a peak indicator should activate a few dB below clipping so that you have time to grab the relevant fader before things do reach clipping. Anything from 1dB to 3dB below clipping is typical.

That's it for peak indication.
Topic seems to have got a bit confused with nominal and internal operating levels rather than the max output level that defines the 'clip point'. Of course you do have to take into account any +6dB level boost introduced at output stage if that applies and you are measuring level before it.

But I do think it's more than worthwhile to have some indication that signal is hitting around nominal level.
Else it's all a bit " You have signal" ...possibly followed by "Too much signal" with nothing inbetween.

 

[JohnRoberts]

But I do think it's more than worthwhile to have some indication that signal is hitting around nominal level.
Else it's all a bit " You have signal" ...possibly followed by "Too much signal" with nothing inbetween.

I have found it extremely useful to know a) you have signal, and b) you have too much signal (I set the O/L threshold at 3 dB before saturation). As I have shared before I designed a one op amp bi-color LED driver (that old circuit supported sampling diodes so I could detect at multiple nodes in a given audio path). Green indicated signal present, and red for too hot.  The trend already last century was for consoles to have far more audio paths than meter bridge meters.

I found putting bi-color LEDs on every internal path was extremely useful, but I may be biased. Just knowing that signal is reaching a given stage of the console helps when setting up a mix or troubleshooting. The O/L indication is icing on the cake.

These don't replace proper meters for L/R master and perhaps subs, but for multiple audio paths anything is better than nothing.

JR 

PS: Sorry over the years I wasted too much time thinking about this stuff. My old bi-color circuit was based on components already in the Peavey manufacturing system and were machine inserted. A scratch design might find better options.

 

[Newmarket]

I have found it extremely useful to know a) you have signal, and b) you have too much signal (I set the O/L threshold at 3 dB before saturation). As I have shared before I designed a one op amp bi-color LED driver (that old circuit supported sampling diodes so I could detect at multiple nodes in a given audio path). Green indicated signal present, and red for too hot.  The trend already last century was for consoles to have far more audio paths than meter bridge meters.

I found putting bi-color LEDs on every internal path was extremely useful, but I may be biased. Just knowing that signal is reaching a given stage of the console helps when setting up a mix or troubleshooting. The O/L indication is icing on the cake.

These don't replace proper meters for L/R master and perhaps subs, but for multiple audio paths anything is better than nothing.

JR 

PS: Sorry over the years I wasted too much time thinking about this stuff. My old bi-color circuit was based on components already in the Peavey manufacturing system and were machine inserted. A scratch design might find better options.

Hi JR. Reading back what I posted it might not have been clear that I meant that indication of signal around nominal operating level would be good in addition to Signal Present and 'Peak' Indicators. Those are the most important I agree.
But third indicator for 'line level' would be the 'icing on the icing' if you like - where 'full metering' isn't practicable.
I think of it like a 'Goldilocks' system - Too Little...Too Much...Just Right...

In my own usage I'm looking at an audio interface that has indicators for input signals at -30dB and -3dB. Now I can check the actual levels to via the on screen mixer. But it would be a lot more intuitive and quicker to just have some indication that I was in the ballpark level wise just by looking at the front panel. In this particular case it doesn't help that the interface runs at well below '+4dBu' level - probably around the '-10dBu' 'consumer' level.

 

[zamproject]

Hello

Well, maybe the OP should tell us if he want a peak indicator, which to me should represent the nominal operating level, or an overload indicator which represent the absolute level the device can handle.

Best
Zam

 

[dogears]

Hi JR. Reading back what I posted it might not have been clear that I meant that indication of signal around nominal operating level would be good in addition to Signal Present and 'Peak' Indicators. Those are the most important I agree.
But third indicator for 'line level' would be the 'icing on the icing' if you like - where 'full metering' isn't practicable.
I think of it like a 'Goldilocks' system - Too Little...Too Much...Just Right...

In my own usage I'm looking at an audio interface that has indicators for input signals at -30dB and -3dB. Now I can check the actual levels to via the on screen mixer. But it would be a lot more intuitive and quicker to just have some indication that I was in the ballpark level wise just by looking at the front panel. In this particular case it doesn't help that the interface runs at well below '+4dBu' level - probably around the '-10dBu' 'consumer' level.

Soundworkshop put this in their early consoles. Maybe this does what you want?

 

[Newmarket]

Hello

Well, maybe the OP should tell us if he want a peak indicator, which to me should represent the nominal operating level, or an overload indicator which represent the absolute level the device can handle.

'Peak' usually denotes a level near clipping. Typically 3dB before clipping occurs to give a bit of response time as I think Ian has said on this thread. imo an indication of actual max level isn't a great deal of use as you're already  at the limit.

 

[Newmarket]

Soundworkshop put this in their early consoles. Maybe this does what you want?

Thanks. That looks like the sort of thing.
As it happens I can design what I want but it's not practicable to mod the audio interfaces I'm using.
But they make me think how convenient it would be without having to check on the PC.

 

[abbey road d enfer]

As I have shared before I designed a one op amp bi-color LED driver (that old circuit supported sampling diodes so I could detect at multiple nodes in a given audio path). Green indicated signal present, and red for too hot.

JR, honest, you circuit got me puzzled. I tried to sim it but there must be something wrong, since only one led gets enough current to actually light.
What's the orientation of the bi-color LED?
Which polarity do the sample diodes detect?

 

[zamproject]

'Peak' usually denotes a level near clipping. Typically 3dB before clipping occurs to give a bit of response time as I think Ian has said on this thread. imo an indication of actual max level isn't a great deal of use as you're already  at the limit.

Ok, I can agree as overload indicator is preferably done with peak (crest) detection.
Still (to me) a peak level is a value showed on a peak meter, with a peak ballistic, and it dose not necessary show overload as reference.

It still not clear what the JAY X want to show with that "peak" led.

Best
Zam

 

[dogears]

JR, honest, you circuit got me puzzled. I tried to sim it but there must be something wrong, since only one led gets enough current to actually light.
What's the orientation of the bi-color LED?
Which polarity do the sample diodes detect?

Seems to work for me.

 

[JAY X]

Hi!

Have been reading all the discussion....

My objective for the peak led is to have avoid saturating too much the summing amps inputs, with the result of a distorted
waveform at the output of the summing amps, where i place the peak led.

When i mix some of my music through the summing mixer, I have to pay attention at daw fader levels, in order not to saturate the mixbus. I have the "green book" by Douglas Self, and  it talks about mixer internal levels to have the best S/N ratio and headroom.He says -2dbu(615mv rms), gives a headroom of about 22dbu.

Maybe, instead of a red peak led, i have to place a yellow led to signal the level at which i have enough headroom left... or have 3 leds: green/yellow/red.

In the next days i will post some circuit schematics. Also one of a Vumeter buffer with Peak led, with balanced inputs, so i can read the output levels at the balanced output jacks.

Jay x.

 

[abbey road d enfer]

Seems to work for me.

Yes, I had redrawn the circuit to make it more readable for me, and I had one diode in the wrong direction.

 

[JohnRoberts]

JR, honest, you circuit got me puzzled. I tried to sim it but there must be something wrong, since only one led gets enough current to actually light.

That is production schematic so correct.

That circuit is the result of much tweaking to get a good visual display. This was designed decades ago so I will try to recall a few details. I suspect the red was more sensitive (LEDS sucked back then) so green was probably hit with more current to appear similar brightness. IIRC the purpose of zener CR130 was to establish a source of current that the green direction would steal from the zener. This also provides a sharper on threshold and keeps clicky current spikes out of the rails and/or grounds.   

What's the orientation of the bi-color LED?

I suspect + marking is for red direction it was a cheap two lead back to back LED, so lights in both directions.

Which polarity do the sample diodes detect?

The diode cathodes are connected to sample node (SMPLA) so only detects positive peaks but typical console paths with multiple sample diodes and inverting stages in between will detect spikes from both polarities. .

When the sampled input is in green signal present region the op amp works like a simple inverter with gain high enough to light LED green with modest signal,,, but when sample input voltage is high enough it pulls the op amp + input positive reversing the op amps output to up red direction. C156 provides some hold- positive feedback to stretch out narrow spikes to make them easier to see. 

This may seem like a lot of parts for a simple function but those were pretty much all machine inserted and pennies each. The 1/2 op amp, also machine inserted, was maybe 6 cents. Diodes were less than a penny.

JR