audio line out voltage

[Spehro Pefhany]

Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[martin griffith]

On Thu, 04 Aug 2005 13:47:23 -0400, in sci.electronics.design Spehro
Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:

Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany
Thats about normal for pro equipment, it gives about 20dB headroom

which seems to be the professional standard requiremnt.

quite a lot of prosumer stuff is tascam level , say about -10. so you
could get away with +10dBfor your clip level.

Where is the " power dissipation", I hope you are not terminating
everything in 600R


martin

 

[Ban]

Spehro Pefhany wrote:

Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany

Speff,
If it is a XLR input, the voltage is balanced. So the peak voltage is only
+/-4V. I use the INA2137 to convert it to single ended(-6dB) and then DRV135
to amplify and balance it again.
--
ciao Ban
Bordighera, Italy

 

[Pooh Bear]

Spehro Pefhany wrote:

Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

Yes - if don't want the input to clip it has to handle at least that much.

That much level is commonly available from pro-sumer gear. It's what you can
typically get from op-amps running on +/- 15V supplies.My own stuff with true
balanced out actually clips @ +28dBu ( ~ 19.5 V rms )

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...

Are you looking at a very low power design ? If you're simply dealing with hi-fi
you can get away with less I expect.

Graham

 

[Pooh Bear]

Ban wrote:

Spehro Pefhany wrote:
Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany

Speff,
If it is a XLR input, the voltage is balanced. So the peak voltage is only
+/-4V. I use the INA2137 to convert it to single ended(-6dB) and then DRV135
to amplify and balance it again.

Where did you get that idea Ban ?

Most pro-audio outputs are driven by an op-amps supplied from somewhere between
+/- 15V and 18V

The 8V that Spehro mentioned is a typical rms voltage from such an output. The
peak voltage in this case would be about 11V.

Where an output is true balanced both pins 2 and 3 of the XLR are actively
driven - providing a pk-pk output of say 25V.

The INA ( and other ) solutions are simply expensive playthings. You can do the
same with an op-amp and a few Rs.

Graham

 

[Spehro Pefhany]

On Thu, 04 Aug 2005 21:13:25 +0200, the renowned martin griffith
<martingriffith@XXyahoo.co.uk> wrote:

On Thu, 04 Aug 2005 13:47:23 -0400, in sci.electronics.design Spehro
Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:

Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany
Thats about normal for pro equipment, it gives about 20dB headroom
which seems to be the professional standard requiremnt.

quite a lot of prosumer stuff is tascam level , say about -10. so you
could get away with +10dBfor your clip level.

Okay. This is nothing special fidelity-wise. Trimming the supplies to
+/-8 or +/-9V (+/-6.5 or +/-7.5V swing) would make quite a difference.

Where is the " power dissipation", I hope you are not terminating
everything in 600R

No, maybe the 16 high speed buffer amplifiers, but the chip is bipolar
in any case. The data sheet limit is 40mA @+/-12V, for just one set of
supplies, which is almost a watt. Quite a bit for a little TQFP.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[martin griffith]

On Thu, 04 Aug 2005 16:09:11 -0400, in sci.electronics.design Spehro
Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:

On Thu, 04 Aug 2005 21:13:25 +0200, the renowned martin griffith
martingriffith@XXyahoo.co.uk> wrote:

On Thu, 04 Aug 2005 13:47:23 -0400, in sci.electronics.design Spehro
Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:

Hi, all:-

snip
Okay. This is nothing special fidelity-wise. Trimming the supplies to
+/-8 or +/-9V (+/-6.5 or +/-7.5V swing) would make quite a difference.

Where is the " power dissipation", I hope you are not terminating
everything in 600R

No, maybe the 16 high speed buffer amplifiers, but the chip is bipolar
in any case. The data sheet limit is 40mA @+/-12V, for just one set of
supplies, which is almost a watt. Quite a bit for a little TQFP.


Best regards,
Spehro Pefhany

What is the IC, it seems a bit thirsty, maybe there is a better option
or approach. I'm a bit out of touch with whats avaialble, but Graham
et al will be able to suggest something.
Are there space/PCB limitations?


martin

 

[martin griffith]

On Thu, 04 Aug 2005 21:01:49 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

Ban wrote:

Spehro Pefhany wrote:
Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany

Speff,
If it is a XLR input, the voltage is balanced. So the peak voltage is only
+/-4V. I use the INA2137 to convert it to single ended(-6dB) and then DRV135
to amplify and balance it again.

Where did you get that idea Ban ?

Most pro-audio outputs are driven by an op-amps supplied from somewhere between
+/- 15V and 18V

The 8V that Spehro mentioned is a typical rms voltage from such an output. The
peak voltage in this case would be about 11V.

Where an output is true balanced both pins 2 and 3 of the XLR are actively
driven - providing a pk-pk output of say 25V.

The INA ( and other ) solutions are simply expensive playthings. You can do the
same with an op-amp and a few Rs.

Graham
Agreed, 4 10K's and a TL072, CMRR not as good as a INA, but quite

adequate, generally.
But running the internals at -10dBu will prolly be OK. Depends on what
the customer wants.Most mixers I've come across never "mix" at 0dB
internally, mainly for headroom maximisation


martin

 

[Spehro Pefhany]

On Thu, 04 Aug 2005 22:12:38 +0200, the renowned martin griffith
<martingriffith@XXyahoo.co.uk> wrote:

What is the IC, it seems a bit thirsty, maybe there is a better option
or approach. I'm a bit out of touch with whats avaialble, but Graham
et al will be able to suggest something.
Are there space/PCB limitations?

Stuck with the chip.... if necessary I'll put a heatsink on it. I'm
probably being overly conservative in any case, but a supply reduction
would also give a bit more regulator headroom.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Joerg]

Hello Spehro,

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

As others have already mentioned I also believe it could be that much. I
just rigged up an FFT via audio card and when I connected the line-out
of a commercial receiver to it everything saturated, big time. On the
scope it looked like more than 10Vpp. So I had to attenuate a lot.

Regards, Joerg

http://www.analogconsultants.com

 

[John Larkin]

On Thu, 04 Aug 2005 21:22:55 GMT, Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Spehro,

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

As others have already mentioned I also believe it could be that much. I
just rigged up an FFT via audio card and when I connected the line-out
of a commercial receiver to it everything saturated, big time. On the
scope it looked like more than 10Vpp. So I had to attenuate a lot.

Regards, Joerg

http://www.analogconsultants.com

Good grief, what happened to 0 dBm? What happens when you slam 10
volts RMS into a vU meter? Is nothing sacred any more?

John

 

[Joerg]

Hello John,

Good grief, what happened to 0 dBm? What happens when you slam 10
volts RMS into a vU meter? Is nothing sacred any more?

Yes, I was shocked, too. I mean, in my case this could have fried the
built-in sound card which would have been a real pain.

What really irks me is that there is no information about that at all.
The receiver simply said "LINE OUT" next to the jack. No level. The
laptop is so skimpily documented that all I could find after 1/4hr of
forensic efforts was that it has an Analog Devices Codec in there.
That's it. Great. No data whatsoever.

I am not too surprised though considering that most people I know who
own a house never had any technical drawings for that house. Luckily we do.

Regards, Joerg

http://www.analogconsultants.com

 

[Pooh Bear]

John Larkin wrote:

On Thu, 04 Aug 2005 21:22:55 GMT, Joerg
notthisjoergsch@removethispacbell.net> wrote:

Hello Spehro,

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

As others have already mentioned I also believe it could be that much. I
just rigged up an FFT via audio card and when I connected the line-out
of a commercial receiver to it everything saturated, big time. On the
scope it looked like more than 10Vpp. So I had to attenuate a lot.

Regards, Joerg

http://www.analogconsultants.com

Good grief, what happened to 0 dBm? What happens when you slam 10
volts RMS into a vU meter? Is nothing sacred any more?

No-one uses VU meters any more. They always had serious limitations and
drawbacks anyway. The extra headroom has actually always been there. The
first mixing desk I built ran off a regulated 30V single rail supply for
example.

Bear in mind that 0VU = +4dBu - the same in voltage terms as the dBm(600R).
So +3dB on a VU meter is already +7dBu. Add to that the fact that a VU meter
is a slow response instrument and doesn't register peak voltages and ..... I
suspect you see where I'm going with this.

And indeed no-one uses dBm in audio any more except where it's actually
needed ( which is rarely ). All modern specs use the dBu which is a voltage
reference of 0.775V and all equipment is 'voltage matched'.

Another reason for high levels is digital recording. It makes sense for the
digital clip point to be similar to the clip point of the electronics driving
it. 0dBFS is often around +19dBu. There is sadly no accepted norm however.

Graham

 

[martin griffith]

On Thu, 04 Aug 2005 14:34:10 -0700, in sci.electronics.design John
Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:
snip

Good grief, what happened to 0 dBm? What happens when you slam 10
volts RMS into a vU meter?

it increases its accuracy...........................


martin

 

[martin griffith]

On Thu, 04 Aug 2005 23:17:30 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

snip postings by fine people

Another reason for high levels is digital recording. It makes sense for the
digital clip point to be similar to the clip point of the electronics driving
it. 0dBFS is often around +19dBu. There is sadly no accepted norm however.

Graham
Bleeding F*ing stupid, all the audio ADCs' I've seen run off 5v rails,

Wavefront, Ashai, etc and clip at about 2.8V, diff in (from memory,
mine, not the computers)

Good (insert appropriate deity) who the hell designs these I/F's when
you have to pad it down anyway to feed the ADC

Next we'll be having 3 phase DC supplies from the audiophoole society


martin

 

[Joerg]

Hello Martin,

Next we'll be having 3 phase DC supplies from the audiophoole society

Nah, it goes in waves. The next step would probably be a straight
capacitor from the plate of an ECC81 to the line out jack and a 10M
bleeder. That should be good for a 75Vpp swing or so.

Regards, Joerg

http://www.analogconsultants.com

 

[martin griffith]

On Thu, 04 Aug 2005 23:06:57 GMT, in sci.electronics.design Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Martin,

Next we'll be having 3 phase DC supplies from the audiophoole society

Nah, it goes in waves. The next step would probably be a straight
capacitor from the plate of an ECC81 to the line out jack and a 10M
bleeder. That should be good for a 75Vpp swing or so.

Regards, Joerg

http://www.analogconsultants.com
Shit Joerg You beat me to it

I'd just hit send when I thought about ECC83s

3rd phase, phasor current aka filaments!

Analogue ready..........................
bed time, to sober up!


martin

 

[Pooh Bear]

martin griffith wrote:

On Thu, 04 Aug 2005 23:17:30 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

snip postings by fine people

Another reason for high levels is digital recording. It makes sense for the
digital clip point to be similar to the clip point of the electronics driving
it. 0dBFS is often around +19dBu. There is sadly no accepted norm however.

Graham
Bleeding F*ing stupid, all the audio ADCs' I've seen run off 5v rails,
Wavefront, Ashai, etc and clip at about 2.8V, diff in (from memory,
mine, not the computers)

The ones I'm using can take a differential 4V pk signal IIRC. Prabably the same as
your 2.8V rms ?

Good (insert appropriate deity) who the hell designs these I/F's when
you have to pad it down anyway to feed the ADC

Actually it's fine. The resistive pad affects the noise as well as the signal
! ;-)

The operating levels are set by accepted practice.There's good reason for this if
you look at a noise analysis. -10dBV working is *always* noisier for example.

Next we'll be having 3 phase DC supplies from the audiophoole society

Well.... Certain audio companies can already offer you a 10kW amplifer. The
audiophools wouldn't have anything to do with them though since they're transistor
and don't have enough of the *right kind* of distortion to suit them - lol !!

Graham

 

[Pooh Bear]

martin griffith wrote:

On Thu, 04 Aug 2005 21:01:49 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

Ban wrote:

Spehro Pefhany wrote:
Hi, all:-

What's the minimum *peak* voltage without clipping that I should allow
for in a general purpose audio input that accepts a signal from an
audio "line out"? I see numbers like +20dB, which would be almost 8V
RMS. Is this correct?

I'd like to reduce the voltage supplies as much as practical on a
crossbar switch to reduce the power dissipation...


Best regards,
Spehro Pefhany

Speff,
If it is a XLR input, the voltage is balanced. So the peak voltage is only
+/-4V. I use the INA2137 to convert it to single ended(-6dB) and then DRV135
to amplify and balance it again.

Where did you get that idea Ban ?

Most pro-audio outputs are driven by an op-amps supplied from somewhere between
+/- 15V and 18V

The 8V that Spehro mentioned is a typical rms voltage from such an output. The
peak voltage in this case would be about 11V.

Where an output is true balanced both pins 2 and 3 of the XLR are actively
driven - providing a pk-pk output of say 25V.

The INA ( and other ) solutions are simply expensive playthings. You can do the
same with an op-amp and a few Rs.

Graham
Agreed, 4 10K's and a TL072, CMRR not as good as a INA, but quite
adequate, generally.

Absolutely so.

But running the internals at -10dBu will prolly be OK. Depends on what
the customer wants.Most mixers I've come across never "mix" at 0dB
internally, mainly for headroom maximisation

Some older ones ( esp ecially ) do and Mackie got a bad rep for doing exactly that
too.

I typically go for -6dB on the mix bus ref external levels. Indeed I sometimes run
channels at -6dB too. Gives *plenty* of headroom.

The only mixer I've ever come across that ran at -10dB ( both channels and mix bus )
was the Neve 51 series for the broadcast market. Those European broadcasters can be
very fussy.

Graham

 

[martin griffith]

On Fri, 05 Aug 2005 00:20:20 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

martin griffith wrote:

On Thu, 04 Aug 2005 23:17:30 +0100, in sci.electronics.design Pooh
Bear <rabbitsfriendsandrelations@hotmail.com> wrote:

snip postings by fine people

Another reason for high levels is digital recording. It makes sense for the
digital clip point to be similar to the clip point of the electronics driving
it. 0dBFS is often around +19dBu. There is sadly no accepted norm however.

Graham
Bleeding F*ing stupid, all the audio ADCs' I've seen run off 5v rails,
Wavefront, Ashai, etc and clip at about 2.8V, diff in (from memory,
mine, not the computers)

The ones I'm using can take a differential 4V pk signal IIRC. Prabably the same as
your 2.8V rms ?

Good (insert appropriate deity) who the hell designs these I/F's when
you have to pad it down anyway to feed the ADC

Actually it's fine. The resistive pad affects the noise as well as the signal
! ;-)

The operating levels are set by accepted practice.There's good reason for this if
you look at a noise analysis. -10dBV working is *always* noisier for example.


Next we'll be having 3 phase DC supplies from the audiophoole society

Well.... Certain audio companies can already offer you a 10kW amplifer. The
audiophools wouldn't have anything to do with them though since they're transistor
and don't have enough of the *right kind* of distortion to suit them - lol !!

Graham
I'lll prooly give a reply when I'm sodbder, dudu, keyboorrsd cant spel

writre, but think weeere in the same ballpark,

hic


martin