Mixing Console Ground Configurations-Layout

[SIXTYNINER]

Hi ,
start this thread after a research inside Gdiy and not found dedicated specific thread about as well ,
where include and talk about  :
The basic ground - gnd - earth info / concepts
The common errors (to avoid) ,
Channel strip power supply and audio signal ground layout
Console chassis ground layout , and power distribution as well
Correct ground layout for a console in studio and in a live stage

and other related aspects .

thanks for any info-help  post

 

[ruffrecords]

That is a tall order. I don't think it is all in one place. You could start here:

https://drive.google.com/open?id=0B_n67A1hN3qtSWduMTdWRUtYQ0U

Cheers

Ian

 

[SIXTYNINER]

That is a tall order. I don't think it is all in one place. You could start here:

https://drive.google.com/open?id=0B_n67A1hN3qtSWduMTdWRUtYQ0U

Cheers

Ian

Thanks so much for post the very good doc link , Ian ,
please keep it available as possible ,

and about a question regarding  old (70 age)  modules with internal wires made with cables ,
how they "manage" the grounds / gnds ?
with cables shields _?

also do you remember type of cables used for inside module connections ?
was something with quality cables like mogami or standard cables ?  ( esoteric "Hi-FIers"  apart)

 

[ruffrecords]

Thanks so much for post the very good doc link , Ian ,
please keep it available as possible ,

and about a question regarding  old (70 age)  modules with internal wires made with cables ,
how they "manage" the grounds / gnds ?
with cables shields _?

also do you remember type of cables used for inside module connections ?
was something like mogami ?

Shields and grounds are mostly different things with different purposes. I can only tell you how it was done at Neve when I was there back in the 70s.

All plug in modules were fully screened. Apart from balanced or sensitive signals, most of the module internal wiring used regular hook up wire.

Outside the modules, power and ground were distributed using large copper or  brass bus bars. Each module was locally decoupled with 1000uF. Buses were carried in a special extrusion that consisted of a number of U shaped channels, one for each bus. Each module was hand wired to the bus via a bus resistor inserted in a feed through in the extrusion. When the bus wiring was complete a plate plate was attached to the top of the extrusion so each bus lived in its own little screened box. Balanced wiring was done with tape lapped twin screened cable with a drain wire. I do not know what make it was.

0V and chassis were kept completely separate and provision was made so they could be connected (at one point only) at the Technical Earth.

Cheers

Ian

 

[nielsk]

I have wondered what the thinking was behind the 1,000 uF decoupling on each module, I do not know of anyone else who did this.
Much more common to find 100 uF or 10 uF.
It did not seem to give them a lower noise floor than other designs...

 

[ruffrecords]

I have wondered what the thinking was behind the 1,000 uF decoupling on each module, I do not know of anyone else who did this.
Much more common to find 100 uF or 10 uF.
It did not seem to give them a lower noise floor than other designs...

It serves several purposes. First as a local energy store. The average current in a class A circuit is constant but there will be short peak current demands as the level changes. A local energy store supplies these so the power supply does not have to. Secondly, this also reduces cross-talk by reducing current changes in the power supply distribution. Lastly it does  improve the noise figures slightly..

Cheers

Ian

 

[nielsk]

Ahh soo!
I wonder if this was empirically tested, as in measuring current consumption on a channel with a steady state tone going through it while an adjacent channels had, say a gated  kick drum slamming? Tried with different values and so on...

 

[ruffrecords]

Ahh soo!
I wonder if this was empirically tested, as in measuring current consumption on a channel with a steady state tone going through it while an adjacent channels had, say a gated  kick drum slamming? Tried with different values and so on...

A complete set of test results for each console was included in its handbook. Cross-talk was measured at 100Hz, 1KHz and 10KHz if I remember correctly.

As I said in my previous post, the average consumption in a class A circuit is constant and independent of signal level. For every increase in current draw there is a corresponding decrease in the next half cycle. You just need sufficient C to provide the increases and the cap recharges on the corresponding decreases, Roughly speaking, the reduction in supply current variation is given by the RC circuit consisting of the ac load and the capacitor. Worst case in a Neve console would be a BA283 driving a 150 ohm load, With 1000uF this gives a 40dB reduction at 100Hz.

One problem with all these 1000uF capacitors across the power supply was that the standard Coutant power supplies Neve tried to use would go into current limit at power up due to the inrush current into all the caps. Trouble was, the leakage current from these caps was greater than the power supply current limit so the supply stayed permanently in current limit. Neve got Coutant to make special versions with a higher current limit to overcome this.

Cheers

Ian

 

[SIXTYNINER]

Outside the modules, power and ground were distributed using large copper or  brass bus bars.

first time i saw the copper big bar on the ssl and other consoles (in the bottom of chassis)
with quite big cable from any 8ch section connected on the same bar ,

the question about was if they really need that ground bar dimension for 15V and 48V phantom voltages ?

it is quite surprising    about how much Neve designers had to engineer themselves with what available
for build the 80 series

… also  it give a good quantity of  curiosity to  see what inside the early 70' apollo space rockets electronic circuits 

 

[Monte McGuire]

had the chance to see copper big bar on the ssl and other consoles (in the bottom of chassis)
with quite big cable from any 8ch section connected on the same bar ,

a question about is if they really need that ground bar dimension for 15V and 48V phantom voltages ?

If the power supply system has little isolation between modules, and conducts rectified signal currents among the positive and negative supply rails, and thus within the ground system, then yes, extra copper will diminish the IR voltages generated by all of these rectified signal currents.

Similarly, if the signaling within the device is purely unbalanced, then the ground system is also asked to conduct linear signal return currents among all of the modules and attached IO gear, so again, throwing copper at it will reduce the induced IR voltages.

It's not the elegant solution, but a large console using bipolar power supplies and unbalanced internal signaling is quite the party of random signal currents, rectified or linear. Adding metal is certainly a lot simpler, and results in a far smaller BOM than moving to balanced internal signaling and a more complex power supply system with effective isolation between all of the sub-circuits.

Have you ever measured the mix bus noise of a large classic SSL? ;-)

 

[ruffrecords]

It is a fact of life that the majority of the signals inside a console are unbalanced. Even if you use balanced mixing this is still true. The only way to minimise interaction between modules is to make sure the power and ground conductors are as close to zero resistance as possible.

Cheers

ian

 

[john12ax7]

Have you ever measured the mix bus noise of a large classic SSL? ;-)

I'm curious about this.  If you put for example 48 faders at unity what is the output noise floor? -100 dBu,  -80 dBu, or worse?

 

[abbey road d enfer]

I'm curious about this.  If you put for example 48 faders at unity what is the output noise floor? -100 dBu,  -80 dBu, or worse?

-100  is only a dream, -80 would be the summing amp noise.
In a typical 48 ch desk, you have about 60 stems to the bus (some FX returns and sub groups), which amounts to about 35dB noise gain. If you start with a very good summing amp with -120dBu EIN (Equivalent Input Noise), that computes at -85dBu. And that would be with zero contribution from any channel.
A channel with the fader down (only its post-fader amp) has a noise of about -105dBu. When you route 48 of them, that adds up to about -72dBu. Now raise just one fader and the mic pre noise will become dominant in most cases; obviously it depends on what gain has the mic pre, and very likely the mic noise will be the dominant factor.
As JR has often theorized, the final noise performance is dominated by the noise performance of the dominant channel (generally the vocals), however, it's not a reason not to try to improve all sources of noise. Putting some figures allows somme perspective on the subject.
I must say that what makes the difference between boys and men is in the character of noise. The first time I heard a Harrison desk, I was impressed by the steadiness of its noise, devoid of any background rumble or spikes.

 

[john12ax7]

Thanks.  Having some real world numbers certainly helps a lot.

One question though,  the numbers suggest a 6 dB noise  increase per doubling of channels.  Should this be 3 dB since they are uncorrelated signals?

 

[ruffrecords]

Thanks.  Having some real world numbers certainly helps a lot.

One question though,  the numbers suggest a 6 dB noise  increase per doubling of channels.  Should this be 3 dB since they are uncorrelated signals?

Yes, it is 3dB. 10log(48) = 16.8dB

Cheers

Ian

 

[SIXTYNINER]

When you route 48 of them, that adds up to about -72dBu. Now raise just one fader and the mic pre noise will become dominant in most cases; obviously it depends on what gain has the mic pre, and very likely the mic noise will be the dominant factor.
As JR has often theorized, the final noise performance is dominated by the noise performance of the dominant channel (generally the vocals), however, it's not a reason not to try to improve all sources of noise. Putting some figures allows somme perspective on the subject.

…. Channel mute automation highly  required as well ,

I must say that what makes the difference between boys and men is in the character of noise. The first time I heard a Harrison desk, I was impressed by the steadiness of its noise, devoid of any background rumble or spikes.

it was an MR series ?

same on Mci 500 series ?

in attached an old Soundcraft tech specs from a flyer .
( will it reliable ? )  :

 

[abbey road d enfer]

Thanks.  Having some real world numbers certainly helps a lot.

One question though,  the numbers suggest a 6 dB noise  increase per doubling of channels.  Should this be 3 dB since they are uncorrelated signals?

Did I really make that noob mistake? Doh!
This is indeed true for summing signals; however for the increase of noise gain in summing amps, the increase is linear.

 

[abbey road d enfer]

it was an MR series ?

Whatever series existed in 1977...

same on Mci 500 series ?

Didn't check the 500, but I remember the 600 to be quite good in that respect.

in attached an old Soundcraft tech specs from a flyer .
( will it reliable ? )  :

Honestly, the 500/600 and 6000 were a PITN regarding noise, operating at -8dBu internal and using a weird grounding scheme (4 different grounds) that was supposed to compensate for the relatively high resistance of ribbon cable.
I never saw one of them performing in accordance with noise specs.

 

[ruffrecords]

Did I really make that noob mistake? Doh!
This is indeed true for summing signals; however for the increase of noise gain in summing amps, the increase is linear.

Noise gain in VE summer depends on whether the bus resistors of unused channels are left floating or grounded. In most cases they are grounded so the noise gain is constant, irrespective of how many channels are routed. Channel noise depends on whether the channel is routed or not and hence on the number of channels routed.

it seems to me that noise gain, and the noise resulting from it is often misunderstood. The noise gain of a VE summer is exactly the number of bus resistors plus one (assuming the feedback resistor is the same value). But the noise due the the bus resistors themselves does not increase at the same rate because the noise resistance gets small the greater the number of sources. The net result is the noise due to bus resistors also increases at the 3dB rate.

Often forgotten is that the noise gain also amplifies anything on the + input of the VE amp. As this is usually connected to 0V,  any 0V noise gets amplified considerably.

Cheers

Ian

 

[abbey road d enfer]

Noise gain in VE summer depends on whether the bus resistors of unused channels are left floating or grounded. In most cases they are grounded so the noise gain is constant, irrespective of how many channels are routed. Channel noise depends on whether the channel is routed or not and hence on the number of channels routed.

it seems to me that noise gain, and the noise resulting from it is often misunderstood. The noise gain of a VE summer is exactly the number of bus resistors plus one (assuming the feedback resistor is the same value). But the noise due the the bus resistors themselves does not increase at the same rate because the noise resistance gets small the greater the number of sources. The net result is the noise due to bus resistors also increases at the 3dB rate.

Indeed, all the different noise contributions should be evaluated separately; in practice, I have found the bus resistor noise to be about an oder of magnitude below the summing amp noise gain (in fact the bus resistor noise decreases quadratically with the number of stems when the summing amp noise increases linearly).

  Often forgotten is that the noise gain also amplifies anything on the + input of the VE amp. As this is usually connected to 0V,  any 0V noise gets amplified considerably.

+1. Actually, that's often what generates the most disturbing part of the bus noise.