Reducing cross-talk in an 80 series Neve

[nielsk]

I am in the process of moving a clients 8048, and this seems like an ideal time to try to improve the cross talk. All the gobbldy-gook (reed relay matrix and re-bussing) has been removed and the buss outs are wired directly to patch, but there is still considerable cross talk between busses.
Has anyone devised improvements in this area that do not change the sound of the console?

 

[ruffrecords]

I am in the process of moving a clients 8048, and this seems like an ideal time to try to improve the cross talk. All the gobbldy-gook (reed relay matrix and re-bussing) has been removed and the buss outs are wired directly to patch, but there is still considerable cross talk between busses.
Has anyone devised improvements in this area that do not change the sound of the console?

First question is, how have you measured crosstalk and how does it vary with frequency? Both will tell us something about the likely causes.

Cheers

ian

 

[nielsk]

Empirical data! always a good idea, too bad I do not have any. As the console is disassembled currently for the move, I can't take any measurement, but when it is back in one piece again any modifications to the frame sections will not just be difficult, but insanely difficult! (ask me how I know...)
Suffice it to say that the amount is sufficient to be a problem, a track of decent level is audible on all busses.

 

[ruffrecords]

Empirical data! always a good idea, too bad I do not have any. As the console is disassembled currently for the move, I can't take any measurement, but when it is back in one piece again any modifications to the frame sections will not just be difficult, but insanely difficult! (ask me how I know...)
Suffice it to say that the amount is sufficient to be a problem, a track of decent level is audible on all busses.

OK, understood. Can you remember if the crosstalk sounded full bandwidth or rather toppy?

Is the original extruded aluminium mix bus still in use?

Cheers

Ian

 

[nielsk]

The sound of the cross talk is pretty much full frequency, it lacks top end and low end a bit but is not overall thin (as in having a gradual high pass filter sound).
The mix signals leave the routing modules at "0" level (the same level as feeding the direct out transformer) and connect to the resistor mix buss. This is comprised of long sections of circuit board (one for each section of channel bucket) with a horizontal trace for each buss, that trace having a resistor connected for each buss on each channel, one end of said resistor poking out of the long aluminum case that encloses the entire "buss" circuit board. This resistor end is connected by un-shielded wire to the routing module connector.
The aluminum case has a cover that exposes the back side of the circuit board when removed.
These sections of "buss board" connect in series across the console, and each buss is connected to it's 1272 mix amp via a shielded wire

 

[fazer]

It been along time but the 8036 NEVE we had back in the 70's and 80s was very low crosstalk on the bus.  I would suspect that what sounds like a modification ( reed relays removed / direct outs from busses to patch bay) has something to do with crosstalk problems.  Also these have a technical ground to the chassis which is well documented in the neve manual. You also need to bond racks to the tech ground and float shields on 1 end of interface cables.  When there installed correct were talking -90 on signal to noise.  If you have multi ground paths your gonna have noise crosstalk, radio stations RF and a host of problems.

 

[ruffrecords]

The sound of the cross talk is pretty much full frequency, it lacks top end and low end a bit but is not overall thin (as in having a gradual high pass filter sound).
The mix signals leave the routing modules at "0" level (the same level as feeding the direct out transformer) and connect to the resistor mix buss. This is comprised of long sections of circuit board (one for each section of channel bucket) with a horizontal trace for each buss, that trace having a resistor connected for each buss on each channel, one end of said resistor poking out of the long aluminum case that encloses the entire "buss" circuit board. This resistor end is connected by un-shielded wire to the routing module connector.
The aluminum case has a cover that exposes the back side of the circuit board when removed.
These sections of "buss board" connect in series across the console, and each buss is connected to it's 1272 mix amp via a shielded wire

OK, that is the standard Neve passive mix bus technique. It provides extremely low crosstalk. I am with fazer on this one, I suspect the mods that have been done to it are the most likely cause. There's a lot more to building a mixer than just wiring modules together. Do you have the original Neve 2-wire diagrams showing the lead dressing to and from the mix bus amps? That could give you a clue.

Cheers

Ian

 

[nielsk]

I am glad to hear reports of the original performance being satisfactory in this regard. Unfortunately the switch and interconnect choices employed do not stand the test of time, the modifications done have turned this console into a reliable and useable one from on that was completely un-useable because of constant noise and signal interruption.
This particular console was designed to work specifically with a Studer A-80, and MTR signal flow, monitoring and metering has very little to no use for modern recording and production techniques (specific to the demands of the owner). The entire reed relay matrix deals with the monitoring and metering via the "jukebox" and has no bearing on the 16 buss mix.
Properly wiring the outputs of the 1272 mix amps direct to patch has no bearing on the issue, it only lowered noise and increased reliability.
Yes, the technical ground scheme using the aluminum frame to connect the copper ground buss bar to the ground lug is not optimal, but this is easily cured. Yes, the "tie every shield to ground" scheme is not optimal (and cost them a small fortune that could have been easily saved) but this can be dealt with by proper termination of connections to the console.
One only need look at an SSL to see how grounding, shielding and termination can be implemented in a robust and cleanly functional manor.
Absolutely nothing has been changed in the implementation of the mix busses from the output of the routing modules to the 1272 buss amp inputs.
I am confident that Ian is correct in the source impedance being the root of the issue, and the mods to the routing modules relating to this.
The routing modules were modified to remove the quad / stereo / mono switching (which cut in and out with a breath on it), pans, and by doubling the poles used on the switch banks.
Now the switching and operation is quiet and reliable, but the load presented by the pan pot is no longer on the output of the 208 op amp on it way to the routing switches.
So the real question is, is 5K the ultimate resistance to put there?

 

[ruffrecords]

Since the mix bus system is intact we should be able to rule out crosstalk occurring within it. In which case, the only way crosstalk can occur is is by a common impedance within the modules. To cause crosstalk, this would have to be effectively in series with the driving amplifier output impedance. I can think of three components making up this impedance. The first is the output impedance of the driving amp. Unlikely to have changed unless the bias is way out of kilter or the transistor gains have dropped drastically. Both seem unlikely to me. The second is the output coupling capacitor. This is a possible culprit. You need a low ESR type here. The third is the 0V wiring from the module. If I remember correctly, this was also bused in the aluminium extruded bus bar system. If these have been altered in any way it could explain your symptoms.

Cheers


Ian

 

[nielsk]

In the routing modules the 208 amps were re capped with Nichicon KA seies (I can't find ESR figures in their docs?)
The output from the amp caps are original (or replaced over time as required) and are the blue Phillips (?) ones, for the most part.
There is a 470uF feeding the routing switches and a 1000uF feeding the direct out transformer.
Would the ESR of the 1000uF affect the routing switches?
I can replace all the 470s and 1000s, before all the modules get loaded back in the frame.
Any suggestion on the ultimate caps to use?

 

[Joechris]

In the routing modules, did you take a measure between Audio 0V and Chassis ground? There should be no connection.

 

[ruffrecords]

I have no idea how ESR changes with age. These regular 1000uF caps have a quoted ESR or 0.199 ohms which is negligible in comparison to the amp output impedance. A pot divider from a 15K bus resistor and 0.2 ohms to 0V attenuated by 97dB so ESR at 0.2 ohms should be fine. They are not expensive caps so I would suggest replacing them.


Ooops!! Forgot to post the link to the caps:

http://uk.farnell.com/illinois-capacitor/108tta035m/capacitor-alum-elec-1000uf-35v/dp/9680861

Cheers

Ian

 

[nielsk]

Frame is back together in it's new home (still under construction!)
I will measure at the routing module connector between B- and audio ground, but it ties inside the routing module (clearly shown on the schematic, on the 1993 module pin E1 is tied to pin H1, and on the 1985 unit pin V1 is tied to pin A1)
All the B- runs from the edge connectors to a (maybe 1/4 " square) copper buss bar that runs the length of the console, and the buss bar is tied to the chassis.
All the shield lines tie to the chassis.
I need to fully explore the path the audio (referred to as "audio B-) grounds run.

 

[ruffrecords]

Frame is back together in it's new home (still under construction!)
I will measure at the routing module connector between B- and audio ground, but it ties inside the routing module (clearly shown on the schematic, on the 1993 module pin E1 is tied to pin H1, and on the 1985 unit pin V1 is tied to pin A1)
All the B- runs from the edge connectors to a (maybe 1/4 " square) copper buss bar that runs the length of the console, and the buss bar is tied to the chassis.
All the shield lines tie to the chassis.
I need to fully explore the path the audio (referred to as "audio B-) grounds run.

I think he said check between B- and CHASSIS which is not of course audio ground.

I don't think the B- bus bar should be tied to chassis at any point in the console except really close to the technical earth terminal. Check out this Neve missive:

Cheers

Ian

 

[fazer]

You might have already considered doing this.  Just thought I would share procedures that have worked for me in the past with perplexing problems.

Seeing the empty frame of the Neve made me think of a problem I have had with my API before.  I have had noise usually on the mix bus.  Even though I have the assign switches off and faders off, I still would get noise from an unknown module/source into the mix bus. 

Since the Neve is empty now, you might want to install the mix monitor module(s) for a monitor path and then check one channel installed and then check crosstalk as you plug in each module to see if something in the newly inserted module is causing the issue.  Noise problems can be an issue with faulty power supplies or low mains voltage also.  I know yours is a crosstalk problem but since the console is empty, you can check as you go for issues that may be from 1 faulty module as well. 

Just a thought.  Good luck!

 

[nielsk]

Thanks for all the helpful input everyone!
The noise floor on the console is very low, the mods I have done to it have considerably improved this area.
The cable connecting the copper B- buss bar to the frame is original, it lands on a frame web one over from the web the tech gnd lug is on, and a few feet away. I will replace it with one that goes to the gnd lug. I can see no way this can do anything but decrease any resistance between the 2 points.
Unfortunately, the console must be fully loaded to completely exhibit the cross talk, and every module has to be treated if any changes are made, but at least now (after 100s of hours of work) the modules are easily removed and replaced without the mere act of dong so causing problems!
BTW, just listen to the #1 rock record in the world right now to hear it, both recorded and mixed on this desk... part of the reason why it must be back in service very soon!

 

[picker999]

sorry for off topic. do you still have zetaIII remote?

 

[ruffrecords]

#1 rock album or single?


Cheers

Ian

 

[nielsk]

Album, hit #1 in the UK in October...
Can you enlighten me as to why the 10468 transformer is used on the input of the routing modules instead of the 31267?

 

[Joechris]

I once read that the "EMI" Neve desk did have 10486 wiered backwards as line in transformers, because they liked them "better".
How are they wiered in your console?