Studio Cue System Project

[Voyager10]

OK - if it is a cable capacitance problem, here's a sketch of the active buffer idea:


The new bits are C1, R3, R4 and Q1 in the cue box, and R5, C2 and R6 in the mainframe box. (V2 is just the power supply, not a battery - that's how LTSpice rolls!).

R2 is simulating the other 4 47K mixing resistors for that output. Note that it's no longer a virtual-ground mixer. R6 sets the make-up gain so that it's roughly unity from V1 to the output, but if you don't need that much overall gain it will further benefit the crosstalk by increasing its value.


Obviously this is just a sketch, and I've no idea whether it will solve the problem.

 

[Voyager10]

I did some more thinking. CAT5 cable seems to be about 50pF/m (between any 2 conductors? not sure) so a 5m length will be of the order of 250pF. Here's a Spice simulation:



and the results are pretty grim:



in other words, at 1KHz the bleed at the op-amp output is -32dB of the signal being fed down the cable.


I've also realised that the wanted signal level is going to be ~ -10dB relative to what went down the cable (at full volume, and hard pan to one side), from the 47K / 15K resistor ratio, so the actual crosstalk will be -22dB at 1KHz.

(This means R6 in the previous post can be 8.2K to keep the same overall gain).

 

[Voyager10]

With a buffer, here's what the crosstalk sim looks like.

Schematic:



Sim results:


So in a like-for-like comparison the buffer makes it 50dB or so better.

(I'll stress I'm just playing around with simulations here, the real world may therefore be wrong).

 

[Potato Cakes]

Thanks for taking the time to run this in Spice. I am using shielded CAT6A cable so I don't know how much that changes the capacitance calculation.

Also, won't the 8k2 resistor in front of the NE5532s used for makeup gain and balancing reduce the gain that is needed for the insertion loss of the cue mixer?

EDIT: CAT6A is about 50nF/km so pretty much the same as CAT5

Thanks!

Paul

 

[Potato Cakes]

Wouldn't V3 in the cue box just be the ground connection since that is more than likely the source of the unwanted signal?

Thanks!

Paul

 

[Voyager10]

Re R6 - in the existing circuit the gain (from pan-pot wiper to NE5532 output) is set by the ratio of the 47K and 15K resistors, giving about 0.32 or -10dB.

In the buffered circuit the gain from pan-pot wiper to Q1 emitter is a little under 0.2, so it needs some extra gain from the NE5532 - about 1.7x IIRC - to get back to what you had. This is set by the 8.2K & 15K ratio.

(You will need some resistor on the input for an inverting op-amp amplifier - in the unbuffered circuit the 47K resistors were doing the job).

 

[Potato Cakes]

Gotcha. Also the 47k bus resistors were intended to be changed after I did some more testing. 47k / 6 = 7k83 which I think is too high of an impedance for the NE5532 for a 1:10 input impedance ratio. I'll have to double check that.

 

[JimJhn]

Poor planning is the cause.
I always have a basic usable stereo feed that is used as a starting point. Knob-fearing musicians can control the volume, if nothing else.
Some musicians are cleverer, those that have a personal studio, and manage quite well the system (behringer P16).
Good labelling of the sources is paramount.
For technically-challenged ones, I do it for them, not too different than the old way when cue sends were managed from the aux sends or an auxiliary mixer. More convenient actually, because I can check the mix on headphones standing by the musician.

An auxiiary mixer has the big advantage of zero-latency. I know many studios where the auxiliary mixer is in the control room. Not the most ergonomic arrangement IMO.
Useful only when the console has not enough Aux.

I'm completely with you about good-labeling.

And I forgot to even mention the zero-latency thing, which is I think why the aforementioned studio did it the small analog console thing to begin with.

The zero-latency software mixers are beyond confusing to me, and just primed for more confusion.

I will say that one of the studios I use has the (Aviom I think?) but uses them for multi-day rock tracking sessions.

For jazz dates they don't put personal mixers out in the tracking room, just simple headphone volume boxes.

Again it's not that people are too ignorant to use them, it's just to nullify any additional potential confusion.

 

[Potato Cakes]

I think I may have found what I am after, I just haven't had a chance to try it yet. The way the mixer currently wired is the wiper feeds the pan then the bus. But the incoming signal is never shorted to ground, so whatever noise there is on the ground potential is getting into the NE5532s as the + input is connected to ground. It may not completely kill the signal but it should get me closer to what I am trying to accomplish. I will try this sometime next week.

Thanks!

Paul

 

[abbey road d enfer]

I think I may have found what I am after, I just haven't had a chance to try it yet. The way the mixer currently wired is the wiper feeds the pan then the bus. But the incoming signal is never shorted to ground, so whatever noise there is on the ground potential is getting into the NE5532s as the + input is connected to ground. It may not completely kill the signal but it should get me closer to what I am trying to accomplish. I will try this sometime next week.

Thanks!

Paul

So you want to short the incoming signal. Think it over.
Your reasoning is utterly wrong.