Help with mixer bus switching

[lucidtone]

Hi list

I'm using a simple summing amp for my band, it mixes together a whole bunch of drum machines and synthesisers.  Drums go on one bus, synths on another.  It works great, one thing I'd like to improve tho is the switching between buses.  (Sometimes a synth or two join the drum bus for a song, as percussion instruments.)

Excuse my crappy schematic skills  =)

Each input is actually on a DP3T switch, with one pole unused and one throw unused, for future expansion. The third bus will eventually be used as an alternative bus to the synth bus, with different effects and compression.

Anyway, when you switch between buses, you get an almighty CLICK, which would be great to get rid of;  I think the solution will probably be a DC-blocking cap on each input, but I'd like to avoid that if it's possible basically because that's heaps of soldering  :

Any other suggestions for improvements would also be welcome.  I'm using OPA604 opamps, and I have some AD823's which I'll also try at some point, probably when I put the third bus in.  Power supply is pretty straightforward 7815/7915 design, and each opamp is locally decoupled with 33-ohm resistors and 10uF tantalum caps, cos that's what I happened to have lots of lying around when I built it.

Thanks for your thoughts!

 

[abbey road d enfer]

Anyway, when you switch between buses, you get an almighty CLICK, which would be great to get rid of;  I think the solution will probably be a DC-blocking cap on each input, but I'd like to avoid that if it's possible basically because that's heaps of soldering   :

First you have to assess the cause of the clicks.
There are two causes:
-If DC offset is the culprit, clicks should be heard even with no signal. Is that the case? In that case, you need to block DC at the virtual-earth mix node.
-If clicks happen because you're switching live signals, these appear only when there's active signal, not on silence. The only way to avoid those clicks is to either switch on silence, use a zero-crossing switch or use a Vactrol for progressive switching.

 

[lucidtone]

-If DC offset is the culprit, clicks should be heard even with no signal. Is that the case? In that case, you need to block DC at the virtual-earth mix node.

Yeah, it clicks with no signal.  I actually have some caps in there between the mix node and opamps already, they have jumpers across them so I can try that easily.  Will let you know.  =)

 

[JohnRoberts]

Click if audio signal is stopped or started suddenly.

Click if source added had DC content.

Click if sum amp has DC offset from changing DC gain to bus amp.

Cap coupling can mitigate DC long term but may still click transiently without caps on both ends of the switch.

If you have an unused pole on the the switch you could also forward reference the input grounds to the bus amp.

JR

 

[Samuel Groner]

If it's a make-before-brake switch it will momentarily short the two busses which might upset the summing amps.

Samuel

 

[Twenty Log]

If it's a make-before-brake switch it will momentarily short the two busses which might upset the summing amps.

TRUE!  I just went through the same possibility with the 7x7 routing matrix thread... PRR was most helpful in understanding what happens when 2 busses get (temporarily) shorted... speaker cone destruction...

http://www.groupdiy.com/index.php?topic=37741.0

Also another thing to watch out for with the "clicks" are floating electrolytic capacitors (if you have any) in-line with the audio...  The 1980 Audio Radio Handbook from National Semiconductor shows the use of pull down resistors that prevent the caps from floating up to a particular voltage only to be "popped" when switching the circuit back in...  Pull down to ground on the "floating" leg when not in use...  100k might work well...

EDIT: this may include the DC block capacitor inside the other external equipment if not anti-floated I suppose...

 

[lucidtone]

If it's a make-before-brake switch it will momentarily short the two busses which might upset the summing amps.

I didn't think of that... it's this guy:
http://www.mouser.com/catalog/specsheets/MS-100545.pdf
Given that it's a sliding lever type switch, I can't really tell if it's BBM or MBB.  The docs aren't helpful either.

If you have an unused pole on the the switch you could also forward reference the input grounds to the bus amp.

Do you mean using small value resistors to ground instead of directly tying to ground, like in the Forssell white paper?
http://www.forsselltech.com/downloads/design_discussions/summing_buss.pdf

The buses all ground to the same place unfortunately... it would be a LOT of rewiring to keep each bus ground separate.  Not impossible, just a PITA. 

 

[JohnRoberts]

If you are happy with the results don't worry about it.

I learned from designing consoles that ground is a concept not a voltage.

Forming a parallel ground bus with say 100 ohm resistors from every input ground to the master bus, with a 100 ohm resistor to the ground in the master section also will give a first order forward reference of input ground potential differences.

JR

 

[lucidtone]

If you are happy with the results don't worry about it.

Very happy with the mixer - hey it's pretty damn simple, not much can go wrong!  I've only had it running a short time but it *seems* to sound better than my Soundcraft.  Could be because I want it to sound better, though. 


Some good ideas thanks list    I'll do some measuring and hopefully figure out what's causing the problem.