summing mixer with panner, short or not to short

[orangerec]

Hello,

I'm building a 24x2 mixer with the CAPI ACA cards. I've decided to incorporate a 3 way switch for the LCR panner.

https://www.mouser.com/ProductDetail/Alpha-Taiwan/SR2511F-0403-19R0B-E9-S-W-159?qs=8%252Br4Hz5Xir%2FmSW9A%252BgefJA%3D%3D

My question is should I use the Shorting (make before break) or Non-shorting (break before make) switches? Thank you in advance.

 

[boji]

My common sense answer would be for routing mono channels, MBB.
For summing say, postfade aux buses to mono, or dropping out the sides of a stereo bus, BBM.

 

[orangerec]

Thanks, boji. It's just input transformer (600ohm)>LCR switch>buss resistor>ACA. MBB it is. It's 24 channels. So, I'll be using 47k buss resistors as per CAPI's schematic. I'd like to incorporate a -4.25dB center(mono) pan law for Center position. 

 

[ruffrecords]

It does not matter which type it is for an LCR switch. BBM is more common though.

If you want to be sexy you could do it wit a single pole centre off toggle switch and a couple of relays. Switching would then be so fast it would not matter.

Cheers

Ian

 

[boji]

It does not matter which type it is for an LCR switch

In an extra-normal situation like switching on the fly, mid-mix, might MBB be less disruptive? 

 

[ruffrecords]

In an extra-normal situation like switching on the fly, mid-mix, might MBB be less disruptive?

Debatable a) if it would be and b) whether you ought to operating an LCR switch on the fly in a mix.

In one case the level momentarily increases then goes to the desired level, in the other case it momentarily drops and then returns to the correct level. I find it hard to decide which one is worse.

As I mentioned. BBM switches are much more common. The only place I have seen these present a potential 'noise' problem is in EQ circuits when you switch caps with different dc charges on them. MBB does not solve this problem. The normal solution is just to ensure the dc voltage is distributed to all the caps by adding 4M7 resistor between adjacent switch contacts.

Cheers

Ian

 

[boji]

Debatable...ought to operating an LCR switch on the fly in a mix.

Indeed, it was the only scenario I could imagine where connection type might matter.

Not to belabor the point, but if a MBB switch was "passively" routing towards resistors in front of VGSB rails, why wouldn't the switching noise be less or equal to BBM?

Edit: Guess we now have to worry about the point at which the bus resistors get lifted from ground rail as well.  On triple thought,  OP: K.I.S.S. and go with what is standard (what Ian said)  ;D

 

[abbey road d enfer]

Not to belabor the point, but if a MBB switch was "passively" routing towards resistors in front of VGSB rails, why wouldn't the switching noise be less or equal to BBM?

Switching noise is constituted of at least 3 elements:

• Noise gain/source noise change: these variations are typically inferior to the idle noise level and don't change the noise spectrum significantly. There are debates about how the summing resistors should be left floating  (and what end) or grounded. But I suspect it's not teh OP's primary concern.
• DC transient: As Ian mentioned, the cure is not in how the switching is done, it's in how the various switched nodes are made equipotential DC-wise
• program level change: this is inevitable, and there are two aspects of it. Transient click, due to the mere fact that the signal at the switched end suddenly goes from zero to something ([or [/i]vice-versa[/i]), and can be fixed only by switching at zero-crossing or ramping (which involves a variable transmittance switch, that can be a potentiometer or a VCA).
  Overall level change: there are several scenarii there. "Panning" from one side to the other is bound to produce different level changes depending on the switching arrangement. BBM with create punches between positions , whereas MBB will create humps.
  Actually, the variations will be dependant on the type of switch. In particular, rocker switches are not natural 3T. They are constituted of two mechanically-shifted sections.
  Rotary switches allow any type od switching, except direct shift from L to R.
  Push-buttonswitches allow that.
  Then I don't see why one would suddenly pan an active source in the mix with such a crude arrangement as a 3-way switch...
  

 

[boji]

Thank you for the detailed explanation Abby!

BBM with create punches between positions , whereas MBB will create humps.

humps to punches-- does gender play a role?  ;D ;D

 

[orangerec]

Thank for the info. I've ordered BBM switches. I love how much I learn on these forums/discussions. Yes, it's a basic LCR switch. I need that for the recall purposes. I'll be using 47kohm buss resistors for the L and R rails (recommended by CAPI for it's ACA) but what should I use for the center position to maintain a -3dB pan law? I'd actually prefer -4.25 or 4.5 pan laws. I believe that's what SSL and/or Neve used back in the day.

 

[boji]

https://www.electronics-tutorials.ws/attenuators/l-pad-attenuator.html

Someone correct me, but I think you'll want a drop of 2.125dB per side for a total center reduction of 4.25.

Edit: Also https://www.rfcafe.com/references/electronics-world/resistor-pads-electronics-world-july-1959.htm

 

[abbey road d enfer]

Thank for the info. I've ordered BBM switches. I love how much I learn on these forums/discussions. Yes, it's a basic LCR switch. I need that for the recall purposes. I'll be using 47kohm buss resistors for the L and R rails (recommended by CAPI for it's ACA) but what should I use for the center position to maintain a -3dB pan law? I'd actually prefer -4.25 or 4.5 pan laws. I believe that's what SSL and/or Neve used back in the day.

It depends very much on the topology used for panning. There are 3 most common topologies (see attached).
Which one are you going to use? If you ordered DP3T switches, I would suggest using #3.

 

[orangerec]

https://www.electronics-tutorials.ws/attenuators/l-pad-attenuator.html

Someone correct me, but I think you'll want a drop of 2.125dB per side for a total center reduction of 4.25.

Edit: Also https://www.rfcafe.com/references/electronics-world/resistor-pads-electronics-world-july-1959.htm

Thanks. Reading up on Lpads now.

 

[orangerec]

It depends very much on the topology used for panning. There are 3 most common topologies (see attached).
Which one are you going to use? If you ordered DP3T switches, I would suggest using #3.

Yes. #3 makes the most sense. I got unbalanced signal post input transformer feeding the LCR switch then to the 47kohm buss resistors. I just have to figure out the values the Center position to maintain a 4.25dB panning law.

 

[ruffrecords]

Yes. #3 makes the most sense. I got unbalanced signal post input transformer feeding the LCR switch then to the 47kohm buss resistors. I just have to figure out the values the Center position to maintain a 4.25dB panning law.

The sum of the two resistors  needs to be 4.25dB higher than 47K.

4.25dB is 1.63 times.

1.63 * 47K = 76K7.

76K7 - 47K =29K7

So I guess the additional resistors need to be 30K.

Cheers

Ian

 

[orangerec]

The sum of the two resistors  needs to be 4.25dB higher than 47K.

4.25dB is 1.63 times.

1.63 * 47K = 76K7.

76K7 - 47K =29K7

So I guess the additional resistors need to be 30K.

Cheers

Ian

Oh! I meant the panning law is -4.25 in the center position. Right? we want the center to be down because the two sides sum together.

 

[abbey road d enfer]

Oh! I meant the panning law is -4.25 in the center position. Right? we want the center to be down because the two sides sum together.

Yes, and that's exactly what Ian proposes.

 

[orangerec]

Yes, and that's exactly what Ian proposes.

Makes more sense now. I'm having some trouble figuring out how Ian came about 1.63 times. I just don't want to follow directions, I'm also trying to learn and understand. Thank you!

 

[abbey road d enfer]

Makes more sense now. I'm having some trouble figuring out how Ian came about 1.63 times. I just don't want to follow directions, I'm also trying to learn and understand. Thank you!

4.25dB computes to a ratio of 1:1.63
Standard logarithmic calculus.

 

[ruffrecords]

4.25dB computes to a ratio of 1:1.63
Standard logarithmic calculus.

If you wanted to know what 1.63 times is in dB then on your calculator you would do log(1.63) x 20

and the answer would be 4.25. Do do it the other way round you reverse the process so:

First you do 4.25 divided by 20 which gives 0.425. The on your calculator you do INV log( Ans) and get 1.63

On some calculators the INV button is labelled SHIFT.

Cheers

Ian