Sum bus theory and practice

[JR.]

Yes the derived filter will recombine better if levels of the two band passes are the same, but by definition they won't be. Probably better than just throwing random filters at it,

The apparent shortcoming of the derived filter, is a not very steep slope in one direction, but in general I would avoid high order filters as that is just more opportunity for combining problems due to phase shift.

If you think about it a state-variable filter, like you would use for a parametric EQ section gives you a 3 way derived dividing filter, as LP/BP/HP all sum back to unity (albeit BP is opposite polarity from LP and HP). The SVF also gives better slope symmetry between the bandpasses.

Note: I am not a huge fan of this, perhaps because I recall the old days when NY area FM radio stations were using crap 12db/oct loudspeaker speaker crossovers in front of multiple limiters to hype loudness. You'd have to buy the record to find out what it really sounded like.

JR

[Gold]

Paul - Use a "derived" filter so that they re-combine properly. The old NSC audio handbook has a discussion on that in active speaker crossovers. We should probably start a thread on that.

Sure let's start a thread. I'm only at the end of the first chapter of the Lancaster book. I got stuck. I promised myself I wouldn't move on until I understood everything.

I know Maselec ended up using a single buffered capacitor. Probably tightly hand selected. I doubt they sell more than 10-20 units a year. It sounds wonderful but I want to learn about active filters and I want to make a filter set for the console I'm building. I figured filters were a good place to start before learning more about EQ topology.

[mediatechnology]

For now, jump here: viewtopic.php?f=6&t=26&p=101

And here: http://www.ka-electronics.com/Images/pd ... pdf#page=4

JR - What browser do you use? Do the pdf page arguments work in Safari?

[JR.]

I'm using Safari, and no it doesn't jump to page 4. It just opens to page 1.

Funny to see that old paper written by Larry Blakely.. He was a marketing type and OK technical writer, but not an engineer...(I guess a recording engineer, but not design engineer).

They say the BP of SVF is only 6dB/oct while HP/LP are 12dB... I would need to think about that some more.. but not now..

Last I checked Larry was still living about 25 miles from me...

JR

[Igor Dockx]

Dear John,

Would you comment on how you'd go about implementing the current summing principle in a conventional mixing desk with mono input channels, and a pan circuit between the fader and the current summer?

So actually my question is, what would a "current panner" ideally look like, if it were to be placed after the channel VCA and before the current sum bus, without compromising the advantages of current summing?

Thanks in advance!

Igor

[JR.]

Dear John,

Would you comment on how you'd go about implementing the current summing principle in a conventional mixing desk with mono input channels, and a pan circuit between the fader and the current summer?

So actually my question is, what would a "current panner" ideally look like, if it were to be placed after the channel VCA and before the current sum bus, without compromising the advantages of current summing?

Thanks in advance!

Igor

In theory you could place a pan pot between the two sum bus inputs and send the current to the wiper. This would split the current between the two buses proportionately based on wiper setting.... BUT,, this pan pot would now increase the noise gain of the sum amps, defeating the benefit of using the current sources.

In another thread here somewhere I have discussed the possibility of using two VCAs, one for each bus, and doing the pan and fader function with the control voltage. This gives the additional advantage of low bus noise due to low noise gain, but as signal is panned or fader turned down, the VCA noise is also turned down...

JR

[Igor Dockx]

Hi John,

Thanks for your reaction!

Here's another question; I hope it's okay to post it in this thread since it has nothing to do with current summing.

I have this Amek Channel in a Box; it has a trafo-coupled line input that can be converted into a virtual earth mixer by bypassing the 5K1 resistors in front of the trafo (see image).

My questions: What is the consequence of having a trafo in front of a virtual earth summing stage, other than galvanic separation? I have never seen a practical example of this before.
And second, if I were to run in virtual earth mode and I want to turn this into a 4-input summer, do the four inputs need to have a 20K4 resistance in series with the input signal to make a total parallel of 5K, for optimal source impedance?

Thanks for sharing your knowledge!

Regards,

Igor

[JR.]

I think I have seen some discussion of that approach elsewhere... I messed around a little with putting a transformer inside a closed feedback loop to get some cheap gain, but using cheap transformers it wasn't very stable. Not using cheap transformers for me was a "why bother then?".

Re your question, yes trafo give ground isolation, but IIRRC those were special transformers, and then they are operating in a modified current (?) mode.

An intellectual curiosity for somebody perhaps, I am not that curious or intellectual.

JR

PS: My email/web MAC mini took a dump... so I was lucky to find this place again. Computers suck, but no computer is worse.