Opinions on passive buss loading arrangement

Hi guys-

Ive got a project Ive been going in circles on, wanted to get an opinion on a bussing arrangement.

Im building an 8 channel mixer using basically 8 api 312 cards. I have direct outs on each and then plan to use the unused tap on the output transformer to feed a passive mix buss using (hopefully) transformerless input 325's.

My primary application for the mixer portion is most commonly going to be used for summing 2 or 3 channels (while using the others direct out and muted off the mix buss) much more commonly than mixing 8 inputs to 2.

Ive been thinking alot about the "best" way to setup the mix buss to fit my specific needs. This is where I could use some feedback or suggestions if anyone has any.

Ive been planning on taking the feed from the output transformer to a slugged dual pan pot and running the output of that pot into a DPDT switch. Ive got the panpot feeding two poles on one end. Ive got the middle poles going to a small pcb with the bussing resistors and that feeds the L/R 325's. I had been planning on putting a slugging resistor on the outer poles for constant impedance on the mix buss. Im starting to wonder however, how critical or necessary this is and this is where I could use some feedback.

The way Im thinking, if Im only going to be summing 2 or 3 channels at any given time, wouldnt I be at an advantage to not have the entire mix buss loaded down so I could use less makeup gain from the output amps? I dont mind having to fiddle with the buss output gain as I add or subtract channels from the buss if it means better sounding output from the box. If I dont have constant impedance on the passive buss though, is that the only drawback that I am going to have?

If I dont maintain constant impedance, will this screw with the bussing resistor value Ive chosen to use? Is level the only thing that is subject to change on a passive buss that doesnt have constant impedance? Can frequency response be efffected as well?

If you were in my shoes, for the application Ive described above specifically, whats the best way to tackle this to get the best sounding performance off the buss. I dont want to bring up a bunch of noise if I dont have to, but Im not sure how not having constant impedance will savagely effect things.

This is the last you guys will hear of this, Ive been posting about this stupid project for like 18 months or so, Im finishing this thing up very soon, this is the last stumbling block I hope.

thanks in advance

dave

I'm not quite grasping your description, but I can visualize enough of it to tell that you're going to have problems if you allow the mix buss impedance to shift up and down. And that's even if you don't care about pops, clicks, crosstalk and having to vary your makeup gain.

So what's the problem? The panpots. The loading on the panpots is fairly critical, and the mix buss impedance forms part of that load. In a mixer with only a few channels, the mix buss impedance is fairly high and becomes only higher as you remove channels. The whole feel/action/law/whatever-ya-wanna-call-it of the panpots will change as you assign and unassign channels. So yes, you want to keep the mix buss impedance relatively constant, and that means substituting a suitable "dummy load" when a channel is switched off the buss.

Frequency response does play into as well, since your gain makeup amp will behave differently in that regard (especially if it's transformer-coupled) as the source impedance varies.

What panpot circuit are you using? What value of summing resistors? What are the source and load impedances you need to present to the amps?

Hi dave-

ok, that makes sense.

The pots I have are actually from an api console, so Im gonna try their configuration before I go and reinvent the wheel- its a 2.5k pot with 1K slugging resistors. The buckets have both 10K and 15K summing resistors, Im planning on having to figure those out on test though.

According to the API literature, the input inmedance of the 325 amp is 12K and the output impedance of the 312 is "less than 15 ohms". If Im understanding this correctly, all I need to do is insert a 15 ohm resistor to properly load the buss when I mute (remove) the 312 output to the mix buss?

dave

Did the buckets originally use active summing or passive? Any diagrams of the original config?

I'm trying to figure out what you need as far as summing resistor value.

api always used active summing so far as I know, Im pretty sure the value of the pot they used probably reflects that, but Im gonna try to make it work in a passive buss. I dont have a diagram but do have a pc board, the signal came into the pot and then went to the board terminating at a summing resistor, no idea what was happening down the line in the console. For a thing as small as Im building, 8 channels, Im hoping it wont be too much of a big deal and Im also really designing this for summing during tracking so it doesnt need to be entirely perfect the way I would want a playback mixer. In a way Im almost tempted to do L/C/R assign swithing instead of a pot, but have the pots so want to try to make it work.

dave

...

The sort of arrangement John describes makes a lot of sense for a more complicated mixer, with routing, auxes and variable and sometimes fairly high source impedances. The approach of forcing the buss impedance low solves a lot of potential crosstalk problems and makes interfacing to the gain makeup stage easier. But in my opinion, for the kind of very simple mixer Dave wants to build, with source impedances being rather low to begin with--and with no need to interface the mix buss directly to the outside world--it seems we have the luxury of taking the "minimum loss" approach instead. See below:



I took the panpot values as a given since you expressed a desire to stick with what API used. 15K mixing resistors work out well since 15K in series with the buss impedance provides a proper load for the "slugged" pot to give about 3dB (3.5dB in this case) attenuation at center. The output of each pan pot actually sees its 15K mixing resistor in series with 15K/7 (2.14K) in parallel with the input Z (12K) of the API 325 amp.

The impedance looking back into the panpot is so low (about 400 ohms) compared to the value of the mixing resistors, that muting by simply grounding the input side of the resistors will not cause a level/impedance shift worth worrying about in the mixing buss. I put the mute after the panpot so that the load reflected to the API 312 would not change much when the channel was muted.

EDIT: The "cw" legends were at the wrong end of the pot. I fixed the drawing this morning.

thanks for the input john, Im trying to see how little I can get away with but will work think about your post carefully.

Dave, thats basicallly the drawing Ive scratched out, my concern was with grounding the end of the mute switch instead of putting a load resistor there. Im gonna try that first (grounding) and see if everything doesnt get too wacky. thanks for all the input!

dave

Ha... That's funny. I guess NOW I can picture what you were getting at :wink: Well, consider your plan now verified by Mr. Calculator.

Anyway, grounding at the mutes should work just fine. The output impedance of the pan pot varies between zero and about 400 ohms. That's miniscule compared to the 15K resistance in series between the panpot and the mixing buss.

...

cool john I'll ring you.

looking more closely at this API pot, Im at a loss once again. I took apart a neve mixer earlier this year and the pan pot is exactly like dave's drawing, this api pot is different. Before I redo it, Im hoping someone can help me to understand what they were doing because its not cllicking just yet. Im gonna try my best to describe how the pot is wired.

Its obviously a dual deck pot with 3 terminals per deck, standard 2.5K dual pot. We'll say A B C are posts on the outer deck and D E F are posts on the inner deck closest to the shaft with posts A and D on the counter clockwise rotation, C and F on the clockwise and B and E are the wipers.

outer deck: A B C
inner deck: D E F
|
|
|
shaft

hopefull that makes sense.

Unfortunately I just have the pot and a piece of buss rail it was hooked up to, so I dont have a ton to go on, but here's how its wired-

A has a jumper to F

C has a jumper to D

1K resistor between B and D
1K resistor between C and E

Hookup wire from harness to A, B, D, E. The wire from D definitely goes to the ground buss, this is what has me thrown off. I think that A might be the input judging from the bits I have here.

Short of poking around and trying to make this work, does wiring a pan pot like this make any sense at all? Im not getting how this works split between the two decks? Im assuming its not just a fancy interpretation of the diagram above. Im gonna see if I can make it work as is and surely its easy enough to modify into the above schem but would like to try to understand how this works as a panpot as I just dont get it.

also, as if Im not full of quesitons, anyone see any problem with using a shared ground buss for both the pot and dpdt switch in dave's diagram?

thanks

dave

Dave,

The panpot wiring makes sense if D/C is input, A/F is ground, B is left output and E is right output. The jumpers and resistors between decks are just a matter of convenience in wiring. If you draw out the circuit, you'll see it matches my diagram.

The mutes should most definitely short to the same ground buss as the pot. I'll draw up a diagram later of how I recommend wiring the thing if I have some free time.

Hold the phone... Those panpots, are they dual linear, or some other taper? That's VERY important to this discussion...