Mixer Resistors

[blown240]

How do you do the mixer resistors for an amp that has more than 2 channels?    I was talking to a guy that said you cannot have more than 2 channels without a way to switch them out of the circuit, because you cannot have more than 2 mixer resistors.

Is that true?   Seems to me that you could have as many as you want, as long as they are all rated the same and come together at the same point before the power stage.  But I have been WAAY wrong before!

[RicharD]

Summing can get messy.  As you stack em up, you're doubling the load each time so your overall volume is dropping.  At some point, the load impedance becomes too much.  Are you limited to only 2?  No.  the Fender PA135 sums 4 into 1 twice, once for the main path and once for the reverb path.

http://www.el34world.com/charts/Schematics/files/fender/pa135.pdf

[blown240]

OK, thats why he was saying that they had to be switched.   Couldnt you run 3 or 4 channels together with grid resistors and then into a Cathode Follower.   Seems like that would correct the load issue.

Something like this:

[kagliostro]

As I can remember the max number of resistors I've seen is 3

however you can read here to have some indication about mix resistors

Mr. Merlin say that probably 4 is the max number

http://www.freewebs.com/valvewizard/resistive.html

Kagliostro

[RicharD]

http://www.el34world.com/Forum/index.php?topic=11433.msg105936;topicseen#msg105936

PRR schooled me on this topic back in April.  I bread-boarded 1 "how not to" circuit and 1 almost "how to" circuit.  Read PRR's posts.  I did make his suggested tweaks but never documented the final circuit.  It's an 8 to 1 circuit.

[HotBluePlates]

Don't use a passive mix circuit, then.

If you're going to use a cathode follower, and accept zero gain (even a slight loss) from the tube, why not configure the tube stage as a virtual-earth mixer? Gain can still be zero (or even some portion less than full typical stage gain), and you can use more than 2-3 inputs.

Virtual Earth Mixer

Look hard at the schematic diagram. There is a 10k resistor running from output-side of the coupling cap to the grid. Each of the input resistors are 10k. Feedback from the plate to grid is negative, and sets gain as the ratio of plate:input resistors. With the plate resistor (after the coupling cap) equaling the input resistors, gain for each input is 1.

But you're not going to use a 6922 (6DJ8). For a 12AX7, you'll probably have a 1M feedback resistor, and the mixing resistors could be 1M (gain=1), or 220k (gain=~5) or 100k (gain=10) for a mild adjustment of gain between different source levels.

[blown240]

Thats pretty interesting.  thanks for the links!!

[RicharD]

Here are some more tube mixer circuits.

Altec 1567a
http://www.sotxampco.com/Schematics/Altec/1567a.pdf

Ampex Mx10
http://www.sotxampco.com/Schematics/Ampex/MX10.pdf

[PRR]

16-input passive mix networks are very common.

> into a Cathode Follower

No, you don't need high impedance. But you do need GAIN (with low noise). With 16 inputs the output is 1/15th of one input. You first need to boost-up weak signal by 1:15; or mix strong signals and have gain of 15 after the mix network to get back to a strong signal.

> Mr. Merlin say that probably 4 is the max number

When you have weak (pickup) signals and NO pre-mix boosting, as shown in his input-jack mixers.

You can build 2, 4, 16 preamps (tube-stage or two with gain control) and then mix them all together just fine.

Richard's "Altec mixer" with 5 inputs is a good model. Five 330K (near-enough Fender's 270K or 220K) with preamps on weak sources (just a pot on In5 which comes from Musak with strong level) into a simple gain stage.

Somewhere north of 8 or 16 the losses get large, or signal dips toward noise; also a 64-input console is physically very-very wide. Huge mixing jobs are sometimes broken into stages: eight 8-into-1 mixers, boosters, a final 8-into-1 mixer, booster.

Don't be too attracted to Merlin's "feedback mixer". It does reduce interaction between controls. Some mix enginers do not want the singer to fade when you boost the cow-bell. A badly proportioned passive mixer sure will do that. A well-proportioned passive mixer may dip slightly and perhaps "naturally" (less singer leaves more room for cowbell). The interaction should be less with more channels. A good feedback mixer, you can run 0VU in channel 1, then slide ch 2-64 to ANY position, and the ch 1 signal stays 0VU spot-on.