Would this work for simple tube summing?

[Adam Smith]

I'm putting together a 4ch set of NYD one bottles. I though it would be nice to have a simple summing circuit installed in the unit.  The idea would be to be able to use the obp's as single, unbalanced pres, or sum between 2 and 4 of them to a mono, fixed tube gain stage. No volume control on the sum circuit, that would be done with the individual obp's.

I looked at the schematic for the ampex mx-10 and that wasn't exactly what I needed, but it seemed simple enough. I also searched the forum on how a summing network is typically set up.

I came up with the idea to use the second stage of the MILA-1 as my summing stage.

So, my math for the resistors used in the summing network was as follows...
9k=(4-1)/(4+1)*15k

my math for the loss in dB is as follows...
log(1/4)*20=-12dB

I knocked up an example of what I'm thinking in schematic form. There are only 3 obp stages to save space. The circuit would be built for 4.

Would what I have here work at all? Or am I barking up the wrong tree totally? Would it be problematic to set up the summing network for four inputs but only use two of them?

Lastly, I'm figuring the output z of my summing network would be 2.25k, which is too low. I'm assuming I'm looking for about 15k to match the grid of the second stage, so I need an output resistor. I'm not sure how to calculate that. Is it (desired output z) - (network output z)? In my case 12.75k.

Any help on this would be great. Thanks!

 

[ruffrecords]

You are on the right track but I think you need to  change your bus resistor values. The input impedance of the summing stage is very high so you do not need to use resistors as low as 9K. Also you do not want to load the one bottle stages too much and 9K is probably a little on the low side. Also, the summing resistor needs to be large compared to the output impedance of the driving stage otherwise there could be a noticeable level change when you switch channels in and out. On the other hand, you do not want the resistors to be too high as they get noisy. Somewhere in between there is a good engineering compromise. I would try 47K to start with.

Also you need to have a resistor from the grid(s) of the summing amp to ground to maintain dc conditions when all the switches are deselected. A 1Meg ohm resistor will be OK.

Lastly you need to change the logic of the switches. A passive summing system expects each summing resistor to be fed from a low source impedance whether it is selected or deselected. The normal way to do this is the ground the bus resistor when it is deselected. So you need to turn your switches round so the wiper connects to the bus resistor and connects it wither to the one bottle stage output or to ground.

Cheers

Ian