Summing network and Forssell article

In this article about the mixing network:

http://www.forsselltech.com/summing buss.htm

Fred says ("resistor" is the summing resistor for each channel):
"With such a resistor in place for each channel, the impedance seen by any single channel is now the value of its summing resistor plus the parallel value of all the channels hanging on the buss.
If we use 1k ohm resistors for these resistors in a 10 channel mixer we would have a parallel source impedance of 100 ohms (1k/10= 100). This is the buss source impedance. Each channel will now see its summing resistor in series with the buss source impedance. That means each channel will see 1k (the summing resistor) plus 100 ohms (the source impedance of the buss) or 1.1k ohms as its load. Using larger value summing resistors will increase the load impedance seen by each channel."

To my eye, looking at the schem, a single channel should see its summing resistor in series with the parallel of all the other channel resistors, not all the channels in parallel. So the impedance seen by any channel should be 1k + (1k/9)...

Where's my mistake?

Does the output impedance of the other channels play a role or I can ignore it as its value is much lower than that of the summing resistor?

Many thanks,
Frank

nydave (whatever happened to him?) posted a lot of notes on this subject. try a search on his posts. he also had a mix network design diagram which i cant seem to find at the moment

Does the output impedance of the other channels play a role or I can ignore it as its value is much lower than that of the summing resistor?

Click to expand...

from what little i understand about this, i think you can ignore it if it is much lower (like 10X lower) than the value of the summing resistor. if its higher than that it will definitely have an effect although the effect is less as you add more channels.

> a single channel should see its summing resistor in series with the parallel of all the other channel resistors, not all the channels in parallel.

You are correct. Fred seems to have taken a simplification which is darn-close for 10 channels, but not exact.

And yeah, you can account for source impedances too, but life is mighty short and in most cases the error is too small to think about.

Many thanks guys,
this was the only thing I didn't understand in that great article and was really making me crazy... :grin:

Frank