Unbalanced 5 to 1 passive mixer questions

[InBetween]

Hello All,

  I know I know... search function. Well I feel like I've done about a thousand hours of search function on this and various other forums, but I still can't seem to find the answers I'm looking for. I'll start by explaining the project a bit:

  -Unbalanced, 5 channel into 1, PASSIVE mixer, just level pots for each channel, no switches or pan
  -Will be feeding a LINE amplifier, not a MIC amplifier (most likely will be seeing 10k input load)
  -All inputs guaranteed very low impedance (~100ohm or less)
  -Channel interaction is less important to me than overall S/N loss and distortion
 
  Right now I have built a dinky little prototype 2 to 1 version, but I am finding some problems and can't quite seem to figure the solutions out on my own. The circuit I'm using right now has each input (1/4" TS) going tip to terminal 3 of a 10k pot, and sleeve to common ground. The pots both have 5k resistors on terminal 2, and terminal 1 goes to ground. All grounds are connected.

NOTE: I realize that using 10k pots and 5k mix resistors doesn't make sense, but its what I had laying around at the time.

  Now for my issues / questions!

  -There seems to be some overall high end loss, which I'm guessing is due to improper impedance bridging from the output to the input of the line amp... the problem is I've seen so many equations for figuring the output impedance of the passive mixer that I can't quite seem to straighten things out... So, I need to know precisely the steps taken to determine output impedance of a circuit like this.

  -I intend to try and get away with 1k pots and 5k mix resistors in the end. I realize that's a very low number for the pots, but I'm assuming since they will be receiving guaranteed very low impedance signals, it will be okay (~10:1 rule?). From what I gather, having the mix resistors several times the pot value will help to keep the overall output impedance more stable. Is this correct? Am I crazy? I'm open to any suggestions.

  -I was thinking originally that I would not need a shunt resistor as proposed in NYdave and others' designs, due to the fact that I am not feeding a mic pre. However, it seems that due to impedance loading issues, it may be best to include one. Again, the output will be seeing a 10k line amp input impedance, so I'm thinking overall output Z should be less than 1k. Any additional information on this, and how to calculate the optimal shunt resistor value would be extremely helpful. Keep in mind this is unbalanced, so the resistor values will not be doubled.

  -On that note, how would I calculate resistance assuming pots in-line with single resistors... the whole subject has confused me as I'm no electronics expert. I keep thinking I have found the answer, and then I'll test it with some math or a multimeter and find myself completely confused.


Basically... I know this topic has been beaten to death, but due to the fact that my design goals are quite different from the average 16:2 balanced summing mixer blah blah blah, I figured it may be reasonable for me to seek guidance.

Many thanks in advance!

-InBetween

 

[ruffrecords]

If you imagine a 1K pot connected to a source with zero ohms output impedance and the pot set so that it is 500 ohms from wiper to top and from wiper to bottom - because the source impedance is zero then the largest impedance the pot can appear to be is 500 ohms in parallel with 500 ohms or 250 ohms. If you do the sums you will find that at any other wiper position the pot will appear as less than 500 ohms. If the pot is fully clockwise or fully anti-clockwise, the pot will look like zero ohms. As a rule, when fed from a (near) zero source impedance, a pot never looks like more than a quarter of its value to following circuit. If the source is non zero you just add the source impedance to the pot value and divide by four. So if the sources are guaranteed to be 100 ohms of less then the impedance the pot will appear as is no more than 1100/4 = 275 ohms.

Now take five pots like this with 5K series wiper resistances. One end of each 5K is connected to at most 275 ohms due to the pot and source. The other ends are the bus. So each resistor and pot combination looks like 5K + 275 ohms = 5.275K across the bus. You have five of these in parallel so the bus impedance is never more than 5.275K/5 = 1.055K. This should be fine for driving a 10K bridging input. If all the pots were at 100% or 0% then the bus impedance would drop to 1K. This represents a worst case change in bus level of less than 0.5dB.

The bus has a loss equal to 1/N where N is the number of channels, in your case 5 or 14dB of loss so your line amp will need to be able to provide this much gain.

You need to remember that a 100 ohm output impedance is no guarantee that the source can drive 1K to a useful level but if you sources are specified to drive a 600 ohm load then you should be OK.

At these low impedances there is no reason for any HF loss unless you are using incredible lengths of very high capacitance cable.

Hope that helps.

Cheers

Ian

 

[InBetween]

Ian,

  Thank you so kindly for your reply. Your answers solved several apparent mysteries for me. My sources (tape machines) are spec'd to drive a minimum load of 600 ohms so I should be alright in that regard, and the line amp will be able to provide at least 20 dB of useable gain, so that should be fine as well. Your explanation of pot resistance as seen by the rest of the circuit was exactly what I needed to know. I'll update with the build results when completed. Hopefully this topic will be useful for anybody else out there taking a different approach to the passive mixing method.

Thanks again!

-InBetween

 

[ruffrecords]

Glad I could help. There are lots of myths surrounding passive mixing which I am happy to dispel.

Cheers

Ian