Poor Man's Passive Mixer with Pan

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
16,158
Location
Norfolk - UK
As a result of several enquiries I have for some time I have been investigating low cost ways of adding a pan control to a passive mixer. The initial idea for reducing cost came from what I call the 'Neve trick' that we used for stereo monitor level controls back in the 70s. The basic idea is shown in the top schematic here:

NeveTricksequence.png


A single pot is wired across a balanced line. Its wiper becomes the hot output and the cold signal goes straight through. When the pot is fully off, although there is no signal between hot and cold, both of them get the full cold signal. This is effectively a common mode signal so how little of it reaches the final output depends on the common mode rejection ratio of the immediately following transformer. This allows a dual gang pot to be used to control the level in two balanced lines rather than the expensive quad pot that would normally be used. The fact that the fully off position is not quite fully off did not matter for monitors.

The middle schematic shows how you might apply this technique to a basic passive mixer. Not particularly useful but the bottom schematic shows how you could then use it to add a pan pot using only a dual pot instead of a quad pot.  I even hand built a prototype for a client to try out although circumstances have so far prevented him from doing so. Listening tests indicated that this actually works reasonably well.  I then received another request for much the same thing so this time I had some simple PCBs made as this client is handy with a soldering iron and was keen to build it himself. I built one of the PCBs and did some tests myself. Again listening tests indicated it worked but I did notice the all off position had some some residual signal. Barely audible but still there and this was using good quality Sowter 600:600 transformers so the CMRR should have been excellent. Also with the level pot fully open and the pan hard left, there was a similar residual in the right channel.

I then went round in circles for quite some time trying to measure the thing. Part of the problem is that there is 40dB loss to the bus so even if the residual was only 40dB below that it would be around -80dBu. So I fed +20dBu into it and found the residual was about -68dBu or only 48dB below the nominal signal. Probably OK for a pan pot but not good enough for a level control that is supposed to be turned off.

So I then looked for alternative balanced attenuators:

Balanced-Attenuators.png


The standard method is shown on the left and uses a pair of pots. At least we don't need a quad pot but at least one of my clients wants slider faders on some stereo inputs to his passive mixer and quad slider faders don't exist. The second schematic shows a single pot fully balanced fader so a dual slider pot would do a stereo fader. The compromise here is that there is a minimum insertion loss of 4.4dB but I can probably live with that.

It then occurred to e that you could probably use the second technique to make a balanced pan pot:

MK2-Passive-Mixer-with-Pan.png


To get a reasonable pan law from a linear pot you need to pad the input to the pot with resistor equal to the pot value. This means there is a loss of nearly 10dB when panned hard left or right. So there can easily be another 14dB loss in addition to the  40dB bus loss which possibly creates a noise issue - it certainly worsens it.

I need good attenuation from the level pot so it needs to be balanced. Four pots for stereo is not an option so the attenuator with a 4dB loss is my preferred route. I don't like the level drop cause by the balanced pan so the only remaining question is if the performance of the 'Neve trick' pan is good enough for a pan pot.

Thoughts, comments?

Cheers

Ian
 
Ian, I think R8 is in the wrong place on the Neve trick pan pot.

You could insert some extra loss on the pan pot adding a resistor from the cold input to the cold side of the pot, so you make a compromise at a mid point between CMRR and desired insertion loss, let's say you add for best performance at max attenuation a 3k3 resistor, you will be having only 2.4dB insertion loss, and pretty good CMRR, depending on the tolerance of the components, 3k2 would be the perfect match for 4k7//10k. You could use lower resistors for lower insertion loss but I think you really need the CMRR when you are fully off, and it isn't a big deal when fully on to have that 3k2 resistor inserted, not so perfectly balanced but you wanted that signal so you could consider that as part of the insertion loss. I didn't make the numbers of how that affect the pan law, maybe you want to take it into account if you want a good match for those 3dB.

Edit: at mid point the loss will be 6.7dB, so about 4.3dB law. The best values for 3dB law are 2k28 for R8 and 1k86 for the new resistor. Using 2k2 and 1k8 you have about 1.44dB insertion loss and 2.9993dB law. The balanced for the best case at full off pan is 1k803 against the 1k8 resistor. I know I used too many significant figures but just to show how good it takes with common resistor values.

Will you be using a balance pot? (or pan pot for the stereo channels)
If not maybe you can solve a way of doing the pan before the level mixer, I don't know. If that's an option let us know so we help you think on that.

JS
 
Yes, R8 is in the wrong position. Thanks for finding that. I will correct it.

I am not sure if the padding you suggests is meant to improve the balance at the mid point or at the hard panned positions. Need to think about that.

The stereo channels do not need a balance pot. They are relatively straight forward.

I like the idea of panning first followed  by the level pot but I am not sure if it helps. Again, further thought is required.

Many thanks for the input.

By the way, I thought of a way of making a quad slider fader the stereo inputs - just use a couple of dual tack sliders and place them side by side with a fit a fader bridge.

http://www.canford.co.uk/Products/21562/58-251_CANFORD-FADER-BRIDGE-30mm

Cheers

Ian
 
With the values I suggest before the balance is better when fully off (when hard panned to L, R is fully off and perfectly balanced) 1.4dB insertion loss at the max position (L when hard paned to L). And 4.4dB attenuation when at the mid point, giving 3dB difference between max position and the center.

Using the values in the attachment I found from input of the chanel to output of the pan -7.7dB and -10dB (pan max and center, level max). From the input of the channel to the bus -47dB and -49.8dB (same conditions). With the pan at min attenuations were insane, over 130dB at the output of the pan and over 170dB at the bus, but that doesn't mean anything with this simulation.

JS
 

Attachments

  • Passive mixer.pdf
    65 KB · Views: 150
Funny how just seeing a schematic makes it all clear! I think this is a very good idea. I will modify one of the existing PCBs and try it out. Thanks for the input.

Cheers

Ian
 
It usually does, on the first post was hard for me to make one, at that moment. Thanks, and you are welcome.

JS
 
http://www.ianbell.ukfsn.org/passivemixer/MK2-Passive-Mixer-with-Pan.png

For -60dB level, R1 must go to <0.6 Ohms wiper contact resistance and end-resistance.

Is that reasonable for a 1K pot?

Numbers I am seeing for good conducty-plastic pots are 1.5% and 4 Ohms.

1.5% is -36db and 4 Ohms on a 1K pot is -48dB.

You may be proving that large consoles (>16 in) "require" low-cost amplifiers and beg for VCAs.
 
ruffrecords said:
I think this is a very good idea. I will modify one of the existing PCBs and try it out. Thanks for the input.

Cheers

Ian

Ian, I apologize for bringing this up after so many years, but I've been researching balanced panning and wanted to know if you ever revisited this. I know NY Dave had also talked about a balanced panning scheme, but since he's largely absent I thought I'd try you first. Thank you for any light you can shed, and thank you so much for all the help you've provided in the past.
 
groselicain said:
Ian, I apologize for bringing this up after so many years, but I've been researching balanced panning and wanted to know if you ever revisited this. I know NY Dave had also talked about a balanced panning scheme, but since he's largely absent I thought I'd try you first. Thank you for any light you can shed, and thank you so much for all the help you've provided in the past.
The mods certainly improve the CMRR but make no improvement to the offness of the level pot. As PRR pointed out, this is almost certainly due to the end resistance of the pot.

NYD's passive mixer and pan used a dual 1K potat the input for the level control so this would have no better offness than my design.

Cheers

Ian
 
Back
Top