Pan circuit for passive mixer

[Fablab]

Hi everyone. After a few years of inactivity a friend convinced me to help him to build a passive mixer with one fader and pan control per channel.
The inputs are unbalanced
I want to realize the pan circuit with a double pot and resistors and I would like to know if anyone knows of a pcb to make it
A thousand thanks

 

[JohnRoberts]

No not aware of any easy solutions. It will be difficult to pan passively.

Perhaps consider hard panning with a 3 way switch L/R/both

JR

 

[ruffrecords]

A lot depends on the source impedance and drive capability of whatever is connected to the inputs. This determines the values of the pots you would use.

Cheers

Ian

 

[Fablab]

I think to use the attached schematic (like in API mixers) with a lower value of the two output resistors (not 47k Ohm but 7k5 or 10k Ohm).
I want to put this pan circuit after a 10k ohm potentiometer for the level control.
Consider that the mixer should have 18 unbalanced inputs.

 

[ruffrecords]

As that schematic states, the circuit is intended to be driven by and op amp or other low impedance source. This means unfortunately you cannot tack it onto a 10K level control potentiometer. Also the bus feed resistors (47K) are an integral part of the design. You cannot just change their values without affecting the operation of the pan.

With a 10K level control you will need to use something like dual 50K LIN pot for the pan with 22K slug resistor from the top to the wiper and 47K bus resistors. With 18 inputs the bus loss will be 18 times (25dB) plus the 3dB loss in the pan pot so your output level will be at least 28 dB lower than the input. The output impedance will be approximately 47K/18 = 2.6K.

Cheers

Ian

 

[Fablab]

Hello Ian, considering I have some very good 10k Ohm Allen Bradley pots (single log. and dual lin.) and I'd like to use them, do you think there might be a good compromise to use them?

If not possible, keeping the 10k ohm pots for the pan, do you think 5k ohm pots are sufficient for the level control?

Meanwhile I found this simpler pan schematic from NEVE with only two resistors

 

[ruffrecords]

What you need to bear in mind is the worst case load you will present to whatever is driving your inputs. Ignoring the bus feed resistors and any effects on the law of the controls, if you have a 10K input pot feeding a pair of 10K pots slugged with 2K2 then the worst case input impedance will be in the region of 1.25K. Many sources will have difficulty driving this. Unfortunately changing the input pot to 5K only makes this worse. A 10K level control followed by 50K LIN pan pots slugged with 22K will have a worst case load of about 5K.

Bottom line, if your sources can drive 1.25K then you are OK with your 10K pots.

Cheers

Ian

 

[JohnRoberts]

Sorry, if it was possible to do this well passively, we could find examples to copy.

JR

 

[Fablab]

I am sorry Ian, but because of my limitations with English I'm not sure I understand.

Considering that the input signals are the outputs from DA converters or other hardware with an impedance not greater than 1k Ohm, is it possible to use both 10k Ohm pots for level control and pan?

Eventually is it possible, in the configuration above which pan circuit do you suggest to me between the API and the NEVE one?

Thank you
Fab

 

[L´Andratté]

Sorry, if it was possible to do this well passively, we could find examples to copy.

JR

+1
It would be so easy and simple with active circuitry, also suitable with your pot values and totally without any drawbacks vs. passive solution, which only *seems to be* easy. If you still want to do it, check out the old threads here, for NYDave´s passive summing mixer and many others and try to understand the impedance/loading issues!

 

[Newmarket]

I'd recommend reading up on panning in Douglas Self "Small Signal Audio Design". I daresay there are people here with the knowledge and experience. Just nice and clear to have the issues and possible solutions laid out clearly.
Imo first stage is to define what your ideal planning law / loss is. It's one of those topics that seems simple at first glance. Then it isn't

 

[JohnRoberts]

You need to decide what is important to you... If doing it passively is paramount I suggest something simple like the 3-way pan switch, another simple alternative is using two faders per input each sending to either L or R bus. Awkward but "Passive".
=====
This is why pretty much all professional mixers use active circuitry. If you give up and decide to use active devices there are many examples of effective pan circuits. Perhaps do a search right here....

JR

 

[Newmarket]

Gotta ask. Why the objective to do it passively ?
Beyond that - listen to JR. I know he's North American but he knows this stuff. fwiw the Peavey desks from his era did face a degree of unfair negative bias.

 

[Fablab]

I know that using op amps or without pan control everything would be much easier, but the friend I have to help asked me specifically to make a completely passive mixer.
I will try to convince him

 

[ruffrecords]

I am sorry Ian, but because of my limitations with English I'm not sure I understand.

Considering that the input signals are the outputs from DA converters or other hardware with an impedance not greater than 1k Ohm, is it possible to use both 10k Ohm pots for level control and pan?

Eventually is it possible, in the configuration above which pan circuit do you suggest to me between the API and the NEVE one?

Thank you
Fab

No problem. Output impedance and drive capability are two different things. The main thing is the minimum load the D/A will drive. Unfortunately just about all interfaces do not specify this. A typical example is the Tascam US-16x8. This has 8 line outputs. The spec for them says the output impedance is 100 ohms, the nominal output level is +4dBu and the maximum output level is +24dBu. What it fails to mention the minimum load impedance it will work with. The assumption usually is that everyone has 10K ohm input impedances so the load is 10K. I would probably work OK into 5K but if the load is 1.25K it is impossible to say but the answer is likely to be probably not.

If you want to make a passive mixer that is compatible with a typical D/A then you have to ensure its input impedance is at least close what it expects to see and in most cases this is 10K ohms. You can build your passive mixer with 10K level controls and dual 50K pan pots and it will have an input impedance of around 5K so there is a very good chance it will work with most D/As. I use this all the time for AUX return inputs of my tube mixer designs. But you cannot use 10K level controls and 10K pan pots.

Cheers

Ian

 

[JohnRoberts]

I know that using op amps or without pan control everything would be much easier, but the friend I have to help asked me specifically to make a completely passive mixer.
I will try to convince him

The customer is always right, so try to understand what his motivation is.

There may be some other way to resolve the conflict.

JR

 

[Matt Syson]

attempting a pasive mixer with summing has a whole raft of problems built into it especially if you have restricted the range of pots you have available (because you had some). Working from known drive circuitry (as Ian comments upon) is a start but then again signal losses will be significant (plenty of unnecessary noise due to the around 30dB of gain you will need and mix resistor values required so you don't overload your input gear and the likelihood of a terrible fader 'law' all in the name of some (likely internet induced) random thoughts about signal purity. Then there is the problem of fader level interaction (moving faders up and down with no signal present on them will alter the levels of other channels that do have signal as the bus impedance will change as level and pan pots are altered.
Audix mixers 'standardised' at a bus level of about -35dBu and buffered signals were fed on through 15K resistors which gave decent performance particularly for the time. A transformer was often used which had around 10db step up which helped with noise figures. Tjhe fader wiper had an amplifier on it which fed the pan pot then two post pan amplifiers with IIRC 3db gain.

 

[emrr]

New York Dave laid out some plans here years ago, find those. I think ruffrecords did too. There are good financial reasons why pan wasn't seen in the antique passive mixing days, doing it right to avoid having levels swim around with changing conditions meant using panning 'T' or ladder attenuators, very expensive and there's not a single ancient custom mixer I've ever seen that had them on every channel, usually just a few that could be patched in as needed.

 

[pucho812]

I have the New York Dave plans for a 16 x 2 passive mixer with pan and volume per channel.
I think Dave pulled all that stuff so if by chance you didn’t get it, it’s gone.

I hope he is well I have not heard from him in a long time.

 

[ruffrecords]

attempting a pasive mixer with summing has a whole raft of problems built into it especially if you have restricted the range of pots you have available (because you had some). Working from known drive circuitry (as Ian comments upon) is a start but then again signal losses will be significant (plenty of unnecessary noise due to the around 30dB of gain you will need and mix resistor values required so you don't overload your input gear and the likelihood of a terrible fader 'law' all in the name of some (likely internet induced) random thoughts about signal purity. Then there is the problem of fader level interaction (moving faders up and down with no signal present on them will alter the levels of other channels that do have signal as the bus impedance will change as level and pan pots are altered.
Audix mixers 'standardised' at a bus level of about -35dBu and buffered signals were fed on through 15K resistors which gave decent performance particularly for the time. A transformer was often used which had around 10db step up which helped with noise figures. Tjhe fader wiper had an amplifier on it which fed the pan pot then two post pan amplifiers with IIRC 3db gain.

I just want to correct one common misunderstanding about passive mixing. It is no noisier than virtual earth mixing. Yes you might need 30dB of gain to restore the level but the same system using a virtual earth mixer will have a noise gain of 30dB. Of course, VE mixing has other advantages and some disadvantages.

Cheers

ian