A mixer with headphone out/ can someone check it'll work?

[sean k]

Straightforward I suppose but will it work? I'll only be using one side of the headphone so I can monitor what I'm sending out while the other monitors the room. Main question is do the + inputs need a resistor between them and the 4.5V?... and is there anywhere I've been overly excessive with caps and resistors where they're not really needed.

Thanks in advance for any replies 

[sean k]

Basically I have a performance next Monday night and I want to run four instruments, a theremin, a synth stick and two stringed instruments which will all run through there own effects  so I need this baby to be able to do that. Just a little verification.

The headphone drivers are needed so much but it'd be good if it works.

Sorry if its kinda uninteresting 

[frequencycentral]

Hi Sean, thought this one might interest you: http://fa.utfs.org/diy/roland100m/schematicsshow12.html

[slacker]

Looks good to me Sean. I think you could get rid of the 0.1u caps between the mixers and adder stages and you don't need the 10k input resistor on 2a. You can just connect the 4 10ks from the mixers straight to the - input.

[sean k]

Thankyou Ian and Rick,

I bought a bunch of TLO84's today, which I've never used before ( only TLO7_'s), so I'll get going on the PCB tonight. I think I'll go for another input channel and also have a separate balanced in to monitor the desk and have it drive the other headphone channel.

So does the headphone amp look fine?, I copied it straight from a book except it was set up for +/- rails so I've put Vb on the + input.

Should I put a 330 ohm resistor between the 1k pot and the 1KB pot and the + input so it won't go below unity?

Sean.

P.S. heres a link to the poster I drew for the event.
http://s4.photobucket.com/albums/y115/quickkiwi/?action=view&current=DrK.jpg

[George Giblet]

- You don't need the 1kA at all on the headphone stage, in fact it will only cause trouble.  Those opamps are running as non-inverting stages so you can't put a feedback resistor in to control the gain like you have done with the input stages.

- A important one is to put say a 470uF electrolytic cap at the output of the headphone amplifier.   This is a single supply circuit.  The opamp outputs will sit at Vb with no signal.  You don't want a Vb level DC signal going through the headphones all the time!  The cap prevents this.

- Another important one is the phase shifter stage has limited output swing.  Since the headphone power stage is unit gain the phase-shifter swing limitation will translate to poor swing at the headphone amplifier output.    You really need to make the headphone stage have a gain of at least 3.

-  There's many ways to jig the headphone stage with parallel opamps.  Some result in an effective series impedance out the output some not.  This one is an example of how to get some gain out of the output stage.  Note there is one opamp wired sort of normal with the exeption that it has series resistance added the the output.  The other opamps act as current boosters only and they feed of the master opamps output (in the example there is one one booster but you can have more).  Don't worry about value of the series resistance you 330R's should be OK.  As shown the gain is much more than 3.

http://www.rane.com/pdf/hc4ssch.pdf

(the example circuit is dual supply, yours is single supply so will need to add the 470uF cap between the current boosted output and the headphones.)

- With a single supply consider changing the top 10k resistor of your VB network to 8k2 for better swing.

- Are you sure you want 10k input impedance?   You won't be able to feed a guitar into that without affecting the tone.

[PerroGrande]

The 10K input impedance was something I noticed immediately, too.  This is far too low if you're going to be using the circuit with guitars or any other device that has a high-impedance output stage.

You have several options to correct this.  You could raise the value of all the feedback components in the initial stage at the cost of increased noise.  Alternatively, you could stick an FET follower on each input at the cost of additional component count and complexity.

George is absolutely correct about the 1KA on the headphone stage. It will only cause problems as drawn.  Run the headphone stage at a sufficient fixed gain, and vary the level of the input signal (which you're already doing with the 100kA the leads up to the phase selection stage).  And yeah, you're going to want a big ol' coupling capacitor on the output so it doesn't have a dc component going to the phones.

And speaking of the phase selection section -- is it really necessary?

[sean k]

thankyou Chaps, I'd figured out I needed a cap on the output but forgotten rolloff so it'll be a big cap. The original circuit had a 470 ohm resistor and I just put a 1k pot there to bring things up or down so do I drop the 1k and put in the 470 ohm as originally spec'd or leave the resistor out completely and use a jumper(circuits already etched).

Phase reversal is because I'll have the out on this going through various fx and loops and stuff so I wanna be able to change phase to suit whatever I'm hearing with the other ear which'll mostly be open to the end amp sounds.

Everything going into the mixer inputs will be from FX outputs or something similar with no pickup coils connected straight into the mixer inputs so I suppose it'll all be quite hot. If it doesn't work as is I could go to 51k with 47k in the feedback loop with a 50k pot. I'm hoping that its all at a low enough impedance going in that I can not amp anything so much as have it at unity or cut and then have enough on the out to just get enough volume to monitor in one ear in an enviroment that isn't overly loud anyways. No where near regular stage volumes. If it don't work I'll add fets on little boards right on the input jacks.

So you mean 8.2k for the resistor from 9V to the bias voltage as the top being one half of the divider to ground to create the bias voltage so it ends up at about 4.8-5V as opposed to 4.5?

[Ben N]

I'm not getting why the inputs are inverting--why not make them non-inverting and control level with pots as votage dividers going into the mixer? As noted above, you are sacrificing input impedance which may not matter for electonic instruments, but will for stringed ones. Also, I think a 5532 (or a pair of them) would do a better job as a headphone amp. In fact, two halves of a 5532 would be more than sufficient, leaving an opamp free for a separate output stage to drive your downstream processors/amps, or an opamp phase switcher (see below). Also, if you want to use a pair of headphones, you should use a stereo jack, with the ring and tip connected together; if you are monitoring with one ear and feeding the signal to additional processing stages downstream you will need a second (mono) jack on the output

I'm also not getting the rationale for the phase section--unless you are going to be running the same signals feeding into the mixer in parallel with it, absolute phase wont make any audible difference--it will affect your headphones and downstream processors equally. You can also do with using an opamp; check out the Adjusticator article at Geofex.

[sean k]

I'm doing what I'm doing because I don't know any better and this is what I found in books around the house 
Seeing the adjusticator and that way simple mixer would have saved me alot of time and effort given a TLO71 would have been all I needed. Oh well, the monies spent and I haven't got enough persulphate left to do another board.

I suppose the easiest fix is to up the input resistors to 100k and then get 500k or 1M pots.

The rational behind the phase thing, and it may not be rational at all, is that I'm using this point to monitor my own sound which may end up in reverse of what the PA sends out. I know that sounds like its silly but I suppose I just wanted to be able to switch if I needed to, or whether I would even need to. In hindsight it seems rather silly... but oh well, I've done it now so I'll be able to see how it works.

The thing is I'll have about five instruments, with just an effect or two on each, going into the  mixer  then go into a parallel effects loop so this ability to turn phase before modulation and delays will give me a perspective on mixing dry and wet I'm keen to check out.

I feel silly now for jumping in so fast and not asking about an easy way which would have been the adjusticator... But this is how I learn... by being kinda right but alot of wrong 

I suppose I just wanted to drill a board with 263 holes in it instead of 30 to 40 

[George Giblet]

Regarding the 1k/470R resistor on the headphone amplifier,  I understand what has gone wrong.

I suspect the original circuit used an opamp to drive the point where the 1k/470R resistor connects to the + inputs of the opamps?  yes?

In that circuit the opamp you have removed is like the "master" opamp in the rane circuit I gave and the other opamps are the current boost slaves.   The 1k/470R resistor allows the removed opamp to also provide some of the current drive.

Removing that opamp upsets the whole show.   The configuration needs the "master" opamp.  With your circuit as draw the 1k/470R no longer makes sense it doesn't belong to the parts left it belongs to the opamp you have removed.  The phase inverter stage you have now cannot supply the current like an opamp.

[sean k]

Thanks George, I'll have a look... indeed it did! 
Now what possesed me to overlok that fact. Valuable lesson in progress.

Anyways I've ditched the whole conflaguration for the moment and have just etched the adjusticator with multiple inputs.

But I will perservere with the schematic I've drawn up and create a subsidary board to slip in an extra opamp as a driver between the j-fet and the impedance converter. Thanks for all your help guys and I think I'll be able to create something that works even if it is overly complicated.

Seeing that we are on the complicated track and that using the inverting mode as the mixer inputs with the 10k after the 1uf setting the impedance any input signal sees then is it possible to use a dual gang 100k pot with one being the input impedance and the other being the gain pot on the loop between output and input.

The reasoning being that the two pots would track each other and so it could then be used to set input impedance while also tracking the gain. I'm thinking that at 10k impedance the gain would be 2 and at 100k it would be very slightly over unity but if the 10k resistor where removed between the input resistance and the feedback resistance then the gain at all impedances would remain at unity. Am I right?

[Ben N]

Hey, Sean, no need to get embarrassed. This is how we learn, and your gumption in taking on this project at all is pretty impressive.

OK, I'm no expert, but I'm not sure how good an idea it is to use the adjusticator as your mixer, because when you switch between inverting and non-inverting operation you will substantially change the input impedance. A better idea is to use one inverting opamp as a mixer (what they call a vrtual ground or virtual earth mixer), and another (the Adjusticator) to switch phase and buffer the headphone amp.

Before you etch anything, you might also want to get yourself a breadboard, so you can try stuff out without marrying it. Also, read up a bit on opamps. There are plenty of decent articles a Google search away that will start clearing things up for you.

Good luck.

[sean k]

Ben, I didn't use the circuit for changing phase on the adjusticator but the simple mixer that is on the same page which is just an adder, without input amps, and a pot on each input.

For the moment I've dropped the headphone option and I'm just doing a simple mixer as times at a premium.

Yes indeed to breadboards, I'm a recent convert, but have still only got two little ones and I've only experimented with timers and switches to run ringing oscillators and kids sampling toys but I've got some ideas for this months contest I wanna breadboard as soon as my broadband comes back and I have some time.

Trouble is that my favourite part of this electronics thing is converting complex schematics, well reasonably complex, to PCBs then drawing them up and etching them. I love doing that even though the failure rate is quite high 

Next on the list is definitely bigger breadboards and a setup to have pots and power supplies plug into it. I love this stuff, electronics, for the rationality of it and its high time I started becoming rational in the way I approach my builds... I hope