Mixer project

[Bobv]

My plan is to make a decent working 4 channel mixer. (supermix?)
I did some searches and found this thread which talked about GGG's mini mixer sucking tone, and how it can be fixed by putting a buffer in front of the inputs that will get guitars, and even a plan for the IC version.
I want to make the same thing but with JFET buffers (also from GGG)
Would simply making the buffer circuits going into the inputs of the Mini Mixer do the job? I threw together a layout of what it would be like - i feel like there are too many capacitors in there - Are the output capacitor on the buffer not needed because of the input capacitor on mixer(edit: other way around)? Anything else that can be improved?

So basically - would this work, and is there any way i can improve it?


GGG JFET Buffer Four of these
GGG MiniMixer  Going into one of these
My Layout

[PerroGrande]

The mixer, as shown in your posting, is very likely to suck tone from a high-impedance source such as a guitar pickup, given that the impedance presented will be dominated by the input potentiometer acting as a voltage divider. 

The idea of putting a buffer stage before each input will correct that issue.  A JFET buffer such as the one posted will present an impedance that will not suck tone and should be able to drive the mixer stage.

You're also correct that you do not need back-to-back capacitors to AC-couple the buffer output to the mixer input.  One capacitor between the stages is sufficient.  Circuits are generally designed to accept an input that might not be "well-behaved" (i.e. may contain a DC offset).  Since you control both stages in this case, you can eliminate one of the caps.

So -- I think you're well on your way to a successful project.  Go for it!

[Bobv]

I breadboarded this thing but with only one input and it sounds much better going through the buffer stage than just into the mixer so I'm assuming it's worth making, thanks for the reply sometimes i make retarded mistakes and not notice it - its nice for someone else to be able to look at it and say it works.
To clarify - should i dump the .1uf caps and just keep the 10uf?

[R.G.]

Save yourself some wear and tear, as well as some time. Go buy a copy of Craig Anderton's book "Home Recording for Musicians". The book tells you how to build your own mixer, in great detail, with all the pieces explained.

Putting together a mixer out of the bits and pieces here will be very limiting.

[Barcode80]

Save yourself some wear and tear, as well as some time. Go buy a copy of Craig Anderton's book "Home Recording for Musicians". The book tells you how to build your own mixer, in great detail, with all the pieces explained.

Putting together a mixer out of the bits and pieces here will be very limiting.

but educational, not a waste. but +1 on that book. i don't own it, but i hear great things...

[Bobv]

I've heard about Craig Anderton's books before - just checked Amazon and saw a used Home Recording for Musicians for $2+shipping, so i ordered that - I'll hold up on building this thing until i read what the book has to offer, thanks for the suggestion.

[R.G.]

Great! I think you'll like it.

[Bobv]

I got the book and there's some great info in here - but the mixer here still uses 10k pots and has unbuffered inputs, the suggestion is to use a preamp (which should do about the same thing as a buffer here?)

Would my original plan work well (is it the four .1uf caps i can eliminate?)
Also, would there be any disadvantage to plugging a low impedance device in through the buffer (the book says noise) would it be noticible enough to add switches that connect input/output of each buffer to bypass it?

Also, would the IC version of the buffer be any better than the JFET version?


neater, smaller layout without the .1uf caps: http://img.photobucket.com/albums/v257/neurd/newboard.gif

[PerroGrande]

On a cursory glance at your board layout (and attempting to draw a mental version of the schematic), things look pretty good.  I don't see any caps back-to-back, so I don't think you've got excess caps in the circuit.   (Again, though, this is from a cursory view, so I'm making the disclaimer that I might have missed something.) 

This should present about a 1meg input impedance to your guitars, which should be more than enough to avoid sucking tone. 

JFET discrete vs. Op Amp?  The op-amp is probably going to have a lower output impedance and arguably some current capabilities that the JFet might not have.  In this case, the JFET buffer is driving a super-high-impedance device (an op amp) over a very short distance (pcb trace).  The difference in output impedance and current capabilities I mentioned probably isn't relevant in this case. On the other hand, it might end up being answered by your ear as much as by electronics.  Some people prefer the sound of a JFET (or especially a MOSFET) to an Op Amp (which some say sounds "colder").  Others make extensive use of op-amps and like the sound.  I think this is largely subjective.  In real production situations, cost, assembly complexity, and PCB real-estate would end up as considerations as well. However, in DIY world, we get to think about the sound

On your question of input impedances:  Well, there are two edges to this knife.  On the one hand, a high impedance circuit does not draw much current.  Ergo, it does not "load down" the output of the previous stage. Nor is it as likely to accidentally create a low-pass filter in the audio range. On the other hand, they will be somewhat more sensitive to noise for precisely the same low-current reason.  JFETs do well, incidentally, when it comes to noise characteristics in Hi-Z buffer applications. Still, proper shielding, etc, is always recommended for noise reduction. On balance, though, I'd rather have high Z input than a low-Z one (in most cases).   If it is of concern, you could always intentionally drop the impedance of one of your inputs and designate it a "CD/Line" input or something like that.  (Making a stiffer voltage divider in R7/R10 could accomplish this).

[Bobv]

Sweet, thanks for the response - i will go over my layout once again (can never check too many times) and then try building it

Should the shielded cable be on just inputs/outputs - or also going to the pots?
as for the shielded cable itself - Should i just buy a roll of single conductor speaker cable?

[PerroGrande]

I like to use shielded cable whenever audio connections leave the circuit board.  That would include pots, yes.  Of course, if your leads are very short and the project is built in a box that can itself be shielded, you might be safe with single-conductor wire.

Speaker cable strikes me as being too thick (and correspondingly hard to work with) for a project like this.  I use thinner (~24AWG) shielded cable -- like the kind one might find on an inexpensive RCA-RCA cable from Radio Shack for short run shielded work.  In short runs, its thinness isn't going to be an issue, and it is much easier to work with. 

[Bobv]

just two more questions: Will these buffered inputs be good for 1/4jack microphones?
Also, if i decide that instead of sending it into my amp i want to send the output directly into my sound card - will this work?

[PerroGrande]

The answer to your second question is the easier of the two...  That answer is yes!  With your output summing/buffer op-amp, you should have no trouble at all driving the input of a sound card.  Even if the sound card has an input impedance of only 10K, the output Z of that op-amp is probably going to be on the order of a few ohms... You're going to be just fine with that.

The first question is more tricky.  From an impedance perspective, most microphones that use a 1/4 plug and an unbalanced signal have a fairly high output impedance (relatively speaking of course, when compared to low-Z mics) along with a fairly high output level.  Even with a higher-Z microphone, you're still going to be using the "low feeds high" mantra when connected to your mixer -- i.e. the input impedance of your mixer is still way higher than a "high Z" mic's output impedance).  So from an impedance perspective, you'll be fine.   The second consideration is the level (amplitude) of the signal.  As I mentioned earlier, a High-Z mic will frequently have a fairly "hot" output, so the lack of a substantial channel gain circuit isn't going to be an issue in all likelihood.  The biggest issue with High-Z microphones is the limitation of line length.  The high-Z of the mic and the shunt capacitance of the cable produces a low-pass filter that starts to noticeably roll-off high frequencies over about a 15-20' cable length.   

So -- that was a lot of writing to say:

Yes (just keep the cable length to within the recommended limit)
and
Yes

[Bobv]

Awesome, Thanks again

[widdly]

If you haven't already, you might like to check out the Rolls website.  They have schematics for a lot of their mixers online.  They are inside the pdf manuals.  Loads of useful ideas there.

[NoFi]

Just a related question, for the inputs, is an op amp buffer followed by a pot really a better/different choice than an op amp arangement with variable gain in that application ? I mean Isn't an op amp also "buffering" when wired as an inverting amplifier (with the correct input resistor for the impedance) ?

[brett]

Hi
yes, an op-amp IC can be used as a buffer in "inverting amplifying" mode.  But... that requires a large input resistor (220k) or moreand an even larger feedback resistor (1M for a voltage gain of just 4).  A feedback resistor of this size starts to introduce noise.
I think a better arrangement would be to use the non-inverting input of an op-amp with high input resistance (e.g. TL072).
cheers