"high quality" audio-splitter

[balkanizeyou]

Hello!
My friend who has a small studio wants me to make a line-level signal splitter for him where one input would be split into 6 identical outputs (gain has to be equal or close to 1). I just thought about using one op-amp as a voltage follower as an input buffer, then adding another 6 voltage followers as a buffer for each output - is there a downside to this approach? Or maybe the input buffer is not even needed? Or maybe there's a better way to do it?
This friend of mine, like a lot of studio guys, is obsessed with the "audio quality" - would using expensive op-amps designed for audio use instead of a good old 5532 or similar opamp even make a difference in this application?

[PBE6]

If you use a large biasing resistor (1M or larger) and configure the opamps as non-inverting buffers, the input buffer is unlikely to improve the performance.

With regard to audio quality, you should be fine with TL072s or NE5532s. They are very widely used in audio. I doubt you'd notice any difference with a more expensive chip, especially in a buffer configuration, but if you do use one make sure you charge triple for the box...it all adds to the mystique you see...


Sent from my iPhone using Tapatalk

[jimilee]

Would an aby box work?


Sent from my iPhone using Tapatalk

[ashcat_lt]

Most line level outputs can split out to quite a number of typical line inputs with nothing but straight wire, and if you or your friend think that a bunch of opamps is going to be better quality than that...

The biggest issue is going to be sorting out any ground loops that are created in the process.  The best ways to solve that usually don't require active circuitry either.

[GibsonGM]

However, if you DO want a buffered output, JFETS are great and you can set up as many outs as you wish.  As stated, quality won't be an issue as long as you were to follow proper build techniques, shielding and whatnot.


http://www.muzique.com/lab/splitter.htm

[balkanizeyou]

thanks for all replies and suggestions!

haha, my friend probably won't be very happy with the idea "ok, so we should probably just split the signal into six paths using a piece of wire", he probably wanted to hear something in the mold of "ok, so we are going to use this hyper expensive audiophile mojo opamp as a buffer, then use the best quality capacitors and resistors 5$ each and a golden wire" 

But considering the fact that the input signal will probably be coming from a mixer and then the outputs will be fed to another mixer, the buffers probably won't be necessary as you say.

I will discuss the idea with him, but if he insists on using buffers - does the JFET buffer suggested by GibsonGM offer better performance in any way then using a JFET input op-amp? It's much easier for me to get a tl072 than a discrete JFET (and cheaper too).

Oh, and I just found out that he wants to split 8 separate signals, each into 6 paths which makes the whole thing a 48 output device  if I wanted to use some crazy audiophile op-amps for the output buffers, it would probably cost a $500 or more just for the ICs 

[GibsonGM]

The only advantage that I see in the AMZ splitter over opamps is the flexibility of just adding a JFET section for each output.  It could be modified later due to this flexibility.    Only 1 bias network, so each is simple to add in.   

I personally love TL072s - which DO have a JFET input! - but don't like going above the dual opamp for things...just my feeling that quads make things a bit more...cluttered.   Only an opinion. 

In the case of wanting 48 OUTPUTS (!!), you're probably better off using opamps, LOL!   You're going to need some quads...if they're socketed, then any problems that could arise are easier fixed by repairing the issue and popping a new chip in     

[grapefruit]

To work out the best way to do it you really need to know the input impedance of the equipment you are going into. Assuming all inputs are  10k ohm the combined impedance is 1.66k ohms.  If the output of the mixer is capable of driving 600 ohms with low distortion then a passive split is a reasonable solution.

The above assumes all your outputs and inputs are balanced. If you use a combination of balanced and unbalanced inputs it will become unbalanced and you have to be careful about the wiring. You  could run the whole thing unbalanced but if he's pedantic about audio quality I'm assuming all the equipment has balanced inputs and outputs.

So, if his equipment does have balanced inputs and outputs and you want to build an active splitter, you really need to design a distribution amp with balanced inputs and outputs, or buy one off the shelf, which may be cheaper if you allow for your time.

Cheers,
Stew.

Sent from my GT-I9506 using Tapatalk

[ashcat_lt]

I personally love TL072s - which DO have a JFET input! - but don't like going above the dual opamp for things...just my feeling that quads make things a bit more...cluttered.   Only an opinion.

If you're actually trying to do anything with the opamp, yeah.  In this case, it's just ins and outs and jumpers, so it shouldn't get too terrible bad.  Do they make octal opamps? 

[GibsonGM]

I personally love TL072s - which DO have a JFET input! - but don't like going above the dual opamp for things...just my feeling that quads make things a bit more...cluttered.   Only an opinion.

If you're actually trying to do anything with the opamp, yeah.  In this case, it's just ins and outs and jumpers, so it shouldn't get too terrible bad.  Do they make octal opamps? 

I like them for low noise, and you don't have to think much about them for audio work - workhorse.   I've never seen an octal, myself...I even did a quick search, just for the heck of it - nothing.    So, some quads...

Like anything, it really depends on how far one wants to take it.    Grapefruit nailed it, sometimes the time is better spent...me, I'd be ok with 741s, LOL!     After all, that was ALL they had back when, and the music came out pretty quiet, ha ha...you could do the XLR thing, really make a cadillac....or just split those signals and be done with it!

[balkanizeyou]

Thanks!
I'm going to see him tomorrow so I hope to find out what his expectations are and the specs of his mixer (which is of course some vintage mojo full analogue blah blah...)

But for now I know that all the inputs and outputs are balanced, so I guess it makes the job pretty easy - find out if his mixer can drive reasonably low impendance load, if it can a passive split is a reasonable option, and if it doesn't - a single op-amp input buffer per channel should do the job, right? Wouldn't be that bad - just 6 single op-amps! (I guess using duals wouldn't be recommended in this case) If that should do the job, then I guess my buddy would probably be willing to spend some more for, say, OPA134 (not that it's going to make a difference anyway).

And I'm not really familiar with dealing with line-level signals, but I assume they shouldn't be hotter than 2V rail-to-rail? So the whole thing could be run from a single +9v power supply?

[ashcat_lt]

TL07xs (apparently) aren't really great at driving very low-Z loads.  At least, this is what I've heard from somebody (PRR?) a couple of times around here.  That 1.6K mentioned above gets close to where theses opamps start to current limit.

Assuming this is "pro" line level gear, nominal level will be +4dbu, so about 3.5V p2p, but most mixers have 18 or 20db headroom over that, which gets close to 30V.  You'd probably want to run at least +-18V bipolar supply, and specify parts that can handle that.

[balkanizeyou]

damn, this is slowly approaching a "dude, buy yourself a pre-made one" area, because I don't know of any available op-amps that are good for audio applications and can easily handle a +-18V input/output, so I would have to settle for some discrete circuitry, then multiply that circuit 6 times, and I would have to make a power supply as well - not saying it's impossible, but buying a commercial distribution amp seems like a cheaper and much less time-consuming thing to do. I'll discuss this matter with him tomorrow, meanwhile does anyone know an op-amp that satisfies the above conditions? It would simplify a lot of things (or I'll tell the friend to keep the levels low  )

[PRR]

> I don't know of any available op-amps that are good for audio applications and can easily handle a +-18V input/output

'5532 will do 18V peak (with max-rated +/-22V supply).

Frankly you only have to be same-as the output you are coming from. Pretty good chance that is powered +/-15V, and very likely a '5532 or one of the JRC similars. Throw six of them together, dun.

[Transmogrifox]

Probably wiser to give a 5532 +/- 18V.  Generally not a good practice to operate at absolute max on purpose.

Either way I agree the 5532 is a good choice for this job.  It will ensure you can drive every line at 600 ohms if necessary.

The passive option would be my choice if the upstream equipment can drive the load he's connected to it.  The most hi-fi amp you can use is none at all.  There may be value in making a tool that doesn't depend on the drive strength of upstream equipment.  It depends whether this is a permanent installation or if he wants some flexibility if he later changes something out.

The active circuit will be best suited as a general solution, like what would be sold commercially when the goal is to make it compatible with most any kind of equipment.  The passive circuit can be designed into a specific system if the specific system is permanently defined.  This is part of the advantage of the DIY approach -- you can build something specifically for your rig without having to worry about whether it would work as well in any other rig.