Suggestions for Pro-Audio Line Splitter Circuit

[dastuff]

My initial thinking is that a good quality transistor with unity gain using (2) 9V batteries (18V) for headroom would make an ideal line splitter for a guitar. Am i on the right track here? I want to achieve lowest coloration of sound and lowest S/N. I would think the transister solution would be better than an op-amp solution? Those with audio experience, what transister would you recommend for the job? Perhaps there is a circuit diagram available?

[Enzo]

Are you talking about something to plug the guitar into directly, then into two amps - as in an AB box?

SInce the guitar doesn't put out over about a quarter of a volt, I don't think headroom is an issue there.

[R.G.]

My initial thinking is that a good quality transistor with unity gain using (2) 9V batteries (18V) for headroom would make an ideal line splitter for a guitar. Am i on the right track here? I want to achieve lowest coloration of sound and lowest S/N. I would think the transister solution would be better than an op-amp solution? Those with audio experience, what transister would you recommend for the job? Perhaps there is a circuit diagram available?

It never ceases to amaze me the depth of the penetration of the audio psycobabble coming out of the "less is more" hifi fundamentalist tweakos. Essentially every single statement you've made is right out of the "Absolute Sound" playbook. And almost entirely either irrelevant or incorrect.

Forgive me dastuff; you may well be a great guy who's simply fallen prey to believing whatever you read a few times. This particular rant is not at you, but at the ones who've filled you with this.

The lowest coloration of the sound is going to be the lowest distortion; single transistor buffers get down to maybe 1/2% if you severely limit the signal range to much less than the available swing. On the other hand, a modern high bandwidth opamp will have distortion under 0.01% for signals up to probably 3/4 of the available power supply voltage. A simple, single emitter follower may have easily measurable distortion for large signals.

Opamps have objectionable sounds **when they clip**. If they never clip, they are really very, very transparent indeed. So the name of the game for clean, transparent sound processing for something like a guitar splitter is to use an opamp stage with plenty of power supply room.

A guitar will have audible treble loss if it drives any significant capacitance in the connecting cable or less than about 1M of input impedance in the buffer at the end of the cable. It is quite difficult to get a single transistor buffer with an input impedance of 1M or over. A darlington, two-transistor-feedback-pair, or bootstrap input, maybe. But then it's getting complicated. Less complexity gives up treble from loading the guitar. And all of these are more colored than a JFET input high speed opamp. The LM833 is in fact designed for this kind of work.

You have to know the guitar's signal before you can estimate headroom. Single coils run to about 100mV, humbuckers to maybe 1V, and "super humbuckers" may have a couple of volts. Bass can run a couple of volts and active bass can be 5-6V. So you have to pick what you're going to allow in the input jack before you can start nailing down headroom.

I think you're actually after the HIGHEST S/N. Lowest Signal to Noise would bury the signal under and avalanche of hiss. In designing low noise inputs, there is much, much more involved than deciding that a single transistor is good. This is one of the places where there are important impedance considerations (as opposed to "matching") to be done. The lowest noise device at one impedance level may not be at another level.

Finally, the worst thing you'll run into is not hiss and headroom - it's hum. Splitting guitar into multiple amps is a classic case for where you get hum, and almost a textbook case of what to do to get rid of it.

There are some isolating splitters on geofex (http://www.geofex.com) that attack the ugly hum problem first, and use low noise JFET opamps to do it. One of them even uses an 18V power supply for more headroom. People have built them and like them very much.

dastuff - don't let the tweakos fill you up. There is more utter crap flying around out there about what makes for good sound than you would believe possible. It's become a gotta-have-a-gimmick world in hifi. I don't want to see water jacketed cables and "improved clarity of sound stage, with more space between the notes, as though a subtle veiling had been lifted" permeate the musical world. It's bad enough in here as it is.

[dastuff]

This "crap" was entirely my own doing. I have seen circuits some years back of buffer circuts made with a single transister. I remember the diagram notes mentioning using decent supply voltage. The notes said these circuits gave very good audio results. Considering an op-amp has 50+ components in its IC with varying tolerances, one can see how a simpler circuit might have less noise, less coloration of the sound. I agree with you that there are a lot of modern day "snake oil sellers" out there. I want the real deal, not crap-ola.

http://www.jimdunlop.com/index.php?p...p_and_e_detail

I think using quality op-amps with a 18V supply (versus 9V) may be the way to go. I have an MXR KFK-1 EQ that uses op-amps and 18V and it is fairly transparent when comparing it with direct. I think it can also run off of (2)9V batteries which would eliminate AC hum (with hi gain guitar sounds). This particular stompbox has two outputs, so i have my split. One side will go into the side chain of a drawmer noise gate. The other goes through multiple guitar modeler processors. My concern is introducing noise at the beginning of the chain before it goes into the modeling. I like using vintage pickups so using active pickups to get best S/N is not an option. I am experimenting with the next generation of high gain guitar sounds. To be as clean as possible before the modelers is important otherwise the high gain sound becomes too grey or muddy (lacks adequate definition).

Maybe all i need to do is to jump/bypass the eq section of the MXR pedal so to achieve input --> volume --> output, maybe get a good enough sound where i do not have to build a circuit myself. Maybe you folks can help me with this? I can soder, have oscilloscope, etctera. I just need to know where to break a connection, where to soder a wire. I figure i would trace the audio path in the MXR and then break it when it goes in and out of the section of eq filters, then soder a jumper wire. Does that sound about right?

[Enzo]

das, the hum is not from the power supply, it is from ground loops that occur when two pieces of equipment get their grounds connected together as inevitably happens when splitting. Batteries won't solve that. That is why RG mentioned isolating splitters on his web site.

[Steve Conner]

Haha, go get 'em, RG! Op-amps don't deserve the slamming they get. Really, all you need is a TL071, two 9v batteries and a 1-Meg resistor, and there's your buffer thingy right there. I swear on a stack of the sacred text of your choice that any distortion introduced by the op-amp won't be audible. They may have 50 components, but thanks to the magic of negative feedback, the tolerances and non-linearities of them are all cancelled. They should rebrand it to "inverse feedback" to get rid of the negative connotations.

It may increase the noise floor a little if you're putting it before some kind of gain monster rig, but so would the transistor you originally thought of. Again, thanks to the magic of NFB, the noise floor of an op-amp is determined only by its input stage, and the contributions of the other 48 components are cancelled.

If you have trouble with hum, you can add a couple of those Mouser #42TM018 transformers that RG recommends. One should probably do it if the splitter is battery powered.

[R.G.]

This "crap" was entirely my own doing. I have seen circuits some years back of buffer circuts made with a single transister. I remember the diagram notes mentioning using decent supply voltage. The notes said these circuits gave very good audio results. Considering an op-amp has 50+ components in its IC with varying tolerances, one can see how a simpler circuit might have less noise, less coloration of the sound. I agree with you that there are a lot of modern day "snake oil sellers" out there. I want the real deal, not crap-ola.

You sound like a reasonable guy again. As Steve points out, the 50-plus parts and their tolerances inside an opamp are almost totally (like to 0.001%) hidden by the feedback. That is, in fact, the whole point of feedback. One can use parts in a design, and the tolerances and quirks are ironed out almost completely.

High gain feedback amps like modern opamps remove the reliance on every part in the chain and force essentially all of the performance to be dependent on the parts in the feedback path. That way, you get down to possibly one or two passive parts controlling the whole mess.

By the way, a transistor emitter follower is also a feedback amplifier. It has 100% negative feedback from the resistor on the emitter. Not many of the hifi tweakos like to talk about that one.

I think using quality op-amps with a 18V supply (versus 9V) may be the way to go.

I vote with you on that one.

I think it can also run off of (2)9V batteries which would eliminate AC hum (with hi gain guitar sounds).

As Enzo notes, that will get rid of power supply hum, but the wiring-interconnection/ground loop hum will still be there.

This particular stompbox has two outputs, so i have my split. One side will go into the side chain of a drawmer noise gate. The other goes through multiple guitar modeler processors. My concern is introducing noise at the beginning of the chain before it goes into the modeling.

There exist opamps designed especially for low noise audio; including the LM381, which can be set up to use only one of the two input transistors to get rid of the square-root-of-two more noise from a diffamp, that get you down to unbelievably low noise levels. It's a 30+ year old part. I had a friend who used one of these for a phono preamp (yep, back when music came on vinyl). He would turn on the power amps (dead silent) then turn on the preamp - again, dead silence - then turn the volume knob all the way up. Still no sound. You could put your ear in front of the tweeter and could not hear hiss greater than your ear's own internal noise. He would then cue the tone arm over the record, and as the stylus touched the record, the pop would shake windows. Anecdotal evidence, and hence worthless, but it's illustrative. I was there. I was impressed.

An LM833 is not much noisier than the LM381 and was designed for audio applications. I think it will be less than the thermal noise of your biasing resistors.

Your real problem will likely be hum.

I like using vintage pickups so using active pickups to get best S/N is not an option. I am experimenting with the next generation of high gain guitar sounds. To be as clean as possible before the modelers is important otherwise the high gain sound becomes too grey or muddy (lacks adequate definition).

I wish you luck. Sounds like a noble calling.

[dastuff]

I think i rememebr seeing a TL071 or TL072 somewhere in my parts bin. I will look into the 381 if the TL circuit is not quieter than the MXR EQ pedal.

Thanks for the guidance.