Buffer thoughts

[phuzle]

So from the reading I have done, it seems the buffer debate ends with the conclusion that having a buffer at the beginning and end of your pedal board is generally a good thing if you have lots of TB pedals.  BUT, this doesn't really solve the problem, I don't think...

Guitar gives off HiZ, which does not run well through long cables, so why bother with the buffer at the END of your cable from guitar?  You'd want to convert to LowZ to drive the first long cable.  Seems to me something like the Tillman pre-amp cable makes more sense (though IMO better without the 3db of gain).

So then, if the amp wants a HiZ input, shouldn't we have something immediately before the input of the amp to convert the LowZ running from our last guitar effect, which runs through our cable, to HiZ for the input of the amp?

What I mean is, we're thinking on the wrong end of our cables!  And I feel that the buffer at the end of the pedal board just doesn't really make sense.  Am I right?

[phuzle]

Does anyone have thoughts on this?  I feel like this could be the absolute solution to this whole debate, but I'd like to hear some input before I go and build it.  Also I am not sure how to go from LowZ to HiZ, does anyone have a schematic for a buffer that could do this?

[spaceace76]

interesting thought!

i suppose it would depend on how long the cable between the guitar and board are, and how long the cable from board to amp is. If you had long cable runs on both, and noticed a significant tone loss, then you would need buffers either in your guitar (or active pups) or at the beginning of the board, and at the end of the board, and within this discussion, a final buffer to convert everything back to HiZ.

If you're using active pickups, you'll only need the two, hi to low and low to hi. but with normal pickups, you're up to three buffers.

So with three buffers, you're slowly moving away from the original signal. The extent of which depends on the circuit you're using, of course. And how much difference you can notice or even care about (even when using a lo-to-hi buffer for the amp) is another story entirely.

So basically, like with all the other arguments for/against true/buffered bypass, the choice is entirely up to the stomboxaholic in question. you should try this out on your own and see what you come up with! anyone have a buffer that changes low Z to hi Z?

[BAARON]

BUT, this doesn't really solve the problem, I don't think...

What "problem" are you trying to solve?  Just because guitars have a HiZ output doesn't mean the amp WANTS a HiZ signal hitting it.  (That's like saying that we shouldn't mash potatoes because they aren't mashed in their default, normal state, despite the fact that it'll taste the same and have the same nutrients.)  The amp doesn't much care what it sees, it'll do its job anyway.  It's what the GUITAR sees that affects what parts of the signal are going to reache the amp, and if you start throwing around HiZ signals for no good reason, you're probably going to end up with some signal loss.

Most amps have a fairly high impedance input, which makes sure most of the guitar's signal gets amplified.  If you have an input before your effects chain with a high input impedance (i.e., the same input impedance as your amp has), it makes sure that the same range of guitar signal is getting passed on to the amp: voila, a transparent buffer that preserves the proper frequency response of the guitar, while also driving a long cable & TB effects chain without high-end loss due to cable impedance and capacitance.

The reason people like a LowZ output buffer at the end of their effects chain is to make sure that nothing gets lost between the pedals and the amp (in case a pedal has poor output impedance), because it's silly to work all sorts of magic with the pedals and then lose it all before it reaches the part of the signal chain that actually makes it audible.

Want to add impedance to the signal just before it hits the amp?  Run your signal through a resistor.  That's exactly what's happening with tube amps where they have "low" and "high" impedance input jacks.  (Read here: http://www.runoffgroove.com/fetzervalve.html#2)


As for buffering at the start of the first cable rather than at the start of your effects chain, yes, you can do that, if you want to start adding extra gear to the guitar end of the process.  I think it's probably just more convenient to have the stompbox on the floor, and it makes sure you get the little bit of cable loss you'd have if you just plugged that same cable straight into the amp... thus making it sound more like there's no buffer at all, just a guitar straight into the amp.

[phuzle]

What "problem" are you trying to solve?  

The problem I am trying to solve is that of treble loss through long cable runs.  Seems to me everyone is against buffers, but hip to the use of them at the beginning of their chain, and possibly at the end.  I just had a thought that it could be done better.  If you've got a dozen true bypass pedals, all off, with a buffer on at the beginning of your pedal board, you've essentially put the buffer in the middle of your chain.  But it should be at the beginning to truly serve its intended purpose, yes?

Then I got to thinking about the use of a buffer at the end of the pedal board chain, which seems to be correctly placed to drive the signal through the cable to the amp.  I just assumed that amps (I am thinking tube amps here) wanted a hiZ signal presented to them since that is what guitars give so why not from a stompbox also?  So if that is wrong then it is good to know    You say that amps do not care what is presented to them?  There is no "ideal" impedance for an amp?

Want to add impedance to the signal just before it hits the amp?  Run your signal through a resistor.  That's exactly what's happening with tube amps where they have "low" and "high" impedance input jacks.  (Read here: http://www.runoffgroove.com/fetzervalve.html#2)


As for buffering at the start of the first cable rather than at the start of your effects chain, yes, you can do that, if you want to start adding extra gear to the guitar end of the process.  I think it's probably just more convenient to have the stompbox on the floor, and it makes sure you get the little bit of cable loss you'd have if you just plugged that same cable straight into the amp... thus making it sound more like there's no buffer at all, just a guitar straight into the amp.

I understand your thinking that having some cable loss before hitting the buffer is more like having the guitar running straight into the amp.  But I am wondering if that is desirable or just a necessary fact that we have to put up with because we don't have buffers at the start of the cable.  What do you think?  Is a little bit off loss a good thing?  Perhaps we should keep the full signal and dial in the amount of treble loss we want, rather than use treble boosters as are so popular these days...

I never understood the point of low and high impedance input jacks, and that article is technical enough that I still don't understand the point - I only understand how its done.  Why are they used, and how do you use them properly?

Thank you!

[Processaurus]

A benefit of the buffer at the end of a pedalboard is that it guarantees presenting the cable to your amp and amp input with both a low impedance signal, and an unchanging impedance signal, despite what effects are switched on or off.  Most DIY effects have a volume pot on the end, that can make the output impedance not so low, say, if the output pot is 100K, and it's turned down halfway, you might be getting a 50K output impedance.  Not as good, compared to a discrete buffer, like most Boss pedals (5K-10K) or an opamp buffer (a few hundred ohms).  Some fuzzes use a 500K pot because they need them to sound right, that can be a weak signal, that is susceptible to noise and interaction with the next device.

Example, there's no out buffer on a pedalboard, and you're quickly setting up for a show and have the cable to the amp going over something electically noisy, like along an extension cord, or by a wallwart (the non switching type put out lots of EMI, that can turn into hum) on the ground.  You have a low impedance output effect on, no hum from the amp, but then when you switch on the vintage fuzz, you get hum, because the current the noisy AC is putting on the cable is a much higher percentage compared to the weak, high impedance fuzz output as your other, low impedance effect.  So it hums when the fuzz is on, weird, it never happened before, and someone might naturally think the problem is in the fuzz.  A buffer makes sure the impedance is always low, after a bunch of different pedals that might be on or off and all have different impedance output stages.

[spaceace76]

i think phuzle meant that if an amp wanted to "see" a high impedance signal at it's input, you could have a lo-to-hi buffer after the long cable from the pedalboard, directly in front of the amp's input. Perhaps in a box sitting on top of the amp or, if the circuit is simple enough, a small box that plugs right into the amp's input jack.

it has been stated though, that an amp does not really need a high impedance input signal. is there a "right" input impedance for any given amp? I play mostly vintage style amps, so i'm guessing running a buffer that puts out lower impedance isn't quite the right idea for this setup, as these amps were invented before impedance issues were engineered out of pedal design.

[Processaurus]

hi, like BAARON was saying, no amp I can think of actually benefits from seeing a weak, hi impedance (IE not being able to provide much current) signal.  The basic rule in designing electronics impedences between stages, to have good voltage transfer with signals, you want the load to be > 10x the source impedance. 

For power transfer (like from amp to cab), you want to match impedances, and for current transfer, you want the load to be at least 1/10th of the impedance as the source.  The 10 to 1 ratio is the rule of thumb because if 9/10ths of something is getting there, you're usually doing "good enough".  So unlike matching cab impedance to amp speaker outputs, matching the impedance of your guitar signal to amp input is a bad idea if your goal is signal (voltage) transfer, as the higher the source impedance (the less current it can provide) the more voltage is lost on the load (the amps input impedance draining current from the signal source).

The only niche exception to lower output impedance being what you'd benefit from, are classic effects like the fuzz face and rangemaster, some wahs (someone else here would know more examples), that get some of their voicing by intentionally loading down the guitar input, and get squashed by the amount of current a buffer can provide.  These could be modified, by a resistor in series with the input, to behave in a similar way, (like I'm planning to add a variable resistor to the input of my fuzz factory, to get it to sound good on keyboards) but that's a whole other discussion.  Normal high impedance input stages are fine getting either high or low impedance signals.  It's the low impedance inputs that don't work well with hi impedance signals, like happens plugging a guitar straight into a line input on a mixer, classic tonesucking.

to hear the effects of making "low-z to hi-z" firsthand you could just solder a >100K resistor in a cable, in series with the tip connection from the center wire to the plug's tip solder connection.  Most likely it would just make things quieter, maybe noisier.  Hope that helps answer those concerns.

[BAARON]

i think phuzle meant that if an amp wanted to "see" a high impedance signal at it's input, you could have a lo-to-hi buffer after the long cable from the pedalboard, directly in front of the amp's input. Perhaps in a box sitting on top of the amp or, if the circuit is simple enough, a small box that plugs right into the amp's input jack.

it has been stated though, that an amp does not really need a high impedance input signal. is there a "right" input impedance for any given amp? I play mostly vintage style amps, so i'm guessing running a buffer that puts out lower impedance isn't quite the right idea for this setup, as these amps were invented before impedance issues were engineered out of pedal design.

No, there's no "right" input impedance for amps.  As long as the signal reaching them isn't phenomenally HiZ they'll be fine.  A vintage Fender amp has the same input stage as a modern Fender amp.

[spaceace76]

Normal high impedance input stages are fine getting either high or low impedance signals.  It's the low impedance inputs that don't work well with hi impedance signals, like happens plugging a guitar straight into a line input on a mixer, classic tonesucking.

ah. so then amps don't necessarily care what the input impedance is, unless you make use of the low impedance input and have a very high input impedance.

so in reality, if you're jumping channels in a vintage marshall, and you aren't using a buffer, you aren't making the most of channel jumping! if you channel jump and are using long cable runs, one channel will always have the sucked out tone delivered to it, unless you're using a buffer for low impedance, in which case the mismatch doesn't occur. that explains a lot! i have to use short cable runs or the tone is wrecked. i thought it was just because i haven't added a buffer to my chain yet, but this explanation rounds out the missing details. thanks!

[CynicalMan]

I just wrote this article today without even realizing there was a thread currently going on about it. 

Because of grid to cathode capacitance and input resistors, amps will sound slightly different given different input impedances. With a hi-Z input you'd here a slightly quieter sound with a bit of treble reduction but it would have to be pretty big for you to notice much difference. See the above article for a model of an amp's front end.

[BAARON]

Normal high impedance input stages are fine getting either high or low impedance signals.  It's the low impedance inputs that don't work well with hi impedance signals, like happens plugging a guitar straight into a line input on a mixer, classic tonesucking.

ah. so then amps don't necessarily care what the input impedance is, unless you make use of the low impedance input and have a very high input impedance.

so in reality, if you're jumping channels in a vintage marshall, and you aren't using a buffer, you aren't making the most of channel jumping! if you channel jump and are using long cable runs, one channel will always have the sucked out tone delivered to it, unless you're using a buffer for low impedance, in which case the mismatch doesn't occur. that explains a lot! i have to use short cable runs or the tone is wrecked. i thought it was just because i haven't added a buffer to my chain yet, but this explanation rounds out the missing details. thanks!

That was the Point of the low impedance input: to reduce gain and smooth out the sound a bit.  It's not "wrecking" the tone or misbehaving in anyway... it was just an idea that ended up being kinda crappy, if you ask me.  That's why you don't see it nearly as often on modern amps.

[spaceace76]

That was the Point of the low impedance input: to reduce gain and smooth out the sound a bit.  It's not "wrecking" the tone or misbehaving in anyway... it was just an idea that ended up being kinda crappy, if you ask me.  That's why you don't see it nearly as often on modern amps.

i use really short cable runs, so i don't hear a huge difference in tone or gain when plugging into either one. i usually just plug into the high and jump the channels, and don't give it a second thought. when i make up a full pedalboard i'll be using much longer cables. maybe i'll install a switch for the input resistor if i feel like throwing caution to the wind and going bufferless.

CynicalMan, that article is a great simulation! what buffer circuit did you use? can you run another simulation with the cable after the buffer?

[CynicalMan]

The buffer was a simple ideal buffer model. ∞ input impedance and 0 output impedance. Because of the 0Ω output impedance, having the cable capacitance after the buffer wouldn't do anything. Maybe I could do a part 2, trying out different buffer circuits once I get the whole website set up.