in simple terms a buffer "buffers"a signal by providing a high input impedance to the preceeding stage and a low impedance to the following stage for driving an audio signal (or network cable,logic if cmos buffer,video) without losses in the signal strenght or attenuation of the high frequencies.
One example would be a passive attenuator used in place of a preamp in an audio system.when you choose the ohm value of theattenuator you try for a value high enough to present a good impedance to the output of a CD player but it must be low enough to drive the following power amp stage impedance.
Most CDPs put out from 100 ohms to 300 ohms but can be as high as 1-5K ohms and power amp intputs from 50K ohms to 100K ohms as the norm.
you want this impedance to "see" a higher impedance ,usually on the order of ten to one,so it does not load down the signal of the CDP and cause it to clip from there not being enough power to drive it .The lower the impedance to be driven the more power you need to drive it and most commercial products are not up to driving a low impedance stage.
So we want the volume control to be approx ten times the value of the output impedance of the cd player.
cool this is easy.
but this volume control must also drive the power amp input impedance so we can only go so high with the volume control value if we are to maintain the ten-to-one ratio !
and not just the amp in but the cable must also be considered as it has its own impedance also .The longer it is the higher this goes and finally you reach a point where you start to lose high frequencies due to the combination of the resistance (impedance) and capacitance of the volume control,cable/amp in acting as a low pass filter which as the resistance increases the cut off slides lower and lower until it is audible.
So you keep the cables as short as you can and hope for the best.
The most common passive volume control value that fits the majority of situations is 10K ohms which is high enough for most output stages to drive but low enough for the power amp to be driven reasonably well.
Not an ideal situation and as with most things careful matching and system requirements come in.
The simple fix is to add a buffer to the output side of the passive volume control.Since a buffer has a very high input impedance you are no longer limited to a 5 or 10 K ohm volume control value but can now raise this to say 100K which is better for the CDP to drive ,the higher the impedance "seen" the less current drive you need.
This 100K pot now drives an input that has a value in the megohm range instead of k-ohm range so this problem is eliminated.
The buffer itself has no voltage gain so 1V in = 1V out but what it does provide iscurrent gain and a very low output impedance.This low output impedance,usually in the single digits or at worst around 10 ohms ,is capable of driving not just the imput of any amp but that with very long cables ! 20-30 feet is nothing with the proper buffer.
Low impedance and high current overcome this by providing the muscle to actually PUSH the signal along the resistance (think about the word "resist" and then consider resistance equals impedance),overpowering anything in its way to get the signal to its final destination.
This method is used for long runs in audio ,pro sound,video,and even in your home network so the signal does not become attenuated or lose frequency integrity.
for headphone use it is not long runs we are concerned with but the actual impedance of the headphone transducer.A high impedance like sennheisers (600ohm commonly) do not put much of a strain on the driving force,the amp,but being that they are not the most sensitive headphones out there they do require voltage gain in most cases to step up the signal.I see many using multiloop buffered headphone amps for such headphones with both an opamp voltage gain stage and a buffer current gain stage but for me personally this makes little sense unless I plan on running fifty feet of headphone cable or i plan on running multple headphones from one amp.
In the case of multiple headphones you decrease the impedance by half when you parallel a second pair so the original 600 ohms becomes 300 and adding more pairs lowers the impedance even more.
Being the KISS type of person (Keep It Simple Stupid) I use only as much "in circuit" as i need for the application and no more so for high impedance can i mostly forego the buffer stage and just concentrate on getting the best sound i can from a gain stage.
But headphones such as those made by Grado which at 32 ohms present a very low impedance to the amp need current ,the muscle to "push' the signal or they will not perform correctly and more often than not overload the output stage of a headphone amp not up to the task.But even though they are current hungry devils the grados are also very sensitive so voltage gain is not a concerne and in fact can sometimes be eliminated all together.Again the K.I.S.S principle of using what i need and no more means i sometimes use a straight buffer stage with NO gain stage for driving low Z cans.
High impedance headphones : simple opamp voltage gain stage
low impedance headphones : simple monolithic or discrete current gain stage/voltage follower (or as it is known a buffer)
If you are a one headphon kind of person then designing the optimum for your chosen cans is really simple and can provide amazing synergy .but being who we are and never happy most here have an entire collection of headphones so a "one size fits all" comprimise is attempted and that is the opamp/buffer headphone amp .
unless it has a low gain factor the volume control range is severly limited to the bottom end of the rotation but if the gain is too low then it will not play loud with cans that have a low sensitivity-comprimise !
If the current drive is too low then your low impedance headphones will not be optimally driven to sound as they should but if you have high impedance headphones that do not require the additional current drive and with it the additional stage the additional circuitry is actaully a negative and not a plus-again comprimise !
Anything designed to handle all situations is and must be a comprimise when it is compared to something designed with a specific end use and matching device
It is like designing a car for the drag strip and expecting it to have good highway or around town performance.It may work and eventually get you there but you would be better off with a car designed specifically for the task it is called on to perform
just my long winded opinion ,means nothing
