ESP Logo
 Elliott Sound Products Project 195 

Guitar 'Talk Box'

Copyright © September 2019, Rod Elliott


The guitar 'talk box' [ 1, 2 ] has been around for a long time, but it was made famous by Joe Walsh on 'Rocky Mountain Way' (1973), and it was also used by Peter Frampton, Aerosmith, Bon Jovi, and many others.  'Rocky Mountain Way' was the first time most people had ever heard one being used, and I was forced to build one many, many years ago when I was the sound operator for a number of bands.  Of course, they wanted one, so I built a few.  This was well before the internet, so it was basically a matter of using 'first principles', and seeing a video of it being used (or maybe it was on TV - it's way too long ago to recall).  The basic idea is very simple ... send the guitar amp's output signal to a suitable enclosed driver, hook up a length of plastic tube which is taped to the guitarist's microphone, and add some switching to divert the amp's output from the speakers to the 'talk box'.

The end of the hose goes into the guitarist's mouth, and as his/her mouth's shape is changed by 'mouthing' the words (no singing is required), the tone is changed.  The effect is to make the guitar 'sing', although it takes some skill to make the 'words' intelligible.  The hardest part is mouthing words normally with a bloody great plastic tube in your mouth (ok, it's not that big, but it still takes some getting used to).  We don't hear them much these days (at least nothing I've heard for some time), but it's still a great effect and can suitably impress the punters at live music events.  There are many other uses as well, and it can be used with keyboards just as readily as with guitar.

Only one driver is suitable, and that's a horn compression driver (ideally a phenolic diaphragm type!).  The amp's output needs to be attenuated so you don't need a 200W (or more) compression driver, and a push-on, push-off footswitch does the change-over.  The circuitry is dead simple, and essentially consists of a few high-power resistors, a capacitor, the switch and a compression driver.  The tricky part is the interface between the driver and the plastic hose, which is a mechanical problem rather than an electronic one.

Given the high sensitivity of most compression drivers (typically around 110dB/W/m), you don't need a great deal of power.  The full output from a typical guitar amp is nearly enough to shake your teeth out unless the signal is attenuated.  It's generally likely to be fine with somewhere between 5A and 20W into the driver - any more risks damage to the driver or the player's dentition.  (And yes, I am serious.)

Basic Circuit

The circuit is shown below.  As you can see, there really isn't much to it, but the resistors acting as the amp's load must be rated for close to the full amp output power.  This makes it a bit of a nuisance, due to the need for good cooling because the resistors will get very hot if used for any length of time.  The resistor feeding the compression driver can be 'select on test', but as shown it should be alright with a typical 50 - 100W driver, driven by a 100W amp.  Using the amp connected directly to a high power compression driver isn't recommended, and operating the amp with a high load impedance isn't a good idea either - especially with valve (vacuum tube) amps which can be damaged if the load impedance isn't within an acceptable range.

Figure 1
Figure 1 - Circuitry For A 100W 'Talk Box' & 50 - 100W Compression Driver

The circuit is configured as a 10Ω load, and this will still provide enough loading to prevent damage with a valve amp set for 4Ω, and it reduces the power requirements for transistor amps that will usually drive a 4Ω load.  That doesn't mean that you can skimp on the power rating though, and I chose 10Ω resistors deliberately because you can simply use an array of 10W resistors bonded to a heatsink to achieve an acceptable power rating.  Eight resistors are used, in a series/ parallel circuit with a total resistance of about 10Ω (including the driver), and a total power handling of 80W.  When attached to a heatsink, this is improved to around 150W (depending on how well the heatsink works), but if used at full power for any length of time, they will still get very hot!

Sw1 is the footswitch, and must be a push-on/ push-off type.  In the position shown, the 'talk box' is isolated, so the amp's output goes directly to the speakers.  When switched, the speaker is disconnected, and most of the power is dissipated in the resistors.  Sw2 is provided to allow the use of the circuit with either more power to the compression driver, or for use with smaller amplifiers.  You will likely have to experiment a little to ensure that you get enough signal at the end of the tube and don't overload the driver.  When switched to the 'High' power mode, the -3dB frequency is nominally 400Hz, vs. 260Hz in the 'Low' power configuration.

The vast majority of the project is mechanical, rather than electrical, and there are no 'electronics' involved.  You can (of course) use a separate small amplifier to drive the compression driver, but that adds cost, and needs a great many more parts.  Ideally, you'll use a compression driver with a phenolic diaphragm, because you need something that can take some punishment without failure.  The 'crossover' is nothing more than a capacitor, and for this application a bipolar electrolytic will be quite alright.  I normally never suggest these due to their variable characteristics and distortion, but for this it doesn't matter at all.

The suggested capacitor value is 22µF, which will roll off the bass frequencies.  Note that the capacitor must be one designed specifically for crossover networks, because it's expected to carry significant current.   If at all possible, I recommend using a polyester or polypropylene capacitor, such as those designed for starting and/or running electric motors.  Bipolar electrolytic caps are incapable of handling significant current for extended periods.  As shown, the -3dB frequency is about 260Hz ('High') or 400Hz ('Low').  Both are somewhat lower than recommended for most compression drivers, but a high power driver with a phenolic diaphragm will survive.  Any driver with aluminium or beryllium diaphragms will probably self-destruct, because they are far less resilient and therefore more easily damaged by low frequency energy.

Note that many of the 'alternative' circuits on-line show direct connection to the compression driver (in some cases a midrange driver is suggested).  While this certainly will work, the issues of excessive SPL in the performer's mouth and likely damage to the driver if driven by a high power amplifier make it somewhat less attractive.  By all means try it if you are 100% certain that the driver can handle the output from a 100W guitar amp (most can't, because they are fed with far too much low frequency energy).  Commercial units may use a direct connection, but power handling is severely limited.

Remember that compression drivers are very efficient, and while distance (in metres) isn't even relevant here, there is some attenuation along the length of hose.  You'll probably get enough sound energy into the musician's mouth with as little as 5W into the driver.  Anything above 50W is likely to create excessive SPL in the player's mouth, making it very uncomfortable to use for any length of time.

Construction Ideas

You need to devise a method for connecting your plastic hose to the driver, and how you go about that depends on what you have available.  Threaded drivers are common, but these are harder to use for this due to the requirement for a female threaded assembly to adapt the hose.  You can just cut off the threaded end of an old horn if you have one handy (or can get one cheaply), and the hose adaptor (i.e. a short length of metal pipe that just fits into the hose with some persuasion).  If needs be, the hose can be clamped in place with a hose clamp (what else ).  The pipe can be glued inside the end of the 'adapter' you've made with epoxy (make sure it doesn't run down into the thread, or it won't screw onto the compression driver).

For 'bolt-on' drivers, a flat plate with a short length of metal pipe (as shown below), which can be welded, brazed, rivetted, soldered or even glued into place (with epoxy of course).  If you can get the hole for the pipe just right, it can even be a force fit.  As long as it cannot be dislodged during use, use the method you are equipped for - I wouldn't expect a prospective builder to buy a MIG welder for one small project.

Figure 2
Figure 2 - Suggested Attachments To Compression Driver

The two ways that compression drivers are mounted (to horns) are a flange or a threaded adapter.  The above drawing shows exploded views of an adapter to suit either method.  If you use a flange mount driver, I suggest that you install a gasket between the horn flange and the adapter plate (remember that it need a hole in the centre to suit the driver's throat opening).  Don't over-tighten the mounting bolts and ensure the seal is airtight.  Screw-mount drivers will be trickier if you don't have a 'sacrificial' horn to provide the mounting threaded section.  At a pinch, gaffer tape can be used to attach the tube to the driver, but that should be considered temporary until it can be done properly.  This is ideal for testing though.

While I've shown a connection where the hose is pushed over the fixed tube, you can use a larger fixed tube that the hose pushes into.  If done that way, there is a small reduction of attenuation caused by the constriction, but it may also allow for greater diaphragm excursion, especially at low frequencies.  I leave it to the constructor to decide which method is preferred.  In reality, there is likely to be little difference.

The hose itself needs to be of reasonable diameter, with somewhere around 12mm inside diameter (approximately 16mm outside diameter).  You need sufficient length to reach from the 'talk box' to the mic with a bit to spare - around 2 metres is a good starting point, and you can cut off the 'messy end' a few times before it becomes too short.  You will need to have separate hoses for each user (assuming multiple people use the same unit, but obviously not at the same time).  No-one wants someone else's saliva!

While on the subject of saliva, the top of the hose should be mounted above the microphone, so it points down.  Otherwise, saliva may run down the tube into the compression driver.  Not only is that likely to be pretty gross to remove, it may damage the driver (especially rust on the steel polepiece and/ or a build-up of 'crud').  I doubt that many people will find removal of dried stray mouth fluids to be a pleasurable undertaking. 

The whole unit can easily be housed in any suitable case, and that's up to the constructor.  If you don't really feel like installing everything into the case, a suitable compression driver will be heavy enough to stay put by itself, and it only needs a lead to plug into the switching (and attenuator) box.  Since the circuitry is only simple, the only reason that a 'decent' sized case is required is to enable stability, heatsinking and/ or ventilation for the power resistors.

Figure 3
Figure 3 - Resistor Mounting Suggestion

The above is one way to mount the resistors, with heatsink compound used between the resistors and heatsink.  If you use the 'cement' type wirewound resistors, make sure that the 'in-filled' part of the resistor doesn't face the heatsink or mounting bracket/ bar, as the dimension is not controlled and pressure will be uneven.  This will cause uneven heating, possibly leading to resistor failure.  The method you use will ideally use whatever you have to hand, rather than buying stuff especially (other than the essentials of course).  Ultimately, you may not need a great deal of heatsinking at all, especially if you run your amp at less than full power by using the master volume so it distorts at lower power.  To get the best from a 'talk box', you will definitely want to use distortion, because it gives a greater harmonic output that makes 'speech' clearer.

There's no reason that you can't use an alternative resistor arrangement, based on what you can get easily or may even have to hand.  The arrangement shown is only a suggestion, and you can make changes to suit the compression driver you intend to use.  You must make sure that the driver can handle the power - a low power driver will fail if subjected to the full output of a guitar amplifier.

Typical Connection

The 'Talk Box' is inserted into the speaker output from the amp to the speaker box.  The amplifier's speaker output goes to the box, and the box output goes back to the guitar amp's speaker box.  The hose from the 'Talk Box' is positioned close to the microphone, which connects to the PA system just like any other mic.  The drawing below shows the general arrangement.  For the hose itself, you can use 'proper' surgical piping, or just a length of PVC tubing.  Make sure that it can't kink during use, as that will cut off the sound almost completely.  The most common way to get the hose in the right position is gaffer tape - it's not pretty but it works.  Remember - mount the hose above the microphone, pointing down.

Figure 4
Figure 4 - Typical Connection For Guitar

In use, the guitarist (or keyboard player) places the hose inside his/ her mouth, positioned so that 'normal' mouth actions as used for singing can be performed comfortably.  It takes practice to get the hose, microphone and mouth in the right places respectively, but once it's been rehearsed a few times if shouldn't be too hard to get right.  Quite obviously, this is anything but a 'normal' procedure for any musician, but the fact that it's been used by so many performers indicates that it can be done.  As always, 'practice makes perfect'.

Needless to say, the techniques used will vary widely between different performers, and I'm not about to provide any specific details.  This is a very personal choice as regards style and application, and everyone will use it slightly (or very) differently.  There is no doubt that the effects obtained can be very good indeed, depending on the musician's abilities of course.

Ideally, the speaker connections will all be Speakon types (or at least XLR), because jack plugs and sockets aren't a good idea for speaker signals - not that it's stopped any of the major manufacturers from using them.  The Speakon connector is specifically designed for speaker cables, and is far more reliable and less likely to short the amp's output.  This can kill transistor amps, and if the circuit goes open that can spell the end for valve amps.  By using the proper connectors, both are less likely.


While talk boxes of various designs and configurations can be purchased ready made, this project is provided with the certain knowledge that many player prefer to build their own equipment.  Project 27 has shown this without any doubt, as a great many guitar amps using the ESP boards have been built.  While it's not one of ESP's most popular projects, it has still provided several hundred guitarists with the facility to build a great guitar amp for far less than a 'brand name' amp, but without sacrificing performance.  As with P27, this project gives you the option for customisation, something that's not available with 'ready made' versions.

As already noted, there is an absolute minimum of electronics involved here, but there is a requirement for a bit more than average mechanical work.  For those constructors with access to a lathe, it should be easy enough to fabricate a suitable connection between the compression driver and hose, but it can still be done with basic hand tools if that's all you have available.  Cheating is perfectly acceptable - provided you ensure that the end result is sturdy enough to be carried around, and performs as expected.

The compression driver that you use is the key to good performance, and it's worth spending a bit extra to get one that will stand up to the abuse it receives in this role.  As already noted, a phenolic diaphragm is preferable to aluminium or any of the 'exotic' materials that are used for high performance drivers.  The driver will receive significant energy at frequencies well below its design limits, and while the constricted throat and long tube do provide significant loading (even at lower frequencies), compression drivers aren't designed for this, and some may not survive.

  1. Talkbox - Wikipedia
  2. Jim Dunlop - Heil Talkbox (Commercial Version)
  3. Building A Talkbox - General Guitar Gadgets

There are no other references, because while there's actually a fair bit on-line, not all of it is especially useful.  Having built a couple (many years ago) I know what works, and workbench experiments were used to verify performance (albeit with caveats, as I don't have a phenolic diaphragm compression driver to test, so reduced power levels were used).


HomeMain Index ProjectsProjects Index
Copyright Notice. This article, including but not limited to all text and diagrams, is the intellectual property of Rod Elliott, and is © 2019.  Reproduction or re-publication by any means whatsoever, whether electronic, mechanical or electro- mechanical, is strictly prohibited under International Copyright laws.  The author (Rod Elliott) grants the reader the right to use this information for personal use only, and further allows that one (1) copy may be made for reference.  Commercial use is prohibited without express written authorisation from Rod Elliott.
Page created and copyright © Rod Elliott, September 2019.