Question everything: Why 9V in the first place?

[rezzonics]

I like to question everything and this is kind of a philosophical question. Maybe things that made sense long time ago don't have sense anymore, technology evolves and improves, maybe things can be made differently now.
My question is: Why 9V in the first place as main power supply on guitar pedals? Why not a lower voltage, say 5V or 3.3V?
Why some pedals even want to double it to 18V?
I guess it's historical, maybe related to the availability of 9V batteries, maybe old transistors and specially op amps required this level of voltage to properly work.
But today there are good audio opamps with low voltage supply, low distortion and rail-to-rail input/output.
The output of a guitar is just several hundred millivolts and in any case usually less than 1Vpp.
Guitar Amps input stages usually are biased in such a way that its input will saturate over 1V.
The goal of using big gains in amplifiers is mostly to generate saturation and clipping.
Why do we want so much dynamic ra0nge?
The only reason I see to use a wider range of voltages is to be sure that saturation happens at the input stage of the guitar amplifier. Is that the goal of overdrive pedals?
But I think that even with a 3.3V supply an input stage could already be overdriven, since saturation happens over 1V.
Another advantage of using a lower power supply voltage in pedals is battery saving... at least if an efficient switching regulator is used instead of a linear regulator.
Frequency switching in today's step-down regulators is so high that even if ripple is not properly filtered it would be inaudible.
Using a step-down regulator a guitar pedal could work even if battery voltage is lower than 8V.
Could I build a guitar pedal boost or overdrive pedal at a much lower power supply voltage than 9V? Will it work OK? What issues could I face?
Could I build a Klon Centaur Clone at 3.3V? Would it sound similar?

[garcho]

short answers:

9V used to be called transistor battery. you're on to something there, plenty of info online about that. at this point, 9V has been standard for so long, and so many people use DC adaptors plugged into the wall, that to make a (commercial, mass-fabbed) pedal that takes a different voltage is bad for business.

lower voltage doesn't *necessarily* mean longer life, life in batteries is about current supply

yes, you can get away with a lot regarding ripple frequency, but you'll find, sometimes, there be dragons...

you certainly can use less than 9V, in fact some famous pedals as well as celebrated DIY projects do just that. check out Tim Escobedo's "circuit snippets"

rarely, can you simply drop voltage down, especially from say, 18 to 3.3, and have everything turn out the same.

others can reply in greater detail, i'll have more time later if you're curious. best thing to do? breadboard and experiment!!!

[rutabaga bob]

Welcome to the forum!  There are schematics of some fuzzes that run on 1.5 or 3 volts...the Heathkit TA-28 is one.  Use the forum search for TA-28.
Cheers!
Larry

[Mark Hammer]

A 9V supply can provide enough voltage swing to do a lot, and maintain sufficient headroom for audio.  Other voltages are generally physically awkward to provide.  I suppose 6V or 9V or 4.5V could have been provided, but AAA cells and "coin" cells are comparatively recent innovations, as batteries go.  You can find 12v batteries that are around the size of an AA, but they don't provide much current, hence shorter lives.

Somebody could have produced a 4.5 or 6V package, using the same slugs or cells found inside of a standard 9V....but they didn't.  So guitar pedals adapted to 9v, although many have used a pair of 9V batteries in series.

It's generally easier to simply start out with a higher voltage than required and regulate downward.  A 9V battery regulated to 5V can last a while until it ceases to function.  A hypothetical 5V battery would stop being useful as soon as its voltage dropped.

[Frank_NH]

I would also think that the signal to noise ratio is improved for line level signals using higher source voltages.  That said, as mentioned above, there are some old fuzz designs that use lower voltages.

[PRR]

Why not 9V?

I'm working on an idea which "could" use "any" supply voltage. Similar gizmos have used 6V to 300V (different designs!). One constraint for me is that the devices I like may only be good for 16V. Another is that the devices it must drive have a legacy of first 48V then 24V and +/-15V power supplies (though they work at lower levels). All of this gets thrown on the Drawing Board to see if it gels.

A killer advantage of 9V is that when you suspect power supply crap getting in, you can put in a still readily available transistor radio battery and see if it gets better.

The problem with AA/AAA cells is they need two contacts each, and a holder, and you (usually) need more than one. Yes, multi-AA(A) cell holders are standard items, except they come in 2 4 and 6 cell, AA and AAA, long or wide. A dozen variants, and the one you want is never handy or stocked locally. OTOH the 9V clip does the job and is so cheap you usually have a 10-pack on hand.

[rezzonics]

What I was actually suggesting is keeping the external input power supply at 9V since it seems to be the standard, but making an efficient step down conversion to a lower voltage, for example 3.3V, that will actually power all the electronic circuits at this lower voltage, having the same behavior even when the battery goes under 8V.
The question is: Do the electronic circuitry really needs to be powered at 9V for audio related reasons like to keep a larger dynamic range or, by selecting the right components, could I run them at lower voltage? Would this even work for boost and overdrive circuits?

[samhay]

If you primary motivation is to reduce current consumption, then regulating 9V down to a lower voltage might make sense if you use an efficient DC-DC converter - a linear regulator will just get warm. This obviously comes at additional complexity and expense, which is hard to justify if most users will plug into a wall wart rather than use batteries.

I would have thought a follow-up question may be - seeing as most people use a DC supply rather than a battery, why don't we use a higher supply voltage (for better headroom) or adopt the 5V standard so that we can run off USB?

[PRR]

We need near 1V for LOUD guitar.

Ideally this says 2.828V supply would do.

Most all-purpose chips lose ~~1V each way. Already we are near 5V.

Most old simple crude circuits are lucky to get a peak of 20% of supply. So 1.4V peak needs 7V supply.

A 9V battery can go to 7V fairly quick.

I don't see a big problem feeding 9V to a chip "only" doing 1.4V peak with high-Z loading. Yes there is power waste, but too small to matter; and it leaves some leeway for battery fade.

Allowing zinc batteries to fade gracefully is important. Nearly half the total energy is still in there at 1.1V per cell (6.6V per "9V"). Being able to sag WAY down without rude clipping greatly extends life per battery.

Switchers are relatively new and expensive (for pedal design culture). Whine is problematic. Conventional switchers will drop 9V to 5V (or boost 3V to 5V), but get marginal when in and out are nearly equal. Switchers ARE often used in commercial pedals, where one design effort covers millions of sales units. Makes much less sense in 1-off or even small Boutique work. Keep it simple, get the design done and produced, let the user deal with it.

Yes, wall-power supplies eliminate battery fade. And mostly makes "efficiency" moot, since wall-power is CHEAP (at our 0.1 Watt 3hrs/day scale). The answer is: many voltages confuse users, which leads to blown-up pedals, unhappy customers. The industry is finally centering on common-negative 9V power for almost all uses not big enough to require a dedicated wart.

[Transmogrifox]

One of the gotcha's with switchers is not the switching frequency but control loop response, stability, noise.

In digital circuits I often see a certain amount of audio frequency stuff happening because of a certain loading condition making it marginally stable -- produces a kind of low-level limit cycle oscillation at audio frequencies.  In one switching power supply I saw a ~1kHz oscillation (right at power supply crossover frequency) happening on the boundary of inductor operation mode.  It was a noise artifact due to synchronous rectifier timing and this was ringing the under-damped control loop.  It never went out of control but it was a low-level periodic noise event that would have been audible if I had attached it to an audio circuit.

Well-designed switchers don't do this, but be prepared for debugging this kind of thing if attempting to apply a reference design on your own. 

This kind of thing is pretty nice though if you don't mind the cost and soldering SMT:
http://www.digikey.com/product-detail/en/texas-instruments/LMZ21700SILT/296-38544-2-ND/5034141

[Rixen]

If we imagine a world where the guitar pedal has just been invented, with no precedence, it would probably have a single rechargeble 3.7V LIPO cell, USB charger and DC-DC converters to achieve whatever higher rails are needed.. like many other small portable electronic devices.

There were good reasons decades ago for *standardising* on 9V. Not all these reasons remain valid, but with old pedals never becoming obsolete it's unlikely to change soon across the (pedal?) board.

This weekend I repair my TS808 clone which someone plugged a centre positive supply into... (and I will install reverse polarity protection)

[R.G.]

The reason for the 9V battery is historical, as you guessed. The first pedals to make "fuzz" ran on one of the 9V "PP3" battery size made common in quantity by the first shirtpocket radios, the so-called "Japanese radio" in the early/mid 1960s; or they ran on one or two carbon-zinc cells, at 1-1.2V or 2-2.4V. Fuzz was not expected to sound clean, so low headroom was not an issue.

Having some pedals on single 9V batteries became the standard because there are somethings the lower multiples of one carbon-zinc cell are not enough voltage for. And having more than one battery is s PITA for mounting and replacement as Paul noted. By the way, battery voltages are semi-quantized by the electrochemistry of the cells. Carbon-zinc has one voltage per cell, lead-acid (car battery) has another, and lithium-ion yet another. 3.?, 6.3, and "12V" were standard-ish because those were the voltages available from the stacking of small numbers of carbon-zinc, the dominant battery cell back when.

To answer some of your other questions, yes, it was for headroom, partially. For things that went beyond distortion, you had to do guitar cleanly, and 3-4V is not enough. Nominal single-coil pickups are about 100mV after the pick attack, but the pick attack can hit a volt, so it would be some "grack" in front of notes with lower supplies. Worse, the humbucker pickup has lots more output, and hot humbuckers can put out several volts of signal.

Even with a 100mV signal, you can't amplify things much in the range of a 9V. Very little gain can be had before bumping into the power and ground rails. There are some subtleties, too, in that historically opamps could not approach the power supplies closely, distorting when their output got closer than 1-2V of either the most positive or most negative supply. Yes, there are rail to rail opamps, but they tended to not be as pleasing sonically as the standard types, and sometimes to cost more.

That "cost more" is a big, unseen elephant in the room. If it costs too much, it's not going into a pedal. This is a little appreciated fact, and it's not just manufacturers trying to be cheap. There's a long dissertation that could be written on this.

But on to your clarified questions:

What I was actually suggesting is keeping the external input power supply at 9V since it seems to be the standard, but making an efficient step down conversion to a lower voltage, for example 3.3V, that will actually power all the electronic circuits at this lower voltage, having the same behavior even when the battery goes under 8V.
The question is: Do the electronic circuitry really needs to be powered at 9V for audio related reasons like to keep a larger dynamic range or, by selecting the right components, could I run them at lower voltage? Would this even work for boost and overdrive circuits?

The late Yogi Berra was a good guy with a quip. He is reputed to have said (and it may predate him) "In theory, theory and practice are the same. In practice, they aren't." That is particularly appropriate here. In theory, any size audio signal can be amplified up to whatever size you need it. In practice, real components and noise floors mean that much below 100mV gets really tricky to keep quiet. In practice, the amounts of hum and intercepted RF get amplified up with a small signal. In practice, guitar signal is a high impedance signal that picks up noise, hum and RF easily, not like low impedance stuff in a studio. Analog engineers get positively twitchy when you suggest going to lower signal levels and amplifying them back up.

Yes, you could do your audio with a 3.3V supply, and many modern digital chips do that. In practice, that's not going to be popular with self-taught pedal hackers who don't have the background to deal with the consequences. And you probably don't want to use a step-down converter. Switching power supplies run on full-input-power level voltage steps that they mess around with and smooth back out. Make one layout mistake (or fifteen of them at the same time) and the power supply bleeds into the delicate, low-level, high impedance audio. Sure - it's entirely possible and maybe even practical to walk between the 20th floors of adjoining tall buildings on a tightwire, but not many people are going to want to do it.

I work for a company that makes pedal-specific power supplies and we often speculate on where the market is going. The 9V battery is about done as a standard. 20 years ago, there was no way an external pedal power supply could be successful in the market if it was NOT 9Vdc. I can't tell you how much has been written and argued to get to where people see 9V as maybe not needed, and feel free to go putting 12, 15V, 18V, even 24V on a pedal designed for 9V because they read on the net that it's a great sounding mod. Never mind that it may have 10V or 16V caps in it.    But as we say here in the south, bless their hearts. I hope the pedals survive, and we now offer supplies with a plethora of output voltages specifically for this need.

In the future, pedals will be designed for non-9V. But I suspect that they'll go to 12V or 18V or all of the above, not 3.3V, at least for the near future. 9V is no longer de rigueur even today.

[thermionix]

I don't know why the others above are so reluctant to tell the REAL story.

6V batteries used to be quite common as many tube filaments ran on 6 (6.3) volts.  But when Jimi Hendrix invented guitar pedals, he was in the process of writing the song "If 6 was 9" and the idea just carried over from that.  I read it on the internet just now, so it has to be true.

[PRR]

> 3.?, 6.3, and "12V" were standard-ish because those were the voltages available from the stacking of small numbers of carbon-zinc

[pedant]6.3 is three 2.1V lead-acid cells, the standard car battery in the US until hybrid radios and huge engines and alternator diodes pushed us to 12.6V (where the UK already was). "6.3" appears a lot in vacuum tube specs because car-radio became a Big Fad in the 1930s, and new 6.3V-heat tubes were produced (and became the standard voltage). Before that were a fair number of "2V" filaments (not heater-cathode) because you could borrow one cell from the farm light windmill storage battery to light your radio. [/pedant]

> 6V batteries used to be quite common as many tube filaments ran on 6 (6.3) volts.

All the 6.3V heaters I know are PIGs. You don't really want to run them "on battery". In a car, despite a heavy battery, you always started the engine if you were gonna listen more than a minute or three. (If you ran the radio too much engine-off, you could not start the engine....) (Less an issue for 2V tubes on the wind-light battery, because the radios were small, it was one cell out of 50, and the windmill would start without juice.)

I do like the Hendrix 69 theory.

Fresh carbon-zinc tends to 1.56V per cell. With batteries we never care about the extra "0.06" because they sag from the start. However if you see well-engineered regulated supplies speccing 9.3V or 9.4V, that's why-- they want to be sure you have a FRESH "battery", yet run no risk of straining your gear.

Voltages change. Cars used 24V and others before 6V then 12V became common. Military and large trucks have long run 24V systems. I have seen it suggested that as hybrid cars gain share, we may switch to higher "primary" voltage because these systems need high voltage to avoid high current (fat wires) yet still make high power. Incandescent and LED headlamps won't like 300V battery voltage, but we have the technology to drop for these "small" loads. Domestic electricity started in the US with 100V to reduce electrocution (no good rubber yet), but had to lean to 105V to cover line losses. AC systems reduced loss but the economy of higher voltage is compelling. UK (after good insulation) settled on 240V. US sticks with 125V because of our 100V legacy. (And we do run 240V to larger appliances.) I know a guy researching solar houses who thinks our homes should have 48V and 96V DC buses. Computers once had odd arrays of odd voltages of both polarities (+26, +8, -13), IBM PC had +12 +5 -5 -12 (one of these for a DRAM technology which obsoleted at the same time), and now everything comes off 3.3V but dropped to 1.1V for the energy-critical core. So change is possible. I guess.

[Mark Hammer]

I recall RG posting some time back about the relationship between the width of oxcarts in Roman times and the scale of contemporary rail-cars; the former being the precedent that set compulsory road widths, which ultimately set the scale of roads, which became co-opted for installation of rails.  Whatever the true sequence of events, the basic premise is that early standards have a way of limiting the landscape for subsequent changes.

I will suggest that the early establishment of "line" levels in amplification, as being around 1V RMS, set the stage for what amplifiers, including those for guitar, were designed to anticipate.  If amplifiers were designed to achieve maximum output with an input signal of 100mv, the 9v battery business wouldn't be as big as it is. 

Once the expected signal amplitude became established, that pretty much dictated what sort of supply voltage would be needed for the things that fed it, which in turn dictated what sort of battery supply could meet the requirements in convenient form.  Where simple germanium transistors were used, one or two AA cells could suffice.  But as op-amps became more typical, the 9V was the one-size-fits-all solution.

To newcomers whose merging with the pedalboard world begins after the emergence of external power bricks, that may seem foreign.  But that's not only why the subway car is the width it is, but why pedals are the size they are.

[bluebunny]

Wait, pedals are the size of an ox??      I think I'm doing something wrong...   

[R.G.]

I recall RG posting some time back about the relationship between the width of oxcarts in Roman times and the scale of contemporary rail-cars; the former being the precedent that set compulsory road widths, which ultimately set the scale of roads, which became co-opted for installation of rails.  Whatever the true sequence of events, the basic premise is that early standards have a way of limiting the landscape for subsequent changes.

...

To newcomers whose merging with the pedalboard world begins after the emergence of external power bricks, that may seem foreign.  But that's not only why the subway car is the width it is, but why pedals are the size they are.

Yep. In recent history, one of the ongoing miseries of the pedal world that cannot now be changed is that nearly all pedal power supplies must supply a DC barrel connector that is 2.1mm ID, 5.5mm OD, and center-pin negative. We do this because Boss was the first to install DC power jacks in many, many pedals and provide the PSA adapter that supplied them. Others copied, until now we simply cannot go back and correct this to the center-pin positive that almost the entire rest of the world uses. For commercial suppliers of pedal power supplies, it is death to NOT support the center-negative history, so new pedal makers make their input jacks center-negative so the existing pedal power supplies work without changes.

In the computer world, this would be a software "deadly embrace" that can't be changed except by killing both processes. Pretty much same here. All pedal owners and pedal power supply owners will simply not change all their stuff to make it be "correct" somehow.

[Mark Hammer]

Wait, pedals are the size of an ox??      I think I'm doing something wrong...   

Yes, you forgot to carry the 1. 

[Mark Hammer]

Yep. In recent history, one of the ongoing miseries of the pedal world that cannot now be changed is that nearly all pedal power supplies must supply a DC barrel connector that is 2.1mm ID, 5.5mm OD, and center-pin negative. We do this because Boss was the first to install DC power jacks in many, many pedals and provide the PSA adapter that supplied them. Others copied, until now we simply cannot go back and correct this to the center-pin positive that almost the entire rest of the world uses. For commercial suppliers of pedal power supplies, it is death to NOT support the center-negative history, so new pedal makers make their input jacks center-negative so the existing pedal power supplies work without changes.

Well, I would say yes and no.
If we were operating in a world that eschewed batteries entirely, I would say yes and agree with you on your assessment of Boss/Roland's dead weight.

But there aren't many 2.1mm jacks where the contact to be dislodged to switch between internal and external power can be moved by the inside of the plug.  It could certainly be engineered, but the traditional design of such plugs is that the shaft of whatever is inserted bumps a leaf contact out of the way to provide power from the outside rather than the battery on the inside.  Where the power always comes from the outside and nobody has to disable any battery, it does become the manufacturer's choice to have outside-pos or outside-neg.  But in this instance, though it was certainly Roland's choice to use that style of jack, it was not necessarily their choice to have the normalized/disabling contact be the one touching the shaft of the plug.

[PRR]

> In the computer world...

For some decades, one 3-letter company dominated the mainframe racket. My father worked for another 3-letter company (one of the 7 dwarfs). All they could do was make side-gear (tape, punch, print) which exactly emulated the Big Company's mainframe plugs, signals, and voltages. "SDB"s tape drive plugged-in same-as "HAL"s tape drive, but was priced different.

A later decade saw many new companies "cloning" the IBM PC's internal and external signals, voltages, and plugs.

(Except even IBM did not fight the dominance of Epson's variant of the Centronics printer port or the AT&T serial terminal interface. Sometimes it is all about who got there first with a good-enough system.)