You want the truth? You can't HANDLE the Truth!! More than you need to know about power supplies
I'll tell ya what: the last fucking thing I want to think about is a power supply. Talk about a glamour-less experience. The best one you can get doesn't do ANYTHING to your sound. Nothing at all. Makes it sound like its not there. Here I am working my ass off to make shit that changes your tone for the better, and the power supply's job is to make sure it doesn't. At all. I should be making power supplies.
If you haven't guess it already, this edition of Straight Jive is going to discuss some power supply theory, the "ideal" power supply, and the harsh realities of life as a musician-- at least insofar as DC supply voltage for your 9V effects pedals. All that other shit, sorry man, you're on your own.
The Ideal Supply
So...
In a perfect world, a 9V guitar pedal (like say, I dunno... a SuperMoon?) would receive a perfect nine volts (no more, no less) of direct current, with an ideal current supply delivering said current. This would be bitchin. We could completely eliminate our power filter circuits (think otherwise? TAKE YO ASS BACK TO CLASS), radically lower the noise floor and deliver a product that would never, ever, ever in 20 million-bajillion-thousand-giga-years (note: a giga-year is bigger than a normal earth year) fail or show any sign of electronic fault. That would be fucking killer.
Sadly, nothing in this world is perfect, but there is a very real beauty in that fact. I like some of my tattoos because the fill goes outside the lines; some spots chose to scar, blow-out and shift as I've aged and worn. Music sounds good on a record, because there's imperfections in the playback and the recording sounds a little different every time you hear it, just like a real, live band. Perhaps we should embrace the shortcomings of our real world power supplies and shun the notion of the ideal supply? Nah. Fuck that. I want good power.
Back to the ideal supply: the closest we get to an ideal supply in everyday practice is a battery. Chemical storage of electrical charge. That's what a battery is, and also why I think hybrid cars are just another disposable razor, but that's a whole 'nother topic.
A typical battery you would use in a guitar pedal is comprised of six individual energy cells, each developing a nominal 1.5VDC across its terminals. When strung together in series, the potential across the terminals simply sum, delivering the nine volts so many effect circuits are based around. This is totally killer, compact, and even relatively in-expensive. BUT! There's a major drawback: chemical degradation. Commonly known as a dead battery. Fuckin bummer man.
A battery is our closest approximation to an ideal supply because it just fucking delivers. I'm not gonna get into the physics and chemistry behind dry-cell theory and operation, there just isn't time for all that, but I encourage you to check it out. It may change your perspective on a few things and nourish some creative ideas!
Back to the battery: it just fucking delivers. A battery kicks out as much current as it can, slowly depleting its reserve storage charge-- read: dying. The beauty in this short-lived flower is its purity: there's no rectification taking place. There's no switching happening. Electrons just line up and march. They march to push, then they march on. Then they die. Lame.
But seriously, think of a battery as the closest practical approximation of an ideal current source you can get your greasy little mitts on, because that's about as good as it gets. At least for our purposes.
Simulating the ideal supply in the real world
As you probably know, the power in our walls is delivered in the form of Alternating Current. Really, there are some advantages to delivered alternating current, but for our lovely little effects pedals, we want direct current! So, we've gotta manipulate the AC and make it DC. Or at least a close approximation to DC. LET THERE BE ROCK!!
I wish I had the time to discuss common rectification techniques and the advantages some systems have over another, but I wanna stay on track here and we'll just look at full-wave DC rectification.
The basic idea is to manipulate the AC supply into only providing positive (or negative depending on the reference point) voltage. It is quite a simple concept really, kinda like primarily using the downward-pressure generated in a reciprocal combustion engine. Of course, there are losses here, but we're humans. If we cared about wasting energy we would probably not be relying on electromagnetism for energy generation would we? If you're really curious about the inner workings of a full-wave bridge rectifier, you should check out Wikipedia's entry on it. Hyperphysics has a pretty good article as well. Read them and change the world.
Once we've got our AC successfully rectified into modulating DC, we've gotta do our best to smooth it out and fill in the gaps (AKA: get rid of as much fucking noise and bullshit as possible). Luckily, we've got some pretty useful circuit elements than can help us with that, namely the capacitor and the inductor. More often than not, you'll find capacitors over inductors, largely due to cost and the physical constraints of inductive circuits. This isn't to say they aren't out there, but they much less common than capacitive filters.
These concepts are the basic building blocks of the familiar, often disdained wall-wart. Everybody has a few shitty ones lying around. You keep thinking it'll work, but its just a noisy piece of shit. Do yourself a favor: get rid of it. Its not serviceable and it ain't gonna fix itself. Lower your blood pressure, your hairline will thank you.
Most wall-warts are just these simple circuit blocks: a bridge rectifier and a simple filter to smooth out the modulating DC into "basically DC." Often times, "basically DC" is good enough for the job. An alarm clock? Totally fine. You're gonna hit that thing as hard as you can and try to break it. It doesn't need a fancy-schmancy super-clean-ultra-linear-power-supply circuit. Just something to get the job done and keep you from being late to work. But our guitar pedals? Well, we need a little more than "basically DC."
Improvements, both inside the box, and outside the box
One of the neat things about guitar pedals specifically is that they are independent from one another, and their respective power supplies. In fact, they're so independent, that lots of guitar pedals have their own power circuits built right into them. Not the type of power circuit that generates power, but they type of power circuit than manages the incoming power. After all, a good manager is key to a successful team! Now I can speak for everyone, but I design the power circuits around here, and I can tell you that careful thought and consideration is given to these deceptively simple circuits.
So even if we're getting a close approximation to DC, how can we improve on it?
Now, it may seem obvious, and for some people the low-hanging fruit of increased capacitance is the natural solution. Lets stop for a sec and talk about why someone would think about increasing capacitance.
I think the line of thinking goes a like this:
We've got dirty power coming in. What to that use to clean up the power? A capacitor!! So if we just add capacitance, we'll clean up that power! Jeez. I'm a great engineer.
Well, kinda. While increasing capacitance does have its place, its not the be-all-end-all method for cleaning up shitty power and ensuring a quiet pedal. If it was, then you'd see huge fucking banks of capacitors in mass-market guitar pedals. It is true that increasing the capacitance of the power circuit will "catch" lots of bullshit on the power rails and further smooth out "basically DC" power, but everything at its cost. Those caps don't ship charged, and they certainly don't hold a charge for all that long (unless the Thévenin equivalent is enormous, in which case you wouldn't care about a power supply because a battery would last two lifetimes). What that translates to is that a bit of current will be used to "charge up" the suggested massive capacitor bank and while it may help provide a layer of insulation against ripple on the positive rail, it also places quite a demand on the supply to deliver a lot of current at initial power-up. If that supply is a battery, well, you just took a big bite out of the life-span of your ideal supply.
And you know what else? There's more.
A huge bank of capacitors makes the power delivery s-l-o-w. I mean d--r--a--g--a---s---s s---l----o------w. Tonally, the effect is a muffled top end, and a stiff, gummy feeling to the circuit. Hey, maybe that's what you want. If so, Mr. Black just broke you off with the secret formula for making dark, murky, molasses-uphill-in-january tone. LOTS OF CAPACITANCE.
The "Daisy Chain"
Probably one of the easiest and cheapest ways to supply power to a number of pedals is the ubiquitous "daisy chain." This is how lots of us end up with isolated supplies. We buy a wall-wart for our favorite pedal because we keep replacing batteries and that shit gets expensive. But that wall-wart only provides power to one of our pedals, and we like to use A LOT of pedals. So instead of buying individual wall-warts for each pedal and a big-ass power strip to plug all those wall warts into, the guy at the guitar shop sells us a daisy chain that we plug our wall-wart into, and then plug each little end of the daisy chain into our individual pedals.
Works great! Except we've got that one fucking pedal that makes a shitload of noise on a daisy chain. You know the one's I'm talking about. And, when we use that polarity inverter adapter (because we need a fucking adapter to use this damned pedal) it wipes everything out and nothing works. Not only that, but we've got this tangled mess of cables, wires, half-smoked doobies, parking lot grime, cigarette butts and all kinds of other bullshit making our board look like something that came out of one of the shitty pawn shops on the wrong side of town. Epic bummer.
To make it worse, for some reason, our amp sounds dull and shitty. Well, this may be the result of excessive capacitance on the power rails.
Enter the realm of Isolated Power Supplies
I personally think reduction in total capacitive load on the power rails is one of the benefits of employing an isolated supply. Maybe this is why they have become popular. There are other reasons as well of course: inverted ground schemes, and of course: NOISE REDUCTION! Right? Well... lets just see how this article shakes out.
Back to capacitive reduction!
When capacitors are placed in parallel, they form a bigger capacitor. Not physically! But functionally. For example, if you have a 100 microfarad capacitor (abbreviated µF) wired in parallel to a 47µF capacitor, the total capacitance in the network is 147µF; 100µF + 47µF. Easy peasy, like greasy fleece. Well, this is one of the shortcomings of a "daisy chain" configuration, especially if you've got a pedal designed with a massive bank of filter capacitors across the power rails.
Isolated power supplies ensure that each power source is separate (or isolated) from one another preventing huge capacitive build-up. Not only that, but you can have inverse polarity between units because the common rail (often called ground) is the only common power rail throughout the entire signal chain (or circuit)-- positive/negative supply is provided individually to each sub-circuit (pedal). Almost as if you had individual power cells for each pedal.
That means your tone is gonna be full, flexible, and not dark and murky and shitty. That sounds like a pretty good thing to me!
But, everything at its cost as you will soon see.
The fun part
This is where things get real. The proving grounds, so to speak. I noticed this really weird thing a few weeks ago:
My good friend Philippe, one of the guys at Caroline Guitar Company, sent me a Maxon OD808. If you don't know about this pedal, it is green, it has three knobs and its got that "classic" overdrive tone. Oh yeah. Its a winner.
So the pedal shows up, and I plug it into my favorite little living room amp: an Orange Tiny Terror. I dial in a little bit of grit and hit the pedal to take it to the top. Well, I'm telling you, I got some fucking righteous tones, but I noticed something: I was getting a lot of extra noise! Could it be the new pedal? I turned the pedal off and the noise went down, but it was still there.
You puzzled yet? I was.
I knew the pedal was a buffered, electronic switching job, but that shouldn't matter, right? Well... I was so puzzled, I decided to figure it out. And I did. Check it out:
I dropped this pedal right onto my Pedaltrain Jr., which is furnished with a pretty bitchin "industry standard" isolated power supply. This power supply has eight isolated outputs, most of which can put out an impressive 100mA each! I have, what I consider to be, a decent little pedalboard setup. No, its not some fancy-ass boutique rig, but its a workhorse with isolated power. Its lightweight, practical and strong. Dude, its solid. But for some reason, I was getting a shit load of noise.
I don't know why, but my first thought was to try subbing in a cheap switching power supply and see if that changed anything. Mind you, I was only running the OD808. No other pedals used. Just guitar, OD808 and amp. Oh, and my fancy-schmancy expensive isolated power supply.
So I hook the OD808 up to the old power supply, a compact, $20 job that kicks out an impressive 1700mA. Would you believe me if I told you that the noise floor went down notably? Well, it did.
This strange behavior left me feeling weird and confused; a state I have become comfortable with, since I'm often confused, and always weird. So I did the most logical thing I could think of: I called Philippe to tell him about my findings.
Philippe and I chatted for nearly an hour and I explained the strange phenomena I discovered. He didn't fully believe me. Shit, I didn't. So I decided I would document my findings and pursue some research. Well, here we are.
I was so intrigued by this strange behavior that I brought the power supplies to the diagnostic bench and took some measurements. I put them bitches on the scope to see what I could see about the noise levels of each, and of course, that of a fresh 9V battery.
Here's what I found:
This is my oscilloscope. It is a 60MHz analog scope. Way more power than you need for guitar pedals, but hey, anything worth doing is worth doing to total excess, right?
This photo is a ground reference, establishing our base-line
I performed a total of ten measurements. Three sets of three, and one final comparison of a fresh 9V battery. The first measurements were taken with the scope configured as below. 20µS / horizontal division, 20mV / vertical division.
A photo of the initial configuration of the scope. The next three photos used these settings.
This is raw AC waveform taken from the isolated power supply's "normal" output.
You can see about 10mV of bullshit sitting on top of the 9VDC output.
This is the raw AC waveform taken from the isolated power supply's "variable" output.
With the variable output disabled-- switch in the "NORMAL" position.
A little more bullshit than the "normal" output.
This is the raw AC waveform taken from the switching power supply's only output. It seems like it has a little less constant noise than the isolated supply, but you can see the spikes when the power supply switches. If my scope is dead nuts on, we've got a surge happening every 80 µS (about four vertical divisions between each spike at 20µS / division). Based on that figure, we can determine that we could have a small 12.5KHz signal sitting on top of our positive supply (0.00008S ^ -1, or 1S / 0.00008S).
In an effort to improve detail and really hone in on my findings, I increased the resolution of the voltage measurement. The scope has been reconfigured to display 10mV / vertical division as seen below.
A photo of the scope settings for the next series of measurements
"Zoomed in" a little, we can see that 10mV bullshit sitting on top of the isolated supply "normal" output even better!
Here's the "adjustable" output on the 10mV / division scale. You can see there's a little more noise than the "normal" output, eh?
Here is our switching power supply. Its actually looking like its got a little LESS noise than the isolated supply!
Minus that switching mark of course.
To really hone in, I took my scope to its most sensitive setting: 5mV / vertical division. For a little perspective, it would take 20 vertical lines to display one volt. One fucking volt. This scope can display 0.04V (40mV) across the entire screen at this setting. Here's a shot of the control panel for reference:
We are now on the 5mV / division setting: the most sensitive this little scope offers.
This is the "normal" output of the isolated supply. It is now really easy to see that' we've got nearly 10mV noise/bullshit sitting on top of the DC supply voltage on this output.
Whoa, baby! This is the "variable" output on the isolated supply, and I'm telling you, that is A LOT of noise.
Obviously, this is the switching supply at 5mV / division. It is actually generating less ambient/background noise than the isolated supply, BUT! it has that pesky switching spike every 80µS or so.
Now this info is all well and good, but we really need a reference dontcha think? I did. So I measured a fresh 9V battery with the scope on its most sensitive settings. Here is the result:
Even a battery exhibits a little bit of noise. A lot less than the measured power supplies, but still a little.
Just goes to show you nothing in this world is "perfect."
Well pictures are all well and good, but seriously, don't you wanna know just what this means to YOU AND YOUR AMP? I sure did. So I recorded each configuration. I even converted them into MP3 format so you can stream them live or download them for yourself to listen to in your reference library. You do keep a reference library, don't you?
All the sound clips were recorded as such:
Guitar (bridge pickup volume all the way down) --> Maxon OD808 (all knobs at noon) --> Orange Tiny Terror (settings below)
Recorded in Adobe Audition CS6 with a plain-jane Shure SM57
Well, you can draw your own conclusions here. Obviously, diming the amp increases the noise level substantially, but it also really brings the noisiest situations to the foreground.
I learned A LOT through this experiment and process, and I think I'm gonna be switching back to a cheap switching supply and selling that fancy isolated supply, since it doesn't really offer much to me and my situation.
I hope this lengthy edition of Straight Jive has been informative, useful and fun for you. It sure has been for me.
Until next time, peace, love and hair grease!
-Jack