How important is power filtering with DC?

[tehfunk]

I'm modding my wah right now, adding a 9v DC input, and have talked about power filtering here before, now I just want to know, how much of a difference does it make? How necessary would you say it is? Are there any audible benefits? I guess that's what's most important, am I degrading the sound of my setup by not filtering? Last do you use just a large electrolytic with positive on +9v and negative to ground, or do you need a resistor in there too? Thanks.

[George Giblet]

It depends on your power source.

If you power it from an unregulated wall-wart the ripple, despite being small,  gets through to the audio.   Transistors circuits, like the wha,  tend to be quite poor at rejecting hum on the power rails.    If you used the same unregulated supply to power the wha and other effects the situation is much worse.

If on the other hand you use a regulated supply you are fairly safe in not using a filter.    For recording you really should be using a regulated supply.

In practice there is always the day when you only have an unregulated wall-wart and that's the day you wish you had put that filter in!   A 100uF cap and a 100ohm resistor is usually a good ballpark. However on that circuit you could do better with a 220ohm and a 220uF cap.  Which will decrease the hum in the order of 30dB.

[earthtonesaudio]

A couple things to ponder:

Pure clean DC doesn't contain any AC noise and thus a filter would make no audible change.  Regulated supplies can be very quiet, but they don't solve all the problems.  Wires=inductors=single coil pickups, so the longer the wire between the wall wart and your effect, the more hum you can pick up from the environment.

So even if you use a regulated wall wart, you can still pick up hum between the wall wart and the wah.  Unless you use a very short power cord and/or shielded power cables (or a battery!), hum filtering on the effect's circuit board is still a good idea.

[petemoore]

  Yupp, the real question is: "How "DC" is this DC...actually.
  There is no 'perfect' DC, not that it has to be: 'not AC-ish' enough to matter is fine, most of the time.

[alanlan]

It helps to know the current consumption of your circuit also.  For example, if your pedal takes 5mA, then a 100 ohm series resistor in the filter will drop 500mV.  That may be OK.  Dropping 2V may be a real problem.  If your pedal has an LED, then don't filter the supply to that; just filter your audio circuitry supply so you minimize the current and therefore volt drop through the filter.  In fact you can take things a bit further and filter certain stages in your circuit separately.

[tehfunk]

Thanks for all the advice, you guys raised a couple of questions in my mind... Does the power filtering affect the amount of power going to the circuit? Also, with the wall wart cable length causing hum, are you saying that I can help cancel that out by using power filtering, or were you saying that even with power filtering, hum from the something such as that is inevitable? Thanks.

[earthtonesaudio]

If you use a resistor/capacitor filter, the resistor drops voltage.  More voltage drop if the circuit draws more current (Ohm's law).  You can use an inductor/capacitor filter, and probably not have to worry about the voltage drop across the inductor at all for most circuits.  But small value inductors are not very effective, and large value inductors are physically large, so there are some trade offs there.

The wall wart cable does pick up hum, but it's not really that bad in most cases.  Usually nowhere near as bad as an unregulated power supply, and you might not even notice a problem at all.

[slideman82]

Keep it simple: just add a 78L09 100mA regulator inside the wah, and some 10V 1000uF or 2200uF cap in parallel with a 4.7n cap, and ripple will be rejected a lot! Try it! Larger the cap (4700uF 10V) better rejection.

[brett]

Hi
also, the larger the series resistance, the better the rejection.

But you don't want a massive cap and a massive resistance, right?

For low power pedals, try 10 ohms and a 470 uF cap.  For a high power pedal (e.g. 200 mA), you'll need to go to something like 3.3 ohms and a 1000uF cap (to keep the voltage drop low - V = I x R.  In this case 3.3 ohms x 0.2 A = 0.66 V.  Or try 2 x 470uF caps (a "trade secret" is that multiple small caps work better than large single caps).
cheers

[George Giblet]

> If you use a resistor/capacitor filter, the resistor drops voltage.

It's not the fact causes the drop it's matter of how much.

That circuit draws very little current, in the order of 600uA.  The voltage drop across the 220ohm resistor I suggested would produce a 0.132V, which is insignificant.  The current is fixed so the voltage drop is constant, known, and small - there is no problem.

In the context of an unregulated wall-wart the actual output voltage and small voltage drops has little point.  An unregulated 9V wall wart could produce anything from 9V to 13V.  If you knew your wall-wart was going to produce say 12V you could purposely select the resistors such that it produced the correct voltage drop to get down to 9 to 10V.

Where keeping the voltage drop small comes into play is if you want to connect unregulated and regulated supplies.  The regulated supply is obviously producing the right voltage so you don't want to change it.  The regulated supply is producing hum so you want to filter it.   The (sensibly) small resistor ensures the effect sees this voltage.   Once you choose to go for a small voltage drop you obviously cannot tweak the resistor to get the right voltage in the unregulated case, you have to accept what it puts out - after all it is unregulated.

[GibsonGM]

100 ohms & 1000 uf (2x470uf in parallel) on the board has always worked wonders for me in desperate cases.  Even killed the clock tick in an Easyvibe.

Best way to do this is to make sure your suppy IS regulated!  Using an LM7809 etc. Filtering on the board is a pain in terms of real estate, and redundant altho occasionally necessary.

There are chapters in books on this, of course.  The simplest way I understand basic regulation is that you need to anticipate how much current your circuits will be drawing, and plan for a 'stiff' idle current in the voltage divider.  Filtering is a secondary problem.

Think of a simple resistive voltage divider, Eout=R2/R1+R2)Ein.    That becomes R1=E1/I2 (I2 is assumed);  R2=E-E1/I1+I2. Where E=supply voltage, E1=desired output voltage taken at the junction of R1,R2.  I1=load current, I2=idle current thru divider.  The higher the idle current, the better the regulation.  This is assuming your PS can deliver all the current you will be asking it to (I1 and I2).   Then, a BFC (1 to 47uf) at the output junction just like in a pedal bias supply will provide your filtering, minimizing ripple (but not totally eliminating it, that never happens).
There's another equation for ripple that I forget; could just be a simple filter eqn.  For better regulation and rejection, an IC is then the best way to go.   

[Paul Perry (Frostwave)]

here's another side of it: you might think that if you are only running off a battery, then you would not need an electro across the power rail to ground at all.
But the circuit designer assumes that the power supply is a battery with zero internal resistance.. and if it is not (eg as the battery gets older) then it looks like there is a resistor in series. And as one knows from deliberately putting resistors in series with batteries, that can lead to all manner of strange effects. Putting a large electro across the power supply, makes the supply impedance (at least at audio frequencies) look much lower, so it is likely to be more stable.

[fogwolf]

I personally have never seen ferrite beads used on a stompbox circuit but there is a very similar thread at the electro-music synth diy forum:

http://www.electro-music.com/forum/viewtopic.php?t=25712

[tehfunk]

Interesting, I have been reading all of this, some of it has gone far beyond the scope of my question, far beyond my electrical engineering "key terms" knowledge and logic, but hey, it's cool. Anyway, to stick with my specific situation, so I can close the open ended nature of my question, I am using a regulated power supply that runs at 9.1v, in this case for a wah which I am aware draws very little current, so does the little voltage drop due to the resistor cause any "problems" for the circuit? And, after reading all of this it has become a little unclear, is power filtering more necessary with regulated power supplies or unregulated, or is the necessity the same for both, as in neither produces an inherently larger amount of hum. Also, for future usage, if it is an effect with a larger current draw, what is the difference in the situation. Should the values be different?

Finally, for my wah, what value resistor and capacitor, or capacitors in parallel would be best, if I filter the power? I've heard lots of values, maybe some relatively definitive ranges would be nice. Thank you so much for all your advice.

[brett]

Hi
ferrite beads used to appear on the base leg of transistors in high-gain circuits to prevent load radio programs appearing (especially during outdoor concerts, for some reason).
Also, I used to have a 1970s hi-fi (Dick Smith Playmaster 40) with BC107s in the RIAA preamp with ferrite beads on the base.
cheers

[earthtonesaudio]

Sam, if you're using a regulated power supply, most of your problems are minimized already.  You might not need power filtering at all.  Most voltage regulator datasheets recommend some bulk capacity on the board if the regulator is far away (like your regulated wall wart), both to reduce effective supply impedance and also to filter any noise picked up by the power cord.  10uF is probably overkill, so try that!

You should add supply filtering ONLY if there's audible hum coming from the wah.  If it's silent, there's nothing to filter out!


Anyway if you do decide to filter:
Since your effect draws little current, you can get away with a bigger resistor before the voltage drops significantly, but adding series resistance with the power or ground rail can lead to unwanted oscillations, because the different gain stages can more easily interact with each other.  You could put separate power filtering on each transistor, but that's a pain.  Most people just use a big cap and as little resistance as necessary.  100uF and 100 ohms makes a low pass filter that is -18dB at 60Hz. 100uF/200 ohms is down 24dB at 60Hz, etc.
( http://www.muzique.com/schem/filter.htm )

Ferrite beads are good for filtering super high frequency noise (think USB cables).  They're basically a single turn of wire inside a little bit of ferrous material.  Useless for 60Hz hum, though.

[Paul Marossy]

You should add supply filtering ONLY if there's audible hum coming from the wah.

In which case filtering isn't going to really help because the inductor itself is introducing the hum into the signal path. The only thing that I have found works with a humming wah pedal is to orient it as far away as possible from the actual source of hum (as in transformers in your amp, etc). Essentially the same problem as a single coil pickup picking up hum from the surroundings...

[tehfunk]

Ok, thank you guys, I'm pretty much set for now.

[RedHouse]

+1 on using the R-C filter.

I resolved an issue with one of my Univibe clones running off a name brand "regulated" power supply which still producing noticable hum (in a recording studio environment) which wasn't as noticable on stage.

It turned out that a dual R-C filter 100-Ohm/1000uF->100Ohm/1000uF->78xxx regulator did the trick.
(probably couold have used 470uF caps but I had 1000uF's handy that day)

[petemoore]

  DC has no AC in it if it's 'true DC'.
  Wallwarts often have some AC...'riding' on the DC [it always wants to get into DC it seems, and keeping it out is what steps need taken to eliminate it entirely, good enoughly, or satisfactorily, iow so it's not heard. 60cycle hum from AC wall supply [120v] is a pretty common thing to hear.
  "Amplification' in the sense here, is like using a steady 'pressure', [DC], to make AC [..pressure/voltage that varies at the frequency of the input]. Using a source like guitar...[which puts out AC...at the frequency the string vibrates], the amplifier controls the release of DC at 'that' frequency [or frequencies], mimicing the source input, but because the DC has so much greater potential [much higher voltage possible than what a guitar puts out], a huge increase in pressure [voltage] can be seen at the 'output'...so the AC wave that was input, looks just like the AC wave that is output, but the output wave is a lot 'larger', and the voltage swings are greater. It controls a DC potential, using the input source to decide when DC is released...
  Oversimplified of course but tries to demonstrate that amplifier power supply is DC [and and opamp is an audio amplifier, as is transistor, tube etc.], they use control of DC potential to 'amplify' AC which goes in small, comes out big.