DC filtering question

[nonoxxx]

Hello,

I looked at DC filtering after getting some noise with tube screamer, and when looking a some schematics I see often in the power supply section of pedals a low value Resistor with a high value cap to form a low pass (that's ok ) but there is some schematics with an added 0.1uf cap added in parallel with the big cap. I had read somewhere that it help for impedance , but someone can explain me more in depth why?

Thanks.

[amptramp]

Electrolytic capacitors can pack a lot of capacitance into a small space but there is one problem: at higher frequencies an electrolytic capacitor begins to look more resistive and can even turn inductive.  If you look at the last chart on this datasheet:

http://eecs.oregonstate.edu/education/docs/datasheets/XC-600178.pdf

the impedance which should be an uncompromising straight line going down to the right against frequency is not straight and the upturn for large values (lower starting impedance) indicates that it has gone inductive.  To make an electrolytic capacitor, you anodize the aluminum and this thin coating of anodize forms the dielectric.  This structure is wound around a core so it looks inductive at higher frequencies because it is wound like a coil.

The ceramic or film 0.1 µF capacitor across the electrolytic does not have problems with high frequencies.  It can look capacitive out to 100 MHz and beyond.  Since the two capacitors are in parallel, the combined impedance is low for an extended frequency range,

[GibsonGM]

^^ Yep.    In the ideal (imaginary) world, a cap is a cap, an inductor an inductor.  But in the real world, they each exhibit properties of the other, at certain frequencies.  Plus some resistance.

At some frequency, a cap will become self resonant .  Clearly we don't want this in our power supply filter!  So the smaller cap (which will have its resonant frequency elsewhere) takes over for that range of junk you're trying to filter, per Ron's description.   

Neat, huh?  Something to keep in mind when experiencing some just CANNOT be figured out problem of oscillation, motor boating, etc. 

[nonoxxx]

Thanks I understand,
So I see  on banzai music ( a electronic part retailer) really high value Electrolithics like 1000 uf/ 1500uf marketed with a "low impedance" and good for audio  like this ones
https://www.banzaimusic.com/Panasonic-FC-1800uF-35V.html
Does it need also the little cap in parallel?

Because if I want to filter a DC theorically  : a 5 ohm resistor with this kind of cap should theorically work well and dont attenuate too much the voltage. For example I made some test with a screamer and putting a 50 ohm resitor with the 100 uf cap do a good job, but the dc after was at 8.5 V instead of 9.2 V it should be nearer than the source with 5 Ohms.

Sorry to bother you with my questions but It's difficult to find good information's source on DC filtering in pedals and all circuit seems doing approximately the same things.

[anotherjim]

Equivalent Series Resistance or ESR. Low ESR in an Electrolytic capacitor is possible, but you pay more for it. Low ESR will not prevent the effect of self inductance with high frequencies.

Low ESR caps become good things in amplifier power supplies or speaker coupling capacitors, but are wasted money in an FX pedal.

[GibsonGM]

It seems that this value cap, over 1000u, is pretty high, Nonoxxx.   Traditionally, one might use a 100R resistor, 470u cap, for a cutoff of 3.4Hz.  Then a small value cap in parallel as we discussed.   It is "overkill".   You need not pay so much for a cap for power supply filtering!!!
You can use a generic electrolytic here, for 1/4 the cost.   Perhaps you just need the .1u?

By "low impedance", I am taking the manufacturer as stating that the cap has a low ESR, or 'effective series resistance', one of the 'real world' properties I mentioned above...so, "it's a good cap".   It's probably TOO good for power applications outside of Hi Fi!   I don't think you gain anything by putting such a tiny resistor and huge cap in place.   I would associate that kind of reservoir requirement with an LFO, for a phaser or something - something which needs to draw current rapidly.  Not a TS.

What noise do you have with your TS?  Try it on battery - if it goes away, it's the power supply.  But the standard 'tricks' to cure this usually are effective....the problem could be elsewhere, such as lead dress and parts placement, if this is excessive.  Is this from a kit, or your own layout?


Edit: I see you got here as I typed, Jim...

[nonoxxx]

THanks ,
I do not have noise problem ( infact it's fixed ,lol ) but It make me studiyng more in depth dc filtering and It was an interrogation when seeing these high value caps, at first I was looking for ordering a 470uf to test and I see this  .

I agree for the 470 uf and the 100 R resistor tips  but in pratice for 9.2v the DC will be attenuate to 8V (tested it with my screamer).
So it was my interrogation to put a resistor with a lower value and a greater cap that will not alter the 9v DC too much . I wandering also how high I could go in Uf.

[duck_arse]

nonoxxx, does your sceamer have a led indicator? if you run the led from the R/C filtered supply, all the led current will add to the audio-circuit current to increase the voltage dropped across the R in the R/C. so, connect the led circuit to the unfiltered side of the supply, before the R/C, then the voltage dropped will be just that from the audio circuit current draw.

[nonoxxx]

 thanks a lot , I will test it this evening,

So I used this kit  : http://diy.musikding.de/wp-content/uploads/2013/06/screamerV2_schalt.pdf and put the 100R just after the dc jack . I forgot that  the LED was also after  and also I don't know too much about voltage drop lol 

At 100 R the voltage dropped to 8V and at 50R 8.5V

[Phoenix]

While all the above info is quite correct, I'd like to point out that the self-resonant frequency of any electrolytic is FAR above the audio spectrum, and so the inductance and rising impedance (which will happen up in the MHz range) is not a problem at all. So adding a film or ceramic capacitor in parallel with a bulk electrolytic cap is pointless with audio. This stuff is important with digital or RF circuits which operate at high frequency, but in audio we don't need to worry.

Now, if you're talking about local bypassing, as opposed to bulk decoupling, that is good practice. To explain what I mean by this, think of it this way. The bulk filter cap (your 100uF or 1000uF or whatever) is like a water tank. You require that water tank for all your water needs - you need to gather up the rain whenever it falls and need it to wash your clothes, water your garden etc. But you wouldn't want to carry it around with you all day just for when you need a drink right? Rather, you'd carry a water bottle with you, and top it up from the water tank. This is what local decoupling is. It's generally considered good practice to add local decoupling to any IC's, usually a 100nF ceramic capacitor as close as you can get to the power pins of each IC as you can. This helps to prevent them from oscillating if they need a small "sip of water", so they don't have to walk all the way to the water tank
If an IC like an op amp does start oscillating, it can cause odd behaviour like excessive or unpleasant distortion, latch-up, phase inversion, or intermodulation products.

About ESR, that's not really a concern in DC circuits, but is important in AC to DC power supply filtering (not the type in pedals, but like the power supply that POWERS your pedals). The ESR causes the capacitor to self-heat with ripple current, which causes losses, and also places an upper limit on the amount of ripple current a capacitor can handle. Switch mode power supplies in particular, but also linear power supplies are the main use for low ESR capacitors. You won't see any benefit from using them in your effects pedals.

Hope this helps.

[nonoxxx]

nonoxxx, does your sceamer have a led indicator? if you run the led from the R/C filtered supply, all the led current will add to the audio-circuit current to increase the voltage dropped across the R in the R/C. so, connect the led circuit to the unfiltered side of the supply, before the R/C, then the voltage dropped will be just that from the audio circuit current draw.

I made a quick test , and it give the same result :/  8.3 V after the resistor instead of 9.2V

Here is a pic of the circuit maybe I miss something.

[nonoxxx]

I think I understand when looking at a build I made of a Big muff with this king of RC filter (the coda music PCB) , the resistor has to be also after the diode ? it's seems  to be the only difference , and the muff is at 9V

[GibsonGM]

Hmm, it looks like the LED IS before the RC filter.  I'll leave to Duck or someone to explain the voltage drop better...

One tip - if you make your resistor and LED leads (or any leads you must connect) into "hooks", you can 'join them up' and pinch them with pliers til they are flat.  Makes a better connection, and you can put heat shrink tubing on there before you solder, slide it over, and warm it up to cover the joint     Then you don't have those little spiky shoots there to get caught on.

[nonoxxx]

Yes the Led is before the Rc filter as explained bu duck_arse the 1 voltage drop occur also if the LED is deconnected.

Thanks for tips GibsonGM I have some heat shrink tubings I will test this

[GibsonGM]

I am surprised that the voltage drop is so much.

I don't put any additional filtering into a pedal unless it is really required.  There may be another way to lessen your noise!  Proper lead dress is one way, although your build looks clean - could move the LED wire over to the chassis and around....  What do you use for a wall wart?  Is it a "good one" that is regulated? 

Adding resistance can INCREASE noise sometimes!  An LFO may create current spikes that cause "ticks", and you need the reservoir cap to minimize this...but a TS should be fairly quiet other than the normal 'hiss' of high gain.   Power filtering may not reduce this much at all....to retain full voltage one may need to go up in power supply voltage, and add a regulator....a lot of work for what is probably just a tiny bit of improvement!

I accept that none of my distortions/TS-type OD will be "on" when I want to be quiet.  They will all make hiss, that is what they do. They amplify ANY noise at all many many times.   But perhaps your noise is more than this?

[nonoxxx]

This noise comme from the my supply power a cheap behringer Multi out PSA-230S (known to not be the quietest) ,
I know it because on a a joyo screamer clone  I own (with no mod)  it's the same ( in fact it's in the dc and  the noise is at 3500 herz ,checked with the EQ of cubase also 3500 herz is part of the boosted range of the screamer tone stack ) and when playing with prototype board whith this kind of cicuit I noticed that occur when putting the 4.5v on the input of the opamp and the filtering remove it.

I am just experimenting so it' not a real issue and I know with a good power supply it will be quiet but I am someone that always need to know why when it talk about this kind of stuffs lol 

I am surprised also by the voltage drop because I build a Big Muf with exactly the same RC filter in the DC (the coda effect PCB)  and on the muff there is only 0.2 V drop after the 100 ohm resistor.

the schematic of the muff : (the value are the same , a 100 r and a 100 ufcap (there is a 0.1uf cap in parallel)

https://up.neissa.org/fZPF9/20170421121944930p1i0.png

It's not a real issue and I bother you only to understand.

[amz-fx]

The voltage drop across the 100 ohm resistor will vary according to how much current is flowing through it. 

With the 0.2v drop on the Muff, it seems like the circuit is drawing 2 ma.  V = I * R , 0.2 = I * 100, or 0.002 amps.

The TS draws more current, so the voltage drop across the resistor is higher.

Best regards, Jack

[Frank_NH]

"I am surprised also by the voltage drop because I build a Big Muf with exactly the same RC filter in the DC (the coda effect PCB)  and on the muff there is only 0.2 V drop after the 100 ohm resistor."

I believe this is simply due to the different current draw between the two circuits (Ohm's law).  You can try reducing the resistor's value to reduce the voltage drop to an acceptable level.  I don't think that will impact the effectiveness of the LF filter for the power.  (edit - I see Jack replied while I was editing my comment )

[bluebunny]

I can't be sure if the colours I'm seeing in your build picture are accurate, but are you sure of the values of some of your resistors?  The "100R" resistor you have before the 9V point looks very similar to the "10K" R17/R18 resistors providing Vb, which in turn look very similar to the "1K" resistor at R5 below the IC.

[nonoxxx]

no it's 100 ohms ( checked with my multimeters )