Power Supply Filter Anatomy

[POTL]

to create their own projects, or copying of the legendary circuits, each faced with the need to add a power supply filter

Tracing the number of projects on the site madbean I noticed that the power from the pedals filters such as overdrive and distortion are very similar, but still there are some differences

This is Proco Rat

This is Tube Screamer

This is Blues Breaker


1)Rat and Tube screamer have a 47 ohm resistor between the capacitor and resistor 4K7, Bluesbreaker does not have this resistor
2)Tube screamer and Bluesbreaker first capacitor is first electrolyte capacitor, a second non-polar (film)
rat first capacitor is non-polar (film), second electrolyte
3) voltage divider has a different resistor values each model
4) Rat and TS last capacitor is 47uf, bluesbreaker have a 100uf

I searched this forum and a few, but have not found an answer to my questions:
1)If you change the component values, how it will affect the sound effect.
Or power filter does not affect the sound?
2)what principles chosen denomination capacitors and resistors?
3)why the location (electrolyte and film) capacitors in the rat differs from TS and bluesbreaker?
4)whether it is possible to replace the circuit, such as use of a tube screamer circuit for rats, and vice versa, or for each effect is calculated proper scheme?
5)what does 47 ohm resistor and why it is not in bluesbreaker
6)in many circuits the cathode of the diode is connected to the power supply, the anode to the ground
Here the diode is not touching the ground, why?

I apologize in advance for the noob question, but I never was able to get detailed information

I think this topic will be of interest to many

[mth5044]

This might help you a bit for the voltage divider:

http://www.geofex.com/circuits/biasnet.htm

The 47R is part of a lowpass filter with the caps. I'll let someone smarter explain that one.

[GibsonGM]

It seems like the ones WITH 47R are doing a better job filtering. Whether it's really needed or not could be debatable; but if your wall wart isn't well filtered, they sure will help!

A lowpass is formed with the 47R and 100n cap (I believe also with the 100u following it as well, but it is very small so I will disregard it).   

Cutoff frequency Fco= 1/2pi R * C    R can be in megohms if C is in microfarads, so

Fco = 1 / 6.28*.00047*100 = ~33Hz if my hasty math is correct   

The 100n is there because the reactance of the 100u and 100n are different for a given frequency, and one will 'channel' higher freq. junk to ground better depending on frequency.   In other words, high frequencies will 'like' one cap over the other (the 100n will be more 'attractive' to them), as it will appear to have less AC resistance...the order they appear in would not seem to me to be important.     In reality, wouldn't we add the capacitance for the above formula, anyway? (to be super-accurate...the 100n is very small).  I don't even use the small values in my build unless they are explicitly called for, but they cannot hurt to be there, certainly.

Without the 47R there, you are dependent on the series impedance of the power supply itself to provide the resistance for the filter.  Adding the 47 assures that you have that amount of resistance at minimum. 

Changing values will not change the sound UNLESS you go so far as to remove filtering that was removing noise from the power supply and it enters the circuit.  OR, theoretically you could raise the resistor so high you starve the circuit for current, but those are extreme things to do...

The denomination of those components is decided upon per the formula above, and where you would like the knee of your filtering to be....

The LED will eventually return to ground in order to work...perhaps the schem. symbol at its bottom is to show it 'goes somewhere', like a switch?

For all normal and foreseeable (to me) effects that are similar to these distortions...that have similar current draw....ANY of those schemes should work fine!   

[287m]

Woah, great timing!
as noob, i just want ask too for next project 

a year ago, i built a 2 in 1 dirt box for my friend who play guitar and bass, the pedalboard start with AB box.
A few days ago, he call me. He want build again for his new brother in law, play guitar and bass too!

I hate doing the same (2 board), decided to build in one pcb/perfboard.
Here i remember the circuit. 1 for guitar, 2 for bass.


the question is, can i use just circuit 1 only to powering? SInce the power is almost same?


and after read Mike post,


combo A+B = C, right?

[POTL]

Thanks for answer

The LED will eventually return to ground in order to work...perhaps the schem. symbol at its bottom is to show it 'goes somewhere', like a switch?

it goes to the switch

I have some difficulty understanding English
made the following conclusions(correct if I'm wrong)
1) changing the values of capacitors and resistors(within reason of course) in the schema will not change sound effect
2) Can be used for similar diagrams(overloads, overdrives) one diagram for power supply filter
3) the resistor is 47 ohms and a capacitor of 100nf form a frequency filter
4) the values ​​of resistors voltage divider in charge of amperage


Here's another question.
1)that changes in the value of the first electrolytic capacitor(100uf)
2)that changes in the value of the last electrolytic capacitor(47uf or 10uf)(those after a voltage divider)
3)Can I for all three schemes(Rat, tube screamer, bluesbreaker)to use the same components for the voltage divider and the last condenser
as an example
47K + 47K 10 uf?
4)what influences the level of current?
as we found out it depends on the values ​​of the resistors in the voltage divider

[POTL]

This might help you a bit for the voltage divider:

http://www.geofex.com/circuits/biasnet.htm

The 47R is part of a lowpass filter with the caps. I'll let someone smarter explain that one.

thanks
left to read

[POTL]

accidentally created a new message

[Beo]

Tracing the number of projects on the site madbean...

It's important to note that madbean is not cloning retro designs exactly, but modifying them, either to simplify, add features, make more compact, benefit from modern components, etc. I think most of his projects are using a fairly unified power filter layout which works well across most/all circuits. But this may be overdesign for some. Just want to make sure you don't think the madbean schematics are representative of original circuit designs... cause they are not. Tonepad is much closer, with mods clearly identified, and there are lots of schematics on this site.

[merlinb]

Series diode: If you use a series protection diode then you automatically lose about half a volt across the diode (unless you use a Schottky diode, possibly). Does this matter? In many pedals, no. A possible advantage is that if the pedals draws extremely little current (microamps), the diode itself will act kinda like a resistor, i.e. you get some filtering effect in combination with the main smoothing cap.

Parallel diode: An alternative protection method that doesn't incur any voltage drop. Needs a power diode. Can be dodgy when used with an unregulated power supply because the PSU and diode will be in a race to the death when connected up wrong. A series resistor (before the diode) helps here, since it will limit the current (or even act as a fuse) as well as being part of the PSU filtering.

Series resistor: Forms an RC filtering circuit with the main smoothing cap. The bigger the better, but you have to trade off how much supply voltage you can afford to lose. The more current the circuit draws, the bigger the voltage drop. 10 to 100 ohms is typical. Sometimes this resistor is omitted altogether, esp in simple opamp pedals which have inherently good power supply rejection. Also see above.

Main smoothing cap: Typically anywhere from 47uF to 470uF. The bigger the better, really, although huge values (thousands of uF) may be avoided because they will drain more battery power at start-up.

Vref resistor: If the Vref is buffered then these resistors can be quite large, like 100k, or even 1Meg when using a JFET opamp like TL072. Bigger means less wasted current. But if Vref is not buffered then it depends how much current get's sucked out (or into) Vref by the rest of the circuit. Smaller means more stable Vref in the face of varying current.

Vref smoothing cap: Not critical. Basically use C = 1/(2.pi.R) where R is one of the Vref resistors.

[POTL]

Series diode: If you use a series protection diode then you automatically lose about half a volt across the diode (unless you use a Schottky diode, possibly). Does this matter? In many pedals, no. A possible advantage is that if the pedals draws extremely little current (microamps), the diode itself will act kinda like a resistor, i.e. you get some filtering effect in combination with the main smoothing cap.

Parallel diode: An alternative protection method that doesn't incur any voltage drop. Needs a power diode. Can be dodgy when used with an unregulated power supply because the PSU and diode will be in a race to the death when connected up wrong. A series resistor (before the diode) helps here, since it will limit the current (or even act as a fuse) as well as being part of the PSU filtering.

Series resistor: Forms an RC filtering circuit with the main smoothing cap. The bigger the better, but you have to trade off how much supply voltage you can afford to lose. The more current the circuit draws, the bigger the voltage drop. 10 to 100 ohms is typical. Sometimes this resistor is omitted altogether, esp in simple opamp pedals which have inherently good power supply rejection. Also see above.

Main smoothing cap: Typically anywhere from 47uF to 470uF. The bigger the better, really, although huge values (thousands of uF) may be avoided because they will drain more battery power at start-up.

Vref resistor: If the Vref is buffered then these resistors can be quite large, like 100k, or even 1Meg when using a JFET opamp like TL072. Bigger means less wasted current. But if Vref is not buffered then it depends how much current get's sucked out (or into) Vref by the rest of the circuit. Smaller means more stable Vref in the face of varying current.

Vref smoothing cap: Not critical. Basically use C = 1/(2.pi.R) where R is one of the Vref resistors.

wow this is a detailed and clear answer, thank you!
As I understand:
1) resistor value is not ibeet large value (for operational amplifiers within 10-100 kilohms to 1 megohms is a FET)
2) The electrolyte and the latter did not play a role? So I can put and 10uf and 100uf?

[merlinb]

1)If you change the component values, how it will affect the sound effect. Or power filter does not affect the sound?

Power filter does not affect the sound, as long as there is enough filtering.

3)why the location (electrolyte and film) capacitors in the rat differs from TS and bluesbreaker?

The locations are the same, but the component numbering is different. The capacitors are in parallel, so it does not matter in what order they are drawn in the diagram.

4)whether it is possible to replace the circuit, such as use of a tube screamer circuit for rats, and vice versa, or for each effect is calculated proper scheme?

They are nearly the same, so would probably work. Only the Vref resistors are different.

5)what does 47 ohm resistor and why it is not in bluesbreaker?

The resistor could be used in the Bluesbreaker too, but in this case the designer 'got away with it'.

1) resistor value is not ibeet large value (for operational amplifiers within 10-100 kilohms to 1 megohms is a FET)
2) The electrolyte and the latter did not play a role? So I can put and 10uf and 100uf?

I don't understand your questions.

[POTL]

4)whether it is possible to replace the circuit, such as use of a tube screamer circuit for rats, and vice versa, or for each effect is calculated proper scheme?

They are nearly the same, so would probably work. Only the Vref resistors are different.

what happens if I replace the values ​​of resistors?
10K for a rat (instead of 100K)
100K for a tube screamer (vmesto10K)

1) resistor value is not ibeet large value (for operational amplifiers within 10-100 kilohms to 1 megohms is a FET)
2) The electrolyte and the latter did not play a role? So I can put and 10uf and 100uf?

I don't understand your questions.
[/quote]

I wanted to clarify what impact changing the value of resistors in a voltage divider (three circuits different values ​​- 10K 47K and 100K, respectively) if I want to collect all three schemes and in all cases use as a voltage divider 100K resistors, how it will affect the tube screamer (the default 10K resistors) and Bluesbreaker (default 47K resistors)


asking about the last electrolytic capacitor, I mean one that is after voltage divider, as a change in its value will be reflected in the scheme?
for example in a tube screamer and the rat is the value at bluesbreaker 100UF is a value of 47 uf
what happens if I replace the value of the capacitor bluesbreaker 100uf or 10uf?

[GibsonGM]

1) the voltage divider resistors set the voltage that is applied to the opamp as a BIAS VOLTAGE, often called "Vref", reference voltage.  Quite often that is 1/2 Vcc, or 4.5V for a 9V circuit.  ANY 2 reasonable resistors of the same value will give you 1/2 Vcc.   The difference is the current that is flowing through them.   Using 10 ohm resistors is very wasteful...so often you see at least 1k and 10k is very common.   

http://electro.uv.es/asignaturas/ea2/archivos/sloa058.pdf    The desire is to balance noise issues with high value resistors, and bias currents.  This is a good source to begin to understand.     To my understanding, values from 1k to 47k are acceptable for these kinds of audio applications....higher values increase noise, but MAY have some kind of 'starved bias voltage' function.

2) The value of the cap after the voltage divider is not very important...I have used as low as 2.2u, and as high as 100u, with no issues.   The common wisdom is to try to have a 'fair value cap' there...at least 10u.   And as Merlin pointed out, higher value caps can take longer to charge....so perhaps 100u is the upper bound for this.
The cap simply provides a reservoir for the bias voltage, to assure a nice, clean supply that does not 'wander' if your power input varies a bit.    If you see that a 10u is called for, and use 33u, you will experience no change, no problems.

[POTL]

1) the voltage divider resistors set the voltage that is applied to the opamp as a BIAS VOLTAGE, often called "Vref", reference voltage.  Quite often that is 1/2 Vcc, or 4.5V for a 9V circuit.  ANY 2 reasonable resistors of the same value will give you 1/2 Vcc.   The difference is the current that is flowing through them.   Using 10 ohm resistors is very wasteful...so often you see at least 1k and 10k is very common.   

http://electro.uv.es/asignaturas/ea2/archivos/sloa058.pdf    The desire is to balance noise issues with high value resistors, and bias currents.  This is a good source to begin to understand.     To my understanding, values from 1k to 47k are acceptable for these kinds of audio applications....higher values increase noise, but MAY have some kind of 'starved bias voltage' function.

2) The value of the cap after the voltage divider is not very important...I have used as low as 2.2u, and as high as 100u, with no issues.   The common wisdom is to try to have a 'fair value cap' there...at least 10u.   And as Merlin pointed out, higher value caps can take longer to charge....so perhaps 100u is the upper bound for this.
The cap simply provides a reservoir for the bias voltage, to assure a nice, clean supply that does not 'wander' if your power input varies a bit.    If you see that a 10u is called for, and use 33u, you will experience no change, no problems.

Thank you
Now everything fell into place and I understood everything

[GibsonGM]

You are welcome.  There is much more to the issue of power supply filtering, but for what we do here perhaps this is enough to understand a little about it.

If you need help understanding what "Vref" (bias voltage) is for, seach the forum or make a new topic!     

[POTL]

You are welcome.  There is much more to the issue of power supply filtering, but for what we do here perhaps this is enough to understand a little about it.

If you need help understanding what "Vref" (bias voltage) is for, seach the forum or make a new topic!     

thanks bro
I know the principle of operation of the filter( more less now I understand the principle of), and that such voltage bias(virtual ground)
I was interested in the exact values ​​and how they work.

Although wait one question still remained 
in schemes often use next diodes:
1) Rectifier diode( usually 1n4001)
2) Schottky diode( usually 1n5817)
3) protection diode( usually P6KE22A)
what type of diode is considered to be more reliable, durable and practical?

if the protection diode on what parameters it choose?

[merlinb]

A power rectifier diode (e.g. 1N4001) is used when the diode is in parallel with the incoming voltage. If the power supply is connected backwards, the diode will conduct and pass maximum fault current until something burns or shuts down. That fault current is why you need a power diode.

If you use a series diode then it does not have to pass any fault current -only the normal working current of the circuit (milliamps). If the power supply is connected backwards then the diode simply blocks the voltage. You can therefore use a small diode like 1N4148, which has a blocking rating of 100V. However, if you want a smaller voltage drop during normal operation then a Schottky is better, but will usually have a lower blocking rating. The 1N5817 is rated for 20V, which is still enough for our 9V pedals.

'Protection diode' or transient suppressor is used when you think someone might connect the power with the right polarity, but the wrong voltage (too high), or a voltage spike on the power supply. Most 9V circuits can handle being accidentally connected to 12V, or perhaps even 20V, but not 50V. So you might use a 20V transient suppressor. But how likely is anyone to plug in a >20V supply? Most builders don't think this is very likely, and it is the user's stupid fault if he does!

[POTL]

A power rectifier diode (e.g. 1N4001) is used when the diode is in parallel with the incoming voltage. If the power supply is connected backwards, the diode will conduct and pass maximum fault current until something burns or shuts down. That fault current is why you need a power diode.

If you use a series diode then it does not have to pass any fault current -only the normal working current of the circuit (milliamps). If the power supply is connected backwards then the diode simply blocks the voltage. You can therefore use a small diode like 1N4148, which has a blocking rating of 100V. However, if you want a smaller voltage drop during normal operation then a Schottky is better, but will usually have a lower blocking rating. The 1N5817 is rated for 20V, which is still enough for our 9V pedals.

'Protection diode' or transient suppressor is used when you think someone might connect the power with the right polarity, but the wrong voltage (too high), or a voltage spike on the power supply. Most 9V circuits can handle being accidentally connected to 12V, or perhaps even 20V, but not 50V. So you might use a 20V transient suppressor. But how likely is anyone to plug in a >20V supply? Most builders don't think this is very likely, and it is the user's stupid fault if he does!

Thanks i stopped on Schottky

Look like this

[ashcat_lt]

Thanks i stopped on Schottky

Look like this

Then you missed the point.  In that version you want the 1N4001.

I didn't see it mentioned, but I think the reason we have both an electrolytic (100u) and a film or ceramic (100n) cap in there is that the electrolytic just doesn't work very well at higher frequencies.  With an ideal theoretical cap, it would be fine on its own, but with real components, it just doesn't for whatever reason, so we put that smaller cap next to it in order to handle what the big one misses.

[LightSoundGeometry]

my power supply is super simple, a rail to rail electro lol ..I may ground the input with shielded cable as well

I sometimes use the amz led anti pop mod