Power Filter Section

[zelcanada]

Hello All,

As I continue to grow my skills in pedal design, I have a question about the power filtering section. I understand it's importance for protecting the circuit from reverse polarity and reducing noise (...?), however I seem to see a lot of different schematics in the power filtering section of a pedal schematic. I'm starting to get lost a little... Is there a "best practice" for this? Do certain kinds of pedals "prefer" certain kind of power filtering?
Any advice would be greatly appreciated!
Thanks!

 

[Robert]

(I realize you're focusing more on filtering than polarity protection, but I had this ready to post in another section)

The three most common methods of polarity protection I see are:

Series - The diode is in series with the load/circuit and blocks any reverse voltage/current. No current flows so no damage is done to the circuit, it simply doesn't work until the correct polarity is applied. The negative is that the series diode drops a small amount of voltage (forward voltage of the diode). This is generally not an issue. A schottky diode can be used to reduce the voltage drop, however these typically have a lower current capability and reverse breakdown voltage. (It takes less current / voltage to damage it)


Parallel / Shunt (Diode to ground)- The circuit receives the full applied voltage, no loss of voltage (forward voltage drop) across the diode. If reverse polarity is applied the diode typically shorts to ground, making the pedal effectively "broken" to the average consumer who can't replace it.



MOSFET - Similar to series, but a MOSFET is used instead due to it's lower forward voltage drop at the cost of a slightly more complex/expensive circuit.




A much less common method is to use a relay. I've never seen this done in a pedal, it's more of an industrial / consumer electronics thing. I've seen this done in a few arcade / pinball machines over the years as well.

Relay - Higher current consumption but no voltage drop. The relay only applies power to the circuit if the proper polarity is applied. D1 protects the relay coil from reverse current.

 

[Robert]

Filtering depends on the expected power supply, and requirements of the circuit.

A DC powered, low current, purely analog circuit typically requires less filtering than a circuit that consumes high current or has digital circuitry.

Most modern power supplies already provide a fair amount of filtering ahead of the pedal.


(Contractors just showed up, I'll elaborate a bit more shortly unless someone beats me to it)

 

[vigilante398]

A much less common method is to use a relay. I've never seen this done in a pedal, it's more of an industrial / consumer electronics thing. I've seen this done in a few arcade / pinball machines over the years as well.

Relay - Higher current consumption but no voltage drop. The relay only applies power to the circuit if the proper polarity is applied. D1 protects the relay coil from reverse current.

View attachment 111610

Effectrode is the only pedal company I'm aware of that uses a relay for polarity protection.

 

[falco_femoralis]

Filtering depends on the expected power supply, and requirements of the circuit.

A DC powered, low current, purely analog circuit typically requires less filtering than a circuit that consumes high current or has digital circuitry.

Most modern power supplies already provide a fair amount of filtering ahead of the pedal.


(Contractors just showed up, I'll elaborate a bit more shortly unless someone beats me to it)

Many designs I've seen have a 100uf electrolytic cap placed from the pos rail to ground. Is this called a filter cap or a bypass cap? (Trying to clarify terminology)

Sometimes there is also a 100nf film or MLCC in parallel with the 100uf, and sometimes the electrolytic value is smaller, like 47uF, or larger, like 220uF.

One circuit I'm tracing (deadendfx mangel wurzel) places an additional 100nf cap right next to the 9v pin of an op amp. I've read placing the filter cap very close to the op amp is best, but is it overkill to have two 100nf filter caps, for example, one at the power input and another at the op amp?

It seems to me that the variance in values from one filter cap to another is up to the electronic background of the designer rather than a hard fast set of rules. Is this true?

 

[jwin615]

Filtering at the opamp is generally considered a best practice. But, it's a pedal not a medical device/aviation equipment etc.
Added filtering at the opamp would be good in circuits with say clock signals. The considerations made in a TS clone vs a BBD delay can/should be different. A sloppy layout and minimal filtering in a TS will likely work. A sloppy bbd layout will likely yield clock noise or oscilations in the audio.
The caps at the input do 2 things,
1)filter hf noise to ground
2)smoothe DC ripple

 

[zelcanada]

(I realize you're focusing more on filtering than polarity protection, but I had this ready to post in another section)

The three most common methods of polarity protection I see are:

Series - The diode is in series with the load/circuit and blocks any reverse voltage/current. No current flows so no damage is done to the circuit, it simply doesn't work until the correct polarity is applied. The negative is that the series diode drops a small amount of voltage (forward voltage of the diode). This is generally not an issue. A schottky diode can be used to reduce the voltage drop, however these typically have a lower current capability and reverse breakdown voltage. (It takes less current / voltage to damage it)
View attachment 111606

Parallel / Shunt (Diode to ground)- The circuit receives the full applied voltage, no loss of voltage (forward voltage drop) across the diode. If reverse polarity is applied the diode typically shorts to ground, making the pedal effectively "broken" to the average consumer who can't replace it.

View attachment 111605

MOSFET - Similar to series, but a MOSFET is used instead due to it's lower forward voltage drop at the cost of a slightly more complex/expensive circuit.

View attachment 111611


A much less common method is to use a relay. I've never seen this done in a pedal, it's more of an industrial / consumer electronics thing. I've seen this done in a few arcade / pinball machines over the years as well.

Relay - Higher current consumption but no voltage drop. The relay only applies power to the circuit if the proper polarity is applied. D1 protects the relay coil from reverse current.

View attachment 111610

This is super helpful! Thanks!

 

[zelcanada]

Filtering depends on the expected power supply, and requirements of the circuit.

A DC powered, low current, purely analog circuit typically requires less filtering than a circuit that consumes high current or has digital circuitry.

Most modern power supplies already provide a fair amount of filtering ahead of the pedal.


(Contractors just showed up, I'll elaborate a bit more shortly unless someone beats me to it)

So considering I'm new to this, any pedal I build right now is going to be relatively simple, DC powered, and analogue (for now)...

You've explained really well what the diode is for. Thanks for that... What about the capacitors. For example, here are two filtering sections I've seen:




In both cases, we have a parallel diode, but I now understand what it's doing there. Both seem to also have a 100u cap... Is that what's filtering the noise? And why does one have a resistor, whereas the other has a second cap?

Thanks!

 

[JTEX]

The second picture will do a much better job of filtering power supply noise because the series resistor R7 forms a low pass filter with C4. This comes at the cost of a slight voltage drop across R7. With most pedals drawing only a few mA though, the drop is very minimal: 10 ohms x 10mA = 100mV, as an example.

These days most power supplies are of the switching variety, not the old transformer+rectifier bridge type (which made 100Hz or 120Hz ripple). Switching power supply noise is usually way outside the audio band (such as 100kHz or more), so even if there's some PSU noise, it won't be audible. It can still interfere with high frequency clock signals and such (if they're present in the pedal), causing audible artifacts in the audio band. Say, your PSU switches at 100kHz and you have some clock at 105kHz > there will be a 5kHz "beating" between the two frequencies that is definitely audible.

 

[drew.spriggs]

I always just keep things simple - schottky for reverse polarity protection, then ferrite bead + caps for filtering. Most power supply noise these days is from switching instead of mains, so stacks of resistance isn't necessary for a HPF. Ferrite bead works to suppress switching noise without really dropping any DC over them. The 100nf MLCC really isn't necessary either as electros these days are pretty good at high frequency, but it's like 4c so easy enough to include just in case. It's not a bad idea to have the LED connected directly to the DC jack to make sure these currents aren't also coming through the schottky (which can increase voltage drop OR can occasionally be audible when switched on), but some LED's will die with over 5v of reverse polarity so I usually put it after with a 10k CLR to make sure LED currents stay low.

If there is something digital happening on the board, you can help keep this noise out of everything else by separating each power feed with a small value resistor (100R or so). Include another electro after this so they're not drawing current spikes from the main filter cap.

If this is going on a PCB, good way to keep noise lower is keep those caps off the ground plane, and run a trace directly to the DC jack. You then tie the ground plane and trace together at a single pad, which is where I add a pogo pin that attaches to the enclosure. Main currents are kept separate from signal grounds, reducing potential noise.

The other thing people miss often is decoupling IC's with a MLCC (either both rails to ground, or +V to GND) as physically close to the pins as you can get - can go a massive way to reducing oscillation if you've got a particularly gainy circuit.