Power Supply Filter Anatomy

[POTL]

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.

Why 1n4001 Schottky here will not work?

about the point I know - I missed, but tomorrow put


that offer to do with an electrolyte?
I reviewed a lot of circuits 100u and the electrolyte film 100n are in virtually every scheme

[POTL]

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

huh
I have always believed that the anti pop enough to put 1M resistor (although it is now more likely to give 2M2) at the input of the circuit and it should be enough 

[ElectricDruid]

Why 1n4001 Schottky here will not work?

1N4001 isn't a schottky diode. It's a power diode. It's the right thing for the diagram you posted. The schottky diode is better in *series* because it has a lower voltage drop.

I reviewed a lot of circuits 100u and the electrolyte film 100n are in virtually every scheme

It should be a 100u electrolytic and a 100n ceramic. A film cap isn't appropriate in this situation. As others have said, the 100uF provides heavy power supply filtering, but its high frequency performance isn't that good. The 100n ceramic (and ceramic is better than polyester film or polypropolene film in this application) is there because it has extremely good high frequency performance and will send any radio frequency interference on the power input directly to ground.
The *exact* values aren't that important, but it's important that there's one large electrolytic cap to act as a power supply filter, and another much smaller ceramic cap to dump high frequency interference. You wouldn't notice if you used 22uF and 220n, or 47uF and 47n, or 47uF and 10n, (to choose a few random variants) although in theory they're all different - just not enough for humans to care.

HTH,
Tom

[POTL]

It should be a 100u electrolytic and a 100n ceramic. A film cap isn't appropriate in this situation. As others have said, the 100uF provides heavy power supply filtering, but its high frequency performance isn't that good. The 100n ceramic (and ceramic is better than polyester film or polypropolene film in this application) is there because it has extremely good high frequency performance and will send any radio frequency interference on the power input directly to ground.
The *exact* values aren't that important, but it's important that there's one large electrolytic cap to act as a power supply filter, and another much smaller ceramic cap to dump high frequency interference. You wouldn't notice if you used 22uF and 220n, or 47uF and 47n, or 47uF and 10n, (to choose a few random variants) although in theory they're all different - just not enough for humans to care.

Thank you
my English is lame (yet) for this I can take a little incorrect information 

1)about the 100n film - I saw at the JHS - they always put the film in the power filter and I thought so right(with your words
I realized, too, 100n value is not the main)
2)values ​​of electrolytes, about 100u I don't quite understand, you mean there's no need to put so much importance and you can get the smaller values ​​such as 47u and 22u
you say about this?
just as with classic circuits(TS808, CE2, FL9,DM2 and many many others) and in current use is boutique(JHS, Wampler) 100u and 100n capacitors (film only)
3) about diodes I realized my mistake
Can you briefly describe what are their benefits as well as the fact that the scheme (or, rather, a diode) is preferable from the standpoint of protection components?

P.S.
I apologize for meticulousness, just want to do everything on a full understanding of, and not just copy

[GibsonGM]

1) I haven't had any trouble with film capacitors for power supply bypassing as long as they were rated high enough...then, I've also totally omitted them and had no problems, either, so...

2) Yes, the value of capacitor is not TERRIBLY important, but you must understand the formula that I provided you above...the level of filtering that you will achieve changes if you change the capacitance.    Using that formula and the other information you have about your circuit, you can determine how good the filtering will be.  If in doubt, use a bigger cap (up to 100u).   You can work this out using a simulator like LT Spice, too, and see how changes affect your circuit!

This filtering is not very critical....most often, lower values of capacitor work fine.     It depends on how much ripple your power supply has when you plug it into your effect!   A battery will have none, and those caps do nothing.

Likewise, if your 'wall wart' power supply is unregulated and has much ripple, you may need to make MORE sections than those shown (a "pi network"), or even refrain from using it because it is just too noisy to be worth the work.   

I recommend you search the internet and read about beginner power supply design, and ripple voltage...this will help you understand how these components work!   

[LightSoundGeometry]

Why 1n4001 Schottky here will not work?

1N4001 isn't a schottky diode. It's a power diode. It's the right thing for the diagram you posted. The schottky diode is better in *series* because it has a lower voltage drop.

I reviewed a lot of circuits 100u and the electrolyte film 100n are in virtually every scheme

It should be a 100u electrolytic and a 100n ceramic. A film cap isn't appropriate in this situation. As others have said, the 100uF provides heavy power supply filtering, but its high frequency performance isn't that good. The 100n ceramic (and ceramic is better than polyester film or polypropolene film in this application) is there because it has extremely good high frequency performance and will send any radio frequency interference on the power input directly to ground.
The *exact* values aren't that important, but it's important that there's one large electrolytic cap to act as a power supply filter, and another much smaller ceramic cap to dump high frequency interference. You wouldn't notice if you used 22uF and 220n, or 47uF and 47n, or 47uF and 10n, (to choose a few random variants) although in theory they're all different - just not enough for humans to care.

HTH,
Tom

awesome ! im going to add a cermaic cap to my PS and should start doing the AMZ mod on every build since its only 2-3 parts to do

[POTL]

1) I haven't had any trouble with film capacitors for power supply bypassing as long as they were rated high enough...then, I've also totally omitted them and had no problems, either, so...

2) Yes, the value of capacitor is not TERRIBLY important, but you must understand the formula that I provided you above...the level of filtering that you will achieve changes if you change the capacitance.    Using that formula and the other information you have about your circuit, you can determine how good the filtering will be.  If in doubt, use a bigger cap (up to 100u).   You can work this out using a simulator like LT Spice, too, and see how changes affect your circuit!

This filtering is not very critical....most often, lower values of capacitor work fine.     It depends on how much ripple your power supply has when you plug it into your effect!   A battery will have none, and those caps do nothing.

Likewise, if your 'wall wart' power supply is unregulated and has much ripple, you may need to make MORE sections than those shown (a "pi network"), or even refrain from using it because it is just too noisy to be worth the work.   

I recommend you search the internet and read about beginner power supply design, and ripple voltage...this will help you understand how these components work! 

Thanks

tomorrow will test LT spice

[duck_arse]

.... I apologize for meticulousness .....

never apologise for meticulousness.

also, in opamp circuits, each opamp package should have its own 100nF ceramic cap soldered as close as possible to its supply pins.

[GibsonGM]

.... I apologize for meticulousness .....

never apologise for meticulousness.

also, in opamp circuits, each opamp package should have its own 100nF ceramic cap soldered as close as possible to its supply pins.

+10,000

[merlinb]

also, in opamp circuits, each opamp package should have its own 100nF ceramic cap soldered as close as possible to its supply pins.

There is no need for this in audio circuits. Several opamps can be supplied from one electrolytic.

[ElectricDruid]

2) Yes, the value of capacitor is not TERRIBLY important, but you must understand the formula that I provided you above...the level of filtering that you will achieve changes if you change the capacitance.    Using that formula and the other information you have about your circuit, you can determine how good the filtering will be.  If in doubt, use a bigger cap (up to 100u).   You can work this out using a simulator like LT Spice, too, and see how changes affect your circuit!

There's a calculator for these simple RC filters over on my website (and many others on the web):

http://electricdruid.net/rc-filter-calc/

If you enter "47R" in the Resistor box, and "100u" in the capacitor box, the results shown us that the filter has a cutoff of 33.9Hz and any ripple above that frequency will get attenuated. That's low enough to provide at least some protection from AC ripple, but not much.

LTSpice lets you do other clever stuff beyond just see a few numbers, but sometimes just a few numbers is enough.

Tom

[POTL]

thanks friends
Now I will shake LT spice and taste

About the ceramics, the funny thing is that we have a good ceramics (Murata, Japan) with a value worth more than a good film (Epcos Germany) 17 cents vs 12 cents LOL
in fact the price does not matter

Can you explain again about the differences 1N4001 and 1N5817
a diode which gives preimoschestva and that reliable protection?


I realized that the right way to connect such

but which option is reliable or 1N4001 1n5817?

[R.G.]

also, in opamp circuits, each opamp package should have its own 100nF ceramic cap soldered as close as possible to its supply pins.

There is no need for this in audio circuits. Several opamps can be supplied from one electrolytic.

Well, maybe. There is a difference between what is provably rock solid and what you can get away with.

National Semiconductor, back in the days when they largely were The Opamp Supplier was pretty clear: one ceramic cap from each power supply pin to ground, as close to the chip as you can get it. This is a difficult thing to do on many layouts, so they allowed that maybe one ceramic per every two chips if you can arrange the power supply distribution grid (!!) so they're both close to the cap. "Close" meant less than half an inch of PCB trace.

Modern electrolytics do keep getting better, and help with this, and in small circuits like pedals, there is often not a lot of interfering junk on the power rails or high currents to cause power-rail feedback.

So yes - you often get away with several opamps supplied from one electrolytic. Sometimes you don't. I have personally had to go back and put in power bypasses where I was not free enough with them. With my boss looking over my shoulder and asking when the design would be (re)done as the production line was waiting. I have all these scars - the very most effective form of education. 

There is a difference between what's rock solid and what you can get away with. Education and skill is required to know the difference in the two conditions.

EDIT: Just as I hit 'save' it occurred to me to say that it's hard to conceive of a power supply being too well decoupled, but easy to come up with conditions where it's not decoupled enough. As always, defining "enough" is complex and variable. Over-decoupling is benign. Under-decoupling is not.

[ElectricDruid]

Can you explain again about the differences 1N4001 and 1N5817
a diode which gives preimoschestva and that reliable protection?

I realized that the right way to connect such

but which option is reliable or 1N4001 1n5817?

The way on the left is the "in series" method, and the way on the right is the "in parallel" method.

For the "in series" method, you need a diode that drops as small a voltage as possible when it is conducting. Imagine you used a blue LED in this position - you might have a 3V drop across the diode, which would only leave 6V for your circuit!
So diodes with a small forward voltage are preferred, that's why schottky diodes are good, and 1N5817 is an example. It has a forward voltage of only 0.45V up to 1A. This is good, but there is always a small loss of voltage because of the diode.
If the supply is reversed, no current will flow and the circuit is protected.

For the "in parallel" method, the rules are different. The forward voltage across the diode doesn't matter, since the diode is reverse-biased (is backwards) during normal use, so it is not conducting. The circuit will get the whole 9V, with no loss because of the diode. If the power supply is reversed, the diode will conduct and no current will go to the circuit. Since the diode is acting almost like a piece of solid wire, it will draw a *large* current and is likely to draw the maximum the power supply can give. If the diode is under-rated, the diode will burn up, and then your circuit isn't protected any more. For this reason, heavy-duty power rectifier diodes are used for this style. 1N4001 is an example, but you might see its even-more-powerful brothers and sisters in the series: 1N4002, 4003, 4004, etc.
The problem with this method is that if the power supply isn't protected against short circuits, it is likely that reversing the supply will kill the power supply rather than the effect circuit.

Both methods have advantages and disadvantages. The in-series method is safer, but you lose voltage. The in-parallel method will save you circuit but might kill your power supply, but could be the only option if you need to maintain the full 9V supply.

I hope this explains it a bit more.

Tom

[amptramp]

I realized that the right way to connect such

but which option is reliable or 1N4001 1n5817?

Note that this is true if you use centre positive power input.  For historical reasons most pedal power inputs are centre negative, so the diodes would have to be reversed.  Schottky diodes would be in use everywhere if it wasn't for two characteristics:

1. Low reverse voltage limits
2. High reverse leakage.

They have gotten a lot better recently but they still do not compare with junction diodes in these characteristics.

[R.G.]

It's important to remember that even tiny ground shifts, the size of a Schottky diode forward voltage, or even a germanium forward voltage, can be unendurable in pedals. Clicks, pop, hum and hash can result from pedal "grounds" being offset from one another, even by a little.

So in wiring series diodes, you must get them in series with the NON ground connector to the power supply.

The ground offset problem is what kept me from popularizing a circuit with four MOSFETs in an active rectifier bridge to make pedals completely insensitive to polarity, even including AC inputs. Alas, the millivolts of offset in ground made this untenable for most pedals.

[POTL]

Can you explain again about the differences 1N4001 and 1N5817
a diode which gives preimoschestva and that reliable protection?

I realized that the right way to connect such

but which option is reliable or 1N4001 1n5817?

The way on the left is the "in series" method, and the way on the right is the "in parallel" method.

For the "in series" method, you need a diode that drops as small a voltage as possible when it is conducting. Imagine you used a blue LED in this position - you might have a 3V drop across the diode, which would only leave 6V for your circuit!
So diodes with a small forward voltage are preferred, that's why schottky diodes are good, and 1N5817 is an example. It has a forward voltage of only 0.45V up to 1A. This is good, but there is always a small loss of voltage because of the diode.
If the supply is reversed, no current will flow and the circuit is protected.

For the "in parallel" method, the rules are different. The forward voltage across the diode doesn't matter, since the diode is reverse-biased (is backwards) during normal use, so it is not conducting. The circuit will get the whole 9V, with no loss because of the diode. If the power supply is reversed, the diode will conduct and no current will go to the circuit. Since the diode is acting almost like a piece of solid wire, it will draw a *large* current and is likely to draw the maximum the power supply can give. If the diode is under-rated, the diode will burn up, and then your circuit isn't protected any more. For this reason, heavy-duty power rectifier diodes are used for this style. 1N4001 is an example, but you might see its even-more-powerful brothers and sisters in the series: 1N4002, 4003, 4004, etc.
The problem with this method is that if the power supply isn't protected against short circuits, it is likely that reversing the supply will kill the power supply rather than the effect circuit.

Both methods have advantages and disadvantages. The in-series method is safer, but you lose voltage. The in-parallel method will save you circuit but might kill your power supply, but could be the only option if you need to maintain the full 9V supply.

I hope this explains it a bit more.

Thanks, everything fell into place.

Method with 1N4001 more like it.
Since I do not lose volts.
Especially I like the idea in OCD with the possibility of using both 9 and 18 volts
Tom

[POTL]

2) Yes, the value of capacitor is not TERRIBLY important, but you must understand the formula that I provided you above...the level of filtering that you will achieve changes if you change the capacitance.    Using that formula and the other information you have about your circuit, you can determine how good the filtering will be.  If in doubt, use a bigger cap (up to 100u).   You can work this out using a simulator like LT Spice, too, and see how changes affect your circuit!

There's a calculator for these simple RC filters over on my website (and many others on the web):

http://electricdruid.net/rc-filter-calc/

If you enter "47R" in the Resistor box, and "100u" in the capacitor box, the results shown us that the filter has a cutoff of 33.9Hz and any ripple above that frequency will get attenuated. That's low enough to provide at least some protection from AC ripple, but not much.

LTSpice lets you do other clever stuff beyond just see a few numbers, but sometimes just a few numbers is enough.

Tom

Wow it is great

[GibsonGM]

I should have mentioned that you can use an online calculator!  There is going to be one available for EVERY electronics calculation you must do.   Search engines will help you find one in the future    

[POTL]

I should have mentioned that you can use an online calculator!  There is going to be one available for EVERY electronics calculation you must do.   Search engines will help you find one in the future   

Yes, I agree it is good Google search
but usually you meet a lot of general information without specifics, and has to spend a lot of time.
In fact, I'm glad that this forum is friendly and responsive, at the largest forum in my country gave me a link to an old book in half a thousand pages and were told nothing