ideal power section?

[jebop]

Hi all,
I was wondering if there is any standard/ideal power section that works for most 9v pedals.
For example, the Rat's power section (from electrosmash.com):
https://www.electrosmash.com/images/tech/pro-co-rat/pro-co-rat-power-supply.png
Also, what makes a power section good/bad?
Thanks,
Miles

[bowanderror]

I was trying to figure this out as well and have a running note with some learnings. Hopefully some others can add more details & correct mistakes:

• Shunt Polarity Protection Diode - Ex: 1N4001 from V+ (cathode) to ground - Prevents a wrongly-connected power supply from blowing up your circuit by shunting current to ground - No voltage drop, but doesn't provide any direct protection to the circuit & can't prevent overvoltage
• Series Polarity Protection Diode - Ex: 1N5817 in series w/V+ - Prevents a wrongly-connected power supply from blowing up your circuit - Directly protects the circuit it feeds, but has ~0.3V drop across it, doesn't prevent overvoltage
• Overvoltage Protection Diode - Ex: 9.1V Zener - Blocks momentary spikes from PSU - Can choose zener voltage based on desired supply voltage (9.1V/12V/18V/etc.), good for protecting expensive components like charge pumps.
• Filter Capacitors - Ex: 100uF or 0.1uF from V+ to ground - Removes noise, prevents interference from other parts of the circuit - Large (33-220uF) caps help with LF noise, small (0.1uF) caps for RF interference. Ideally you want separate smaller filter caps for each opamp, placed as close to where power enters the opamp as possible - How to calculate best values for a given circuit?
• Current-limiting resistor - Ex: 10R-100R in series w/V+ - Prevents circuit from pulling ridiculous amounts of current & blowing something up - Will cause small voltage drop & restrict max current draw. How to determine when circuits require one & what size?
• Bias Voltage Divider - Ex: 2x series 10k resistors from +9V to ground, generating +4.5V Vb (1/2 Vcc) from the node between them - Required for biasing opamp inputs - High value resistors will restrict current & voltage drop can be affected by current draw, but OK for high impedance loads? Small values will use up more current, but may be necessary for low impedance loads? Small values For opamps, feedback resistors should be much larger than bias voltage divider resistors.
• Bias Voltage Divider Filter Caps - Ex: 10uF from Vbias to ground - Removes noise & prevents interference - Is 10uF standard for +4.5V bias voltage?
• Linear Voltage Regulator - Ex: 7805 +5V Linear Voltage Regulator for PT2399 circuits - Provides an exact & consistent voltage, which is not affected by current being pulled by other circuit elements - Require input voltage higher than the regulated output voltage to prevent dropout. See datasheet for specs, filter caps, & other circuit components
• Resettable Fuse - Ex: ? in line with +9V - Switches off power if current draw goes above rating - What's a good polyfuse for pedal circuits? Any other things to keep in mind when using them?
• Charge Pump - Ex: ICL7660S IC which converts +9V to +18V - Voltage doubler generates a higher voltage power rail for running pedals with a higher headroom or more gain. Depending on configuration, can also be a voltage inverter (+9V to -9V) - Ensure oscillator frequency is above audio range (>20kHz). Most are limited to 30mA current draw, LT1054 can be used up to 100mA - Is this only for voltage doubling operation?

[anotherjim]

My favourite is 2: Series diode. Maybe any 9v pedal that can't tolerate even the 0.7v drop from a 1N4001 needs to be a 12v or 18v pedal? However, the diode blocks AC signal/LFO currents via the power supply so adequate (larger) smoothing/bypass caps need to be fitted on the pedals own power supply distribution. This can lead to serious instability with small chip amplifiers. There are designs that only work because they have no series diode and depend on the external PSU capacitors for AC bypass.

9: Polyfuse is interesting, however, the reset time can be an unknown quantity. Can we afford the time to wait on stage?
With potential dangers with high capacity lithium cells, the polyfuse is good protection, but we don't really have that situation.

[garcho]

Re: #4 and #5

Those caps that you correctly said should be close to the IC power pins aren't there for filtering. They act as current reservoirs for the local IC, helping maintain smooth operation despite whatever is going on in the IC, power supply and buses.
Speaking of filtering, that's maybe the main reason for that 100R series V+ resistor. In combination with a large value cap it makes a low pass filter (think of DC as very low frequency) for any AC at the power supply.

[Rob Strand]

9: Polyfuse is interesting, however, the reset time can be an unknown quantity. Can we afford the time to wait on stage?
With potential dangers with high capacity lithium cells, the polyfuse is good protection, but we don't really have that situation.

My call on this one is the Polyfuse is used as part of method 2.   The diode protects the circuit the Polyfuse limits the current (and protects the diode).   As a side effect it limits the short-circuit current in the circuit but really, if you put 60V into a pedal the current will not trip the Polyfuse and a good amount of the circuit will fry.

[marcelomd]

You can also use a P MOSFET for for polarity protection. More components than with a single series diode, but almost no voltage drop. Something like this:



In my ideal world, I'd use a MOSFET for polarity protection with some kind of slow start, to prevent pulling a lot of current.

I like a 100uF+100nF capacitors for the main power, +47uF for Vbias and a 100nF for every opamp. I got some sparks when connecting the power supply, because the capacitors pull too much current. Then I started using a 33 ohm resistor before the capacitors. This is probably too much capacitance, but I like the lack of noise.

[niektb]

Depending on the FET you use, you dó have some ON-resistance. (I use a BS250 which is 8Ohms) but generally the heat capabilities are much better than a standard resistor so it's less likely to blow (i've seen series resistors blow because they got too hot). And the added filtering is nice. I also use the same cap values, no problems with sparks 

[antonis]

In depth dealing with any pedal power supply brings me in mind equivalent medical instance where you care more for good anaesthesiologist than surgeon..
(which, IMHO, is essentially true and correct..)

[Rob Strand]

I like a 100uF+100nF capacitors for the main power, +47uF for Vbias and a 100nF for every opamp. I got some sparks when connecting the power supply, because the capacitors pull too much current. Then I started using a 33 ohm resistor before the capacitors. This is probably too much capacitance, but I like the lack of noise.

You can add soft-starting to the circuit you gave but it complicates it a bit.   You need soft starting (or your current limit resistor) so the MOSFET doesn't end up failing due to the excessive turn on currents.  There's actually dedicated chips out there now which offer this type of input configuration together with soft starting.

[Gargaman]

Re: #4 and #5

Those caps that you correctly said should be close to the IC power pins aren’t there for filtering. They act as current reservoirs for the local IC, helping maintain smooth operation despite whatever is going on in the IC, power supply and buses.
Speaking of filtering, that’s maybe the main reason for that 100R series V+ resistor. In combination with a large value cap it makes a low pass filter (think of DC as very low frequency) for any AC at the power supply.

Re: #4
Will 100R and 47uF form a low pass filter with frequency cut around 34Hz?
You get the same frequency If you use 47R and 100uF. Right?

[antonis]

Will 100R and 47uF form a low pass filter with frequency cut around 34Hz?
You get the same frequency If you use 47R and 100uF. Right?

Same as you use 10R and 470μF or 470R  and 10μF or any other values combination of 0.0047 time constant (RxC product)

The lower the series resistor value the lower the voltage drop and power dissipation across it but the higher the capacitor value (and space & cost..)

[ElectricDruid]

I don't know about "ideal power section", but I've pretty much standardised on the following design from my page of basic stompbox elements:

https://electricdruid.net/wp-content/uploads/2018/12/StompboxElements.jpg


This covers several of the things previously mentioned: series diodes (schottky to reduce volt drop), series resistor to act as lowpass filter with the caps, decent cap values to give at least some hum roll-off (this is 34Hz), and a 100n ceramic in parallel to dump digital or rf noise on the power to ground.

Ultimately you just choose something you're happy with as a balance between cost/complexity and ultimate safety/features and go with it.

[Fancy Lime]

+1 to what Tom said. This is what I use also and it has served me well. On circuits with low current consumption, I increase the resistor to 100R. In high gain circuits, I may increase the electrolytic cap to 220u.
Andy

[antonis]

I've pretty much standardised on the following design from my page of basic stompbox elements:

https://electricdruid.net/wp-content/uploads/2018/12/StompboxElements.jpg

One of these days we shall talk about that "lazy" BJT buffer configuration, Tom..

:cheers: :cheers: :cheers:

[ElectricDruid]

One of these days we shall talk about that "lazy" BJT buffer configuration, Tom..

Oh god no! Then I'll have to try and remember where I stole it from!

TBH, the FET buffer I've never yet used ('cos TH FETs are now awkward to get, so I avoid them if possible) and the BJT buffer only appears once (as a input buffer on Flangelicious). Mostly I stick to op-amps, and even numbers of them at that!

[vigilante397]

• Current-limiting resistor - Ex: 10R-100R in series w/V+ - Prevents circuit from pulling ridiculous amounts of current & blowing something up - Will cause small voltage drop & restrict max current draw. How to determine when circuits require one & what size?

THIS IS NOT A CURRENT LIMITING RESISTOR.

I have NEVER seen a pedal circuit that requires an actual current limiting resistor. Op-amps, transistors, SMPS, etc. will pull a finite amount of current, they are not designed to pull as much as they can (which would have been a terrible design), and if they were, a 10R resistor wouldn't save them.

As alluded to above, the series resistor is the R component pf an RC low pass filter. That is all. It is 100% there for filtering the DC input . This is one of my biggest pet peeves, I apologize for the rant

[antonis]

THIS IS NOT A CURRENT LIMITING RESISTOR.

Not even in case of relatively big reservoir cap..?? 
(I've seen quite a few pedals with 1000μF or bigger caps..)

[Fancy Lime]

• Current-limiting resistor - Ex: 10R-100R in series w/V+ - Prevents circuit from pulling ridiculous amounts of current & blowing something up - Will cause small voltage drop & restrict max current draw. How to determine when circuits require one & what size?

THIS IS NOT A CURRENT LIMITING RESISTOR.

I have NEVER seen a pedal circuit that requires an actual current limiting resistor. Op-amps, transistors, SMPS, etc. will pull a finite amount of current, they are not designed to pull as much as they can (which would have been a terrible design), and if they were, a 10R resistor wouldn't save them.

As alluded to above, the series resistor is the R component pf an RC low pass filter. That is all. It is 100% there for filtering the DC input . This is one of my biggest pet peeves, I apologize for the rant

Alright bucko, let's do this 
OF COURSE IT'S A CURRENT LIMITING RESISTOR! THAT'S HOW RC FILTERS WORK!!!

Sorry, I usually don't tell on the internet but it looked like fun when you did it. Seriously though, isn't limiting the current rushing into the cap exactly what makes an RC LP filter do it's thing? The cap stores a finite amount of charge and the resistor regulates the current that fills it with such, no? Obviously, I am talking about AC and you are talking about DC. But if we leave out the semantic nitpickery and internet flame war douche Baggett, I agree, most circuits do not need the DC limiting resistors because the active elements and/or their implementation limits current consumption. HOWEVER, CMOS inverters pressed into linear operation are a possible exception. I have never tried if a CD4049 blows if you omit the series resistor on the power line and connect it to a low impedance power supply, but I suspect it might.

Ah, this was fun. After a year indoors, I'm itching to use my outdoor Boise again 

Cheerios,
Andy

[vigilante397]

OF COURSE IT'S A CURRENT LIMITING RESISTOR!

I mean it's a current limiting resistor in the same sense that every resistor is a current limiting resistor, so I'll meet you halfway on that one. Technically yes, at higher frequencies the resistor prevents the capacitor from charging quickly before the input changes polarity. BUT the resistor's purpose is not to "Prevent circuit from pulling ridiculous amounts of current & blowing something up". That's what I was taught when I first started building pedals, and the more I learned the less sense it made, so it's become a peeve of mine that this continues to be perpetuated.

My CMOS experience outside of school is admittedly limited, were they that poorly designed?

Sorry I yelled

[PRR]

If you use a cross-rail diode to avoid damage from reverse polarity, that resistor can "limit current" to prevent killing the diode (which leaves you unprotected).

Assuming 1N400x (1A) and 9V supply, anything over 9 Ohms will protect the diode +and+ the rest of the board. If you persist, any pedal-size resistor at nearly 9 Watts will burn up real quick. 270 Ohms 1/4 Watt will sit there too hot to touch for months.

That's hardly the most elegant way, but sometimes it is not about electronic elegance but having objective evidence to deny warranty repair. Or at least a 12-cent part rather than a whole board.