Electrolytic capacitor direction

[SprinkleSpraycan]

My understanding has been that electrolytic caps HAVE to go one direction but I'm seeing in more than one application that isn't true. So, for instance, in the case of the tube bender side of a hoof reaper (schematic included) how does the "backwards" cap c2 affect the sound?

[antonis]

IMHO, it has nothing to do with "sound"..

C2 positive leg looks to about 1/10 of power supply where negative leg looks to GND (via R1 pull-down resistor)..

[anotherjim]

^ That is the correct way for the input cap. There was a time when I didn't know any better and I thought the + end should always face the input and the - end the output because I had seen schematics where the coupling caps were all that way and then wondered why the volume control on my guitar crackled when turned. Wrong of course, the - end always faces closest to ground which a passive guitar having no DC voltage surely is.

[duck_arse]

look at C6 - is it forwards or backwards?

[nocentelli]

look at C6 - is it forwards or backwards?

I thought that was the problem cap (C6)

[antonis]

C6 orientation could be reversed with no issue..

For circuit iddling (no signal), Q2 is in cut-off region (no Base voltage) so cap negative leg is "floating"..
In case of D1 was a resistor (or a Ge diode) and considering Q2 leakage current significant, C6 polarity should be reversed..

[Rob Strand]

For C6 I agree with antonis.

It's not a good idea applying a recipe to determine cap polarity.  You really have to know what the actual voltages are at each end of the cap.  Inter-stage caps are the most likely place where errors occur.   

The only recipe you can eyeball reliably for an amplifiers is:
If you look at a circuit and one end of the cap has a resistor to ground then the voltage at that point is zero.   
The other end of the cap is either zero (due to another resistor to ground), or a positive voltage due to some voltage in the circuit (assuming a negative ground circuit).

In your circuit this rule would apply to: C2, C5, C6, C10 and even C4, C7.
The only tricky one is C6 but even then one side has a path to ground.

When you have diodes and switching power supply circuits that rule can be wrong because you can generate negative voltages.

[diffeq]

C6 and C2 is 1uF, you could use non-polar cap for peace of mind.

[SprinkleSpraycan]

There was a time when I didn't know any better and I thought the + end should always face the input and the - end the output

That's where I'm at then. Thanks everyone for helping me level up.

[Phend]

Look at C12 in these two Foxx circuits.
(Tone pot)
Comments ?

[FiveseveN]

JD's version is correct, but the cap surely doesn't need to be that large, therefore not polarized.

[SprinkleSpraycan]

Look ar C12 in these two Foxx circuits.
(Tone pot)
Comments ?

C12 will have less/more resistance to ground depending on where the tone controls is set. Does going through the c11 cap make a difference? I feel like you would tie negative to r22 so there is a consistent ground path. Is that right?

[antonis]

Voltage difference between C12 legs is about 500mV (Q4 Base bias - 1 Ge diode drop..)
Taking into account that most electros can withstand more than 1V reverse voltage, both orientations should be considered OK..

P.S.
FWIW, "safe" reverse voltage value coincides with that of exceeding the maximum rated working one..

[FSFX]

FWIW - This information about ratings is from an electrolytic capacitor manufacturer:

[duck_arse]

Look at C12 in these two Foxx circuits.
(Tone pot)
Comments ?

Phend, Phend, Phend - what abot the Sustain cap, no mention? any time you see a cap plus pointing towards a cap minus without some DC source in between, youse needs to ask yourself .....

[JustinFun]

If in doubt, stick in a temporary ceramic and measure the voltage to ground either side!

[anotherjim]

FTM scheme is a perfect example of every cap + facing the input although it could be a drawing error. The DC polarity in circuit rule seems simple and obvious but I don't think it was ever spelt out in any of the DIY electronics mags and primer books I read in my youth.

Sometimes, they can't make up their minds although the polarity in those cases probably doesn't matter 'cause no DC difference or very little such as in the Roland 808...

... look at amps 16 & 19. Even spec 6.3v caps on a +/-15v powered circuit since the feedback caps should never see much difference in voltage across them.

[pinkjimiphoton]

i'm being too lame to read the whole thread, but basically, in the end with any electro or marked cap, you want the negative side always <or outer wraps of the cap> to go to GROUND potential to help nuke noise as well as protect the cap. the outer wrap helps shield the cap from noise and hash.
positive side needs to go toward the voltage potential... remember, ground is often just the neutral point between positive and negative potentials, so in some cases, like pnp fuzzes, you DO need to reverse the cap polarities with the electros... sorry if oversimplifying, trying to make it dude speak so newbs can understand it <of which i still am one all these years later, so...>

basically what i do if i have a schematic, is look to see which side is gonna have a positive voltage flowing to it, and that gets the +
usually its either power supply filtering or coupling cap in a lot of circuits, so doing it this way is fairly safe in most circumstances.

or use your meter if you have no schematic, and see what the voltages are where the cap needs to go, or, with stuff off, read the resistance to ground from one or the other nodes if unknown... one side will show significantly less resistance to ground than the other most of the time, and that would get the negative lead.
so far, this has worked for me when i need to sub something in cuz i'm out of something or need a specific value or voltage or whatever.
also remember, you can always fake a np cap by tieing the cathodes of two electros together, and just using the anodes.
in 9v circuits you can usually get away with just the caps themselves... say ya wanted 1u, np... you don't have, so you can use two 2.2's negative to negative and get an approximate value of 1.1uf, which is within a 10% tolerance and should work fine.
higher voltages, you may wanna parallel each cap with a resistor too, maybe 22k or so probably would do, but i'd have to look it up to be sure or mess with it. its just to even out the voltages so one cap doesn't go over voltage and smoke itself.

sorry for the book.

final thought... don't always trust the schematics, cuz often there may be a mistake in orientation on the schem, the layout, even the pcb.
check it out first before building, and it becomes pretty clear USUALLY which orientation is gonna be right, but remember, it will vary depending on function... power supply filtering may be different from coupling caps.

peace

[Rob Strand]

I don't think it was ever spelt out in any of the DIY electronics mags and primer books I read in my youth.

I like the cap symbol with the + marking more than any of the old symbols.  Some of the older cap symbols were ambiguous unless that was the only symbol you knew - I'm pretty sure a few articles and commercial schematics flipped the meaning on occasion (I would look at the power supply cap direction first).

Sometimes, they can't make up their minds although the polarity in those cases probably doesn't matter 'cause no DC difference or very little such as in the Roland 808...

If you want to be pedantic some of the bipolar opamps have DC input bias current and that causes a real DC voltage in the opamp circuit so you can split hairs that way.  For things like the NE5534 the input bias is quite high; some audio circuits are end up with 1V DC offset and you end up getting close to the recommended limit FSFX posted (voltages over about 1V de-plate the electrolytic capacitor).

For input and output caps with 0V DC you could argue the device is more likely to be connected to a positive voltage so the cap + should go toward the external connections.   Even I don't always follow that rule, probably brainwashed from the old transistor circuits.

[ElectricDruid]

+1 what FiveSeven said:

Who needs a coupling cap that large anyway?!? Thus far, I've always been able to design out any need to go over 470n or so. Even 18K impedance gets the rolloff down to 20Hz, so it's generally not hard. Just "Avoid electrolytics in the signal path" is a simple and workable strategy!