a lesson learned ;)

[pinkjimiphoton]

AC

that was the first thing i thought of as being the issue.

I wouldn't tell anyone to drop the series diode. That's the most important one for me.
Now, both the parallel diode - if its wired after the series, then it should never ever do anything, not even with a bad supply and might as well not be there.
Then if the parallel diode is before the series, it's going to smoke with wrong supply.

...but if you do put in a series diode, the pedal circuit must have a supply bypass capacitors even if a published circuit doesn't show any. Without a series diode, the circuit benefits from the capacitance fitted in the psu. With the series diode the psu capacitor is blocked by the diodes one-way action. If anyone refuses to to use series diode protection - because it made the pedal sound "funny" - you have to ask if they fitted an adequate supply capacitor in the pedal.

To re-iterate what's been said many times, old circuits like the FF ARE different.
They only ran on battery and the battery is a very good bypass capacitor. So, they did not fit a bypass capacitor. As a psu capacitor may be a long way from the pedal circuit, the circuit ought to have its own capacitor anyway. I mention supply capacitors because it matters in relation to any polarity protection system fitted.
They are simple circuits, usually with resistors in every path. Reverse power, so long as the voltage doesn't exceed the ratings of the Q's does nothing -  no damage at all. However, a supply capacitor in the pedal, since it's probably got to be a polarized type, will be damaged.
With integrated circuit chips -  any kind - there will be a low impedance path between its supply pins if polarity is reverse. The chip is usually destroyed when that happens.

hmmm, wonder if thats why some of my early things sounded different on power supplies and batteries. i remember one circuit that just plain wouldn't work right at all on a power supply, but sounded great on batteries.

still rather smoke a diode than a pedal.

the midrange booster smoked the chip in it. one of the fuzzes smoked the 4001. again, the ones with both diodes not only didn't make a loud buzz noise, but appear to have been undamaged.
i will check 'em good when i get a chance... need to debug a small "last in line" preamp i built to sit on my amp... just a basic 3 band baxendall and a buffer/boost. makes a huge diff at the end of that 20 foot run from my pedalboard, and makes it easier to tweak the overall sound of the setup..

but anyways, thanks for all the replies, i think i'm learning something.

>the one in series is what i think i'm gonna start adding again.

Do this. The voltage drop is a price worth paying, particularly if you use a Schottky - the 1N5817 is popular.

I agree.. I usually do this.  1N5817 in series, and a 1N4001 in parallel.

MC

If you've got a series diode, the parallel one is completely redundant. If it comes before the series diode, it provides no protection, but will itself burn up (potentially taking other components with it) under reverse polarity conditions. If it comes after the series diode, it serves no function at all, unless the Vrrm/Piv of the series diode is exceeded (possible with schottky's), in which case it will just burn up anyway.

I believe this info was why i stopped doing it, but it appears from my accident the second diode DID make a difference, in that those pedals appear to have been undamaged, still worked on 12vac instead of 9vdc and didn't hum or buzz.

i am gonna mark off anything for power supplies that isn't 9vdc with some shrink wrap on the plugs i think so i can't make this mistake again.

[GGBB]

1N5817 in series, and a 1N4001 in parallel.

I think a parallel diode on its own can introduce potential problems. It's essentially a one-way dead short, so without any current limiting either your unregulated or AC supply or the diode will probably fry eventually. If the diodes fries short, you now have a two-way short and the supply will go next. If the diode fries open then you no longer have the protection it was supposed to provide, circuit damage to follow. It's good for very temporary reverse polarity protection only. If you are using a series diode anyway, I think including a reverse parallel diode without current limiting actually adds risk not reduces.

[pinkjimiphoton]

but don't the two diodes together form a half wave rectifier? sorry for the dumb questions, trying to gain more understanding. in a tube amp, for instance, a dual diode rectifier tube is used, which is essentially the same thing, isn't it? or is it more like a diode clipper, cutting off the spikes to make a smoother square wave?

[Phoenix]

but don't the two diodes together form a half wave rectifier? sorry for the dumb questions, trying to gain more understanding. in a tube amp, for instance, a dual diode rectifier tube is used, which is essentially the same thing, isn't it? or is it more like a diode clipper, cutting off the spikes to make a smoother square wave?

No, it only forms a half-wave rectifier, from the series diode. The reverse bias parallel diode, if it's before the series diode, is just a short circuit and will burn up on an AC supply (potentially causing other damage), or if it comes after, it cannot conduct because the series diode is then reverse biased, so there's no current path.

[pinkjimiphoton]

forgive my stupidity, but isn't the idea of having the second one after the series diode to prevent catastrophic failure if the first diode smokes?

[Phoenix]

forgive my stupidity, but isn't the idea of having the second one after the series diode to prevent catastrophic failure if the first diode smokes?

If the first one (series) fails, the second one (reverse parallel) will DEFINITELY fail, and much faster too, it's just not a robust protection mechanism at all - it can really only protect against *BRIEF* reverse polarity from *LOW CURRENT* sources like batteries, it will not stand up to any period of reverse polarity from a power supply, even using a "high current" diode (for pedals anyway) like a 1N400x. It's been termed the "warranty indicator" for this reason - it doesn't really offer any protection, but makes it very obvious if you've used the wrong power supply.

The series diode on the other hand will ONLY fail if its Vrrm/Piv is exceeded, so maybe 20V on the low end for some schottky's, but otherwise you could leave it plugged into reverse polarity permanently with no damage.

[pinkjimiphoton]

thanks for the explanation greg, i think i get it. so, i should leave it in there to prove when customers use the wrong power and then complain at me, then? 

[GGBB]

forgive my stupidity, but isn't the idea of having the second one after the series diode to prevent catastrophic failure if the first diode smokes?

I always thought - and assumed that's what you did - the anti-parallel diode went before the series diode. And if done properly had its own series CLR, or the power in had a CLR before the two diodes. (And the CLR should be big enough and high enough power to restrict current to a certain expected value and not burn up while doing so.)

If it's after, then it looks like a full-wave rectifier for a center-tapped AC supply, but it actually isn't since circuit ground is not at the tap (0V) but at the negative power in (V-):



I'm not entirely sure what kind of mayhem this might cause, if any for AC power. I think it's harmless for DC. For a non-center-tapped AC supply rectifier you'd need two more diodes:



But then your circuit ground is 1 Vf drop away from both V+ and V-. I can only assume that's not good, and even with DC would probably cause some problems.

And if you've got other pedals in the mix sharing either the power or the signal ground ... I don't want to think about that.

Dropping the parallel diode and using just the series diode is a whole lot easier and safer IMO.

[Rob Strand]

No, it only forms a half-wave rectifier, from the series diode. The reverse bias parallel diode, if it's before the series diode, is just a short circuit and will burn up on an AC supply (potentially causing other damage), or if it comes after, it cannot conduct because the series diode is then reverse biased, so there's no current path.

Small transformers take some effort kill.    The designs are usually limited by regulation rather than heat removal.   Like if you left it on for a day (or days) with a short eventually you will blow the transformer.

The thing about modern plugpacks and small transformer is they have a non-resettable thermal fuse. Once you over heat the transformer and trip the fuse you have to throw out the device.   That was a pretty common scenario for the chargers for rechargeable drills with NiCd cells.  The cells would fail short and then take out the charger.

The other risk these days for parallel diode protection is if someone plugs in a high-current switchmode.   Those simply weren't around years ago but are now common.    You can't rely on the switcmode's current limit.

So maybe that's a good enough reason to change to series diodes.  For parallel protection you really need a resistor or a PTC.   There's also modern ideas like MOSFET protection which have very low voltage drop,

http://www.ti.com/lit/an/slva139/slva139.pdf

It's desirable to add a zener to the MOSFET like this.  (The fuse and AC TVS are optional, these perform crude overvoltage protection.)
http://electricalnote.com/wp-content/uploads/2015/09/Reversed-Polarity-protector-With-MOSFET-300x153.jpg

You might find BJT versions of these reverse protection circuits.   They kind of work but they have a flaw.  If the reverse voltage causes the BE function to conduct (somewhere between 5V and 10V) then the circuit starts applying negative voltage to the circuit, which is undesirable, so they aren't recommended for uncontrolled abuse.