Polarity protection using MOSFETs... question for R.G. maybe?

[earthtonesaudio]

This article:
http://geofex.com/Article_Folders/mosswitch/mosswitch.htm
Briefly mentions this:

...use four MOSFETS in a bridge to make effects insensitive to battery polarity...

I searched, and several posts discuss this idea, but no schematic.
Seems like a really useful trick.  Just specify the max voltage the pedal can take, then don't worry about the other details.

I'd like to put this in my pedals, as I'm constantly salvaging wall warts from strange places, some are AC, some are DC, and even though I check them with a meter before plugging them in, I want to be extra careful in case a friend comes over and plugs in a pedal with the *wrong* power.


So... where's the schematic? 

Thanks,
Alex

[R.G.]

So is this for pedals for your own personal use, or to sell?

[slacker]

You can do this using a full wave bridge rectifier feed it AC or DC of either polarity and you get positive DC out of it. You'll have to add some filtering and maybe some regulation and take the voltage drop caused by the diodes into account when picking your input voltages but it does the job.

[xshredx]

R.G.... If possible I would really appreciate the info too...   I just got 2 pedals from a friend some weeks ago that were fried when he did lend them to one of his friends... turns out that dude used his line6 adapter...   point is: this is not the first time that I repaired fried pedals, and I'm really thinking that it would be absolutely useful to have protection in the pedals I make myself. I make pedals for myself and for friends, but we do lend out pedals often to each other...

edit: slacker: that's good advice (and I already made a pedal like this, inspired by the power supply on the ucapps.de projects), but still I would like to know about the four mosfets method. Because using a full wave rectifier with diodes will lead to a voltage drop, and on R.G.'s page there is written that the four mosfets method works almost like a bridge rectifier with diodes, but 'with almost no voltage loss'

[earthtonesaudio]

So is this for pedals for your own personal use, or to sell?

Eventually both, unless there's a good reason not to (i.e. if it messes with the ground voltage level in multiple pedal setups, or if it's someone's intellectual property, etc).

[earthtonesaudio]

Two versions:
http://www.thetaeng.com/FETBridge.htm

[Ben N]

Cool. I wonder, could this be done with a single cmos chip, say something like a 4007
http://www.standardics.nxp.com/products/hef/datasheet/hef4007ub.pdf (40 cents at Mouser)?

[earthtonesaudio]

I don't think so.  Something about the internal connections makes me think it wouldn't work.

Maybe a dual-complementary pair with separate everything on one chip.  There's gotta be something like that available.

[Ben N]

Look at the bridge schematic, and then at the 4007 internal schematic in the datasheet. I believe that in the 4007 the gates of the complementary pairs are connected, while the sources of the N-channels are connected together, and the sources of the P-channels are connected together. That would be the same as in the bridge, non, mon ami?

[earthtonesaudio]

I was referring to the substrate connections.  On page 8 under "Note" it talks about maintaining electrical isolation between transistors and the substrate by making pin 14 most positive and pin 7 most negative.  The way the bridge rectifier is drawn, those pins will be the DC output connections, which means they won't be the largest voltages on the chip.

I don't know if that means it won't work, just that the datasheet warns against it.

[Ben N]

Wh, as the DC connections, wont those be the largest voltages?

[earthtonesaudio]

The input connections will be the largest voltages.  There's always some loss through the FETs, so the DC output will be just slightly less than the DC in.  If you put AC in, the DC out might be higher than the RMS AC in, but not higher than the peak AC in.

If you have some 4007s lying around, I would still try it and see if it works.  I don't have any, so I'll probably order some discrete transistors just to be on the safe side.

[Ice-9]

In R.G's original article he suggests putting a zenner diode between the gate and source. There is no mention of what voltage zenner should be used. I take it its not really critical but would a 5.1 zenner surfice.

[Ben N]

It sure would be nice if RG or some other PWAK (Person With Actual Knowledge) would chime in here--seems to me that a CMOS 1-chip solution is what would make this, as Owl would say, a Really Practical Idea.

[earthtonesaudio]

In R.G's original article he suggests putting a zenner diode between the gate and source. There is no mention of what voltage zenner should be used. I take it its not really critical but would a 5.1 zenner surfice.

Less than Vgs of the FETs you use, more than the supply voltage.  They're for over-voltage protection, you don't want them on all the time, or you'll waste a lot of power (which would completely defeat the purpose of the FET bridge). 

For a FET with Vgs of 20V and a power supply of 9V, use a 15V Zener.


And this is only for the schematics with the gate resistors in them.  The bridge with 4 FETs and no other components should NOT get Zeners, and should NOT be given high voltages.  9V is okay.

[earthtonesaudio]

1-chip, built-in Zener protection:
http://www.ortodoxism.ro/datasheets/HitachiSemiconductor/mXwvqtr.pdf

[Ben N]

OK, Alex--nice job. I like the SIP package, too. Now, where can you get them, and at what price? Not at Mouser

[earthtonesaudio]

Eh, I looked on Mouser and Digikey, and it seems like all they have are surface mount things.  If you can find the VQ3001J/VQ3001P, that would be a good bet too (but discontinued). 

[Sir H C]

The problem with one chip solutions is that the ESD is common and can have issues, and with everybody on the same substrate, you can get latchup issues if you are not extremely careful in how you put it all together.  Also, somewhere here there was a thread on the 4 FET/Diode solution, problem is that ground is now raised, and that might be an issue.

[earthtonesaudio]

The problem with one chip solutions is that the ESD is common and can have issues, and with everybody on the same substrate, you can get latchup issues if you are not extremely careful in how you put it all together.  Also, somewhere here there was a thread on the 4 FET/Diode solution, problem is that ground is now raised, and that might be an issue.

Now that I've thought about it a bit, 4 discrete FETs wouldn't take up much more space than a 12 or 14 pin IC.  Plus you can arrange them to your liking.

Found that thread, and this:

The very low voltage drop in the MOSFETs makes the ground voltage differences moot.