explaining bipolar supplys

[Eddododo]

I was daydreaming about some circuit ideas and I just realized, or so i think, that I think i get it ( i always avoided them, im a nooby-booby)

compared to a 0-9v supply, would i treat -9v as i did ground previously, and the ground [of the bipolar circuit] as i would have used Vbias?
are there any additional caveats? is there an advantage of +/-9v vs +18 ?          I was thinking about LFO signals as switching signals and a simplified approach to something i am puzzling seems that certain ideas would be simpler if the LFO 'dips' into negative voltage as opposed to going above and below a fixed positive threshold



anyway, ignoring the afterthought, am i right about the comparison of unipolar to bipolar?

Thanks,
Ed

[ashcat_lt]

All voltage is relative.  Usually inside our boxes we can pick about any point to call 0, and measure our other voltages from there.  In a typical 9V circuit with 4.5V bias you could just as easily call the bias point "ground" and refer to the supply as +-4.5V.  Some circuits call the top of the battery 0, and the bottom -9V.  +-9V is exactly the same as 18V until you go to interface with another unit.  Then you have to consider where your voltages are with respect to theirs.

[GibsonGM]

You're on the right track.  I think of "0" as ground in a bipolar circuit.   My signal crosses zero, to go + something and - something.   

Yes, it's the same as riding around a 4.5V bias.   But it actually IS "-9V" as REFERENCED TO GROUND, or zero volts.   There is this very slight difference between the two systems...a sine wave WILL reproduce normally in a bipolar system, crossing zero and all (as long as it is of small enough amplitude for your device to handle, of course).   

When you elevate your signal (4.5V bias...), it does not really cross ground - we are fooling it into thinking it does!  And removing that bias at the output.

It's just a matter of a little mental juggling, which you'll get used to after looking at it for a little bit       Just remember your signal riding on a 4.5V level ABOVE ZERO if you use bias, and crossing zero if you use bipolar...

[Thecomedian]

isn't the idea behind it that you can use a greater voltage swing while staying at 9v + or -, for an IC that can't take 18v, but you'd like more headroom?

[induction]

IC's don't know the difference between +/- 9V and +18V/ground. It's 18V either way. As ashcat said, all voltage is relative.

[Thecomedian]

I thought it was all relative to ground.

Suppose there's an AC signal of 18v. you bias at 0v, and ground is at 0v. with a unipolar supply of DC9v, the AC signal sends the IC up to +9v, and then the chip is cutoff for the negative half of the cycle.

With a bipolar supply, the chip can go -9v. In both cases, the distance in voltage from ground is only 9v, but you can deal with a full 18v signal, as opposed to biasing at 9v and having unipolar 18v supply.

That seems to sound right to me.

[ashcat_lt]

I thought it was all relative to ground.

Suppose there's an AC signal of 18v. you bias at 0v, and ground is at 0v. with a unipolar supply of DC9v, the AC signal sends the IC up to +9v, and then the chip is cutoff for the negative half of the cycle.

With a bipolar supply, the chip can go -9v. In both cases, the distance in voltage from ground is only 9v, but you can deal with a full 18v signal, as opposed to biasing at 9v and having unipolar 18v supply.

That seems to sound right to me.

Honestly, that twists around so many corners I'm not really sure where I am anymore.  

Voltage is relative.  Assuming that all of your devices can actually swing all the way to the rails, you can only get a signal swing at most equal to the distance between the top rail and the bottom rail.  And that's a period.  Doesn't matter what you call ground.  If you want an AC signal to pass your device symmetrically, you must reference (bias) that signal to exactly half way between the rails.  You can do the biasing externally or internally.

[Thecomedian]

Im just working this as a theory.

the brown line represents the entire voltage the IC sees at that stage of AC signal development. When biased at halfway, an 18v AC signal will send the DIFFERENCE of voltage in the IC from 0 to 18v.

When biased at 0, the AC voltage will only send the DIFFERENCE of voltage in the IC from 0 to 9v, and will be rectified or cut off in the negative half of the cycle.

Add a bipolar supply, and the DIFFERENCE will be +9/-9, instead of +18/0. If the chip can take a maximum of +/-12 to 15v, but not +/-18v, the bipolar supply keeps the black/ground potential difference from rail to 9v, + or -.

[induction]

You have mostly the right idea about biasing (few op-amps are rail-to-rail, but that's a detail).

The point you're not getting is that the maximum voltage rating of a chip refers to voltage differences, not absolute voltages. (Technically, it refers to half-voltages, usually. In other words if an op-amp (the TL071, for example) is rated for 18V, that means it accepts up to 36V difference between V++ and V--). So if an op-amp can take +/-9V, it can take +18V/ground. There is no difference to the chip. There is no difference, period. The chip doesn't know or care where this arbitrary 'ground' thing is. You could feed it +1,000,018V/+1,000,000V and the chip wouldn't know.

What a chip knows is 1) the difference between the applied V++ and V--, and 2) the applied bias.

So you get no extra headroom with +/-9V than you get with +18V/ground, assuming you bias correctly.

[ashcat_lt]

I think you're getting confused by trying to cram too many concepts into one thought.

Look at your opamp.  

It has VCC+ and VCC- inputs.  It has no ground or "Vzero" input.  When you're talking about maximum supply voltage all that matters is the difference between VCC+ and VCC-.  Let's say it will take +/-18V, that's a total of 18V - (-18V) = 36V.  If you have a "unipolar" the you connect what you're calling 18V to the VCC+ terminal and what you're calling 0V goes to VCC-.  The total voltage the opamp sees is 18V - 0V = 18V.  That's well within the safety range of the IC.  Even if it can only handle +/-12V you're fine.  I would submit that these specs should be given in total rail voltage, rather than confusing folks with this bipolar thing.  

Honestly the difference between 18V "unipolar" and +/- 9V "bipolar" is that there are three wires coming from the "bipolar" supply -  a negative rail, a positive rail, and one half-way between.  You could take your standard "unipolar" circuit, connect the -9V to the "ground" symbols, the +9V where the positive rail goes, and then remove the voltage divider creating the bias voltage and connect the 0V center tap to anything that comes off of that Vbias point and it would work exactly the same as if you just plugged in the "unipolar" supply and left the bias divider intact.  Except if your next device has it's chassis connected to something with a real potential difference compared to this one.  For example, if your next pedal was a standard 9V pedal, and you used this exact same power supply and connected the 0V center tap to the ground, you'd have a difference of 9V between the two chassis.  You'd be shorting out the bottom half of your supply via loop through the PSU and cable shields.  You'd probably let the smoke out of a component or two along the way.

Now, the audio comes into your box referenced against the cable shield.  And yes if you tie that cable shield to the bottom rail of the supply, and don't later bias the signal up somewhere, then it will not be able to pass the bottom half of the waveform through the opamp.  So, if you have a supply which gives you a bottom rail, a top rail, and a middle point, and you tie the middle point to the cable shield, then you don't have to add any components inside your box to bias the audio signal to the middle point.  But those components were really just moved over to the power supply.  You don't have to worry about them, but whoever made the PSU does.

[GibsonGM]

To see this in action, one should sim up a circuit that uses a bipolar supply, and one that does not.  A simple opamp gain stage would do.  Then show the waveform before the output cap, which is a sine wave riding on a DC bias (thecomedian's first pic). 

With the bipolar supply, you get his second pic - an actual sine wave which has no DC 'mixed in' (theoretically). 

The DC must be blocked in the first case, but the operation of the IC will be about the same in all respects.


The real difference (if, as pointed out, there IS one) is that with the bias supply circuit, your sine wave does NOT cross zero.  That is nitpicking, I know...but is a very real difference.   The "+9/-9" will also give you a wider (2x) range of amplitude...the 4.5 bias obviously limits you quite a bit.   Those numbers are just there because we most commonly use 9V batteries/supplies, of course. 

It's helpful to keep in mind that you can add DC to a signal and take it back out again without altering it, which is exactly what you do when you apply bias.    You are just shifting levels around, but the audio signal stays exactly the same.

[ashcat_lt]

With the bipolar supply, you get his second pic - an actual sine wave which has no DC 'mixed in' (theoretically).  

The DC must be blocked in the first case, but the operation of the IC will be about the same in all respects.

What is this DC referenced to?

The real difference (if, as pointed out, there IS one) is that with the bias supply circuit, your sine wave does NOT cross zero.

This is only true if you insist on calling something other than the bias point zero.

There IS NO absolute voltage.  The closest we get is the earth ground of whatever part of the chain is actually tied to the mains wiring, but everything in fact works just fine without that.

[GibsonGM]

Yes, you are totally correct, Ashcat - however, in the real world we ARE constrained to some absolutes.  These pedals will, I presume, feed an amp??   Then, ground is zero volts.  Unless you've built an amp that uses something else as its reference?   

I appreciate the philosophy, and again, it's true... but unless we have some starting positions, we're not going to be doing much productive work!  If the OP doesn't understand what's going on, NOW is probably not the time to introduce the "there are no absolutes" thing, eh?  Just sayin'...

In THIS realm, you can reference ground (zero volts, earth ground....) or "some" bias voltage, generally taken as 1/2 your supply, 4.5, or what have you.   If you choose to call 10 volts {above what is normally taken as earth ground for the rest of us} *ground*, and 30 volts  becomes 20 volts, well - ok, have fun with it!  It WILL work, but - 

It would be difficult to communicate ideas or have any standardized equipment in that realm, wouldn't it?   You'd have a lot of errors, a lot of ruined stuff due to connecting it up with different levels...it's just not how it works....

No disrespect intended, just that we do have to take SOME things as a 'given'.  The sky is really black (without color), we know that, but since we experience it as blue we refer to it that way, there ya go....

[Eddododo]

Yes, you are totally correct, Ashcat - however, in the real world we ARE constrained to some absolutes.  These pedals will, I presume, feed an amp??   Then, ground is zero volts.  Unless you've built an amp that uses something else as its reference?   

I appreciate the philosophy, and again, it's true... but unless we have some starting positions, we're not going to be doing much productive work!  If the OP doesn't understand what's going on, NOW is probably not the time to introduce the "there are no absolutes" thing, eh?  Just sayin'...

In THIS realm, you can reference ground (zero volts, earth ground....) or "some" bias voltage, generally taken as 1/2 your supply, 4.5, or what have you.   If you choose to call 10 volts {above what is normally taken as earth ground for the rest of us} *ground*, and 30 volts  becomes 20 volts, well - ok, have fun with it!  It WILL work, but - 

It would be difficult to communicate ideas or have any standardized equipment in that realm, wouldn't it?   You'd have a lot of errors, a lot of ruined stuff due to connecting it up with different levels...it's just not how it works....

No disrespect intended, just that we do have to take SOME things as a 'given'.  The sky is really black (without color), we know that, but since we experience it as blue we refer to it that way, there ya go....

actually, fortunately enough, my brain is wired for stuff like this.. as soon as a got a kind-of 'yes' [honestly more like when i asked the question] i had the rest of the thread discussion in my head haha

my real problem is that ive approached electronics with this stupid building blocks approach because basically i am impatient  Im quite good at math and shifting around these kind of relationships in my head; long story short i just have not paid my dues by starting from square one even though i know better.

Im a clever ignoramus, basically

thanks for the vocabulary and the solid thinking points fellas- if y'all see me around here, dont be afraid to toss the confusing stuff at me all willy-nilly, what i REALLY lack is the vocabulary/idioms/basic stuff. though the latter has the obvious solution of me reading a little from the top

again, thanks!

[GibsonGM]

Glad to hear it, and that there isn't anything confusing, Eddododo....I really just wanted to highlight that in the stuff we do here, most everything IS referenced to a specific ground point.  Of course there are tons of ways you can set things up so that they would not be - Ashcat is totally correct - but if someone comes in here that's new and starts down that road, they may be in for quite a lot of trouble before the 'get it'.

Just like - when is a capacitor not a capacitor? When it's an inductor!  ha ha..

[Kesh]

One big difference between a 0 - 9V supply with a 4.5V bias, and a -4.5, 0, 4.5 bipolar supply, is in the latter all three pins are expected to have a low impedance.