Grounding stuff to Vref

[slashandburn]

Ive gone and done it again, "accidentally" put a wire to Vref rather than ground. This time my black wire coming off my battery clip.

Im assuming this is at least one of the reasons my circuit isnt working this time. Voltage across the battery is 1.9v in circuit, 7.5v if i measure it  removed from the clip. Its a basic tube screamer circuit feeding a 386 power amp.

Anyway, ill sort this (seemingly recurring) error out in the morning, before I likely find a few more. I was just wondering more than anything, after talking about it being okay to ground some stuff to vref in another thread (specifically the negative feedback loop bit in a non inverting opamp) Is there anything else I can "get away with" grounding to Vref?   Obviously what Ive done in this case isnt going to work. Im just curious more than anything, why it might be beneficial or desired to ground something there in some cases.

Anyway, just some late night pondering. I figured Id put it out there for something to read in the morning before I go blindly re-soldering things until it makes noises.

Cheers
Iain

[PRR]

Your battery is used-up. Get another, fresher, battery.

Your other question is too broad to even start on.

[anotherjim]

If you want to DC couple stages, you probably Have to use Vref as ground for such as interstage volume control & clipping diodes.
But the times you have to DC couple are few (LFO modulation mostly). AC coupling the audio gives an opportunity to form a high pass filter to control low end mud.

[Seljer]

That big filter cap that is typically present at the voltage divider that's generating Vref basically looks like a short to ground for AC signals.
But placing another cap in the feedback setup of the opamp (and in that case you can run it to ground directly) effectively removes any gain for DC signals which might help against surprises with DC offset issues, as well as the high pass effect for EQing that anotherjim mentions

[slashandburn]

Okay thanks for that! I'll get brushing up on ac and dc coupling, then. And get new 9v batteries!  (perhaps my cheap DMM isnt very.accurate, or those 2 for a buck 9v's are absolute junk).

Sorry for all the protracted questions, still learning here as you can see.

Genuinely appreciate your time and feedback, folks!

Cheers
Iain

[teemuk]

That big filter cap that is typically present at the voltage divider that's generating Vref basically looks like a short to ground for AC signals.

This. And the same applies to power supply rails too: As long as there is enough (filtering) capacitance to bypass AC signals to ground throughout guitar's effective bandwidth all these nodes are also low enough in impedance for AC signals to practically resemble a grounded connection. Keep in mind though that capactive coupling is usually mandatory to isolate different DC potentials from each other. So, you can use all these low impedance nodes as "floating" ground reference for AC signals but not for DC signals.

Different matter is how "noisy" such reference points will be, because any noise signal in them will also appear in the gain stage's output signal. The very same phenomoneon can sometimes be exploited for "hum-bucking" too.

[GibsonGM]

All that, plus this (maybe) easy to keep in mind "kinda general rule":

Vref is like "AC ground" and has nothing to do with "ground"

Ground is DC ground, 99.5% of the time zero volts.

Your SIGNAL operates around Vref - it's the point the audio thinks of as zero for its up and down swings

Ground is where the transistor, opamp etc. 'see' zero volts DC, which THEY reference to do their amplifying jobs.  The two things happening are separate, in a way...the device needs DC to run, and the signal needs a point of reference to allow it to go up AND down in value. 

A little confusing when you're new, I know.  It will start to make sense as you build, and sim stuff on the computer and look at voltages at different nodes, over some time.     From the question, you now see the two points aren't interchangeable altho sometimes it seems that way

[slashandburn]

Your SIGNAL operates around Vref - it's the point the audio thinks of as zero for its up and down swings

And there it is! Awesome, a lot more makes sense now. It might have been a lot easier had i just asked "what is Vref?",if I'd known thats what I needed to know.   Nice one thanks for that Mike! It really cleared up a lot for me. Feel like kicking myself a bit now. Really should've known that!

Cheers
Iain

[GibsonGM]

Glad it makes sense now, Ian!  I also had to really stop and think about that when it was a new concept, and I would get upset when my Vref was maybe 3.8V instead of 4.5, things like that, ha ha!      J

It just biases the amplifying device (opamp...) to 'think' that 1/2 your supply voltage is its signal ground, so it can move from its OWN ground (zero crossing point) above that level AND below it. 

Without Vref, we'd get rectification...the signal would rise from zero to max, then back to zero, and be clipped right off, losing the negative part of the audio.    I think doing this is one of the coolest parts of audio electronics!       They do this with tubes and transistors too, just call it 'setting the bias point' or 'biasing'.   

[karbomusic]

Without Vref, we'd get rectification...the signal would rise from zero to max, then back to zero, and be clipped right off, losing the negative part of the audio.

Which we have to do in our world of unipolar power supplies instead of a V+/V- supply where there would be no tricking of the center point required for a bipolar op amp. Just noting so that it makes even more sense to the OP.

[Tony Forestiere]

It used to drive me batty converting the older Anderton stuff for single battery operation. Mike just shed the veil from my eyes. Simply and elegantly described. Thanks.

[GibsonGM]

Without Vref, we'd get rectification...the signal would rise from zero to max, then back to zero, and be clipped right off, losing the negative part of the audio.

Which we have to do in our world of unipolar power supplies instead of a V+/V- supply where there would be no tricking of the center point required for a bipolar op amp. Just noting so that it makes even more sense to the OP.

+1  Bipolar supplies make stuff so easy on this end, but end up often being a pain on the front end, LOL.   I don't use them due to lack of interest in making things complicated in the power supply.   So now we're stuck with Vref...that little "trick" to "fool" the device into doing what you want.  YOU get to be the boss!! (ha ha)

I spent a couple of years (self-taught and stumbled, still do at times) trying to "get it".  I'm happy when I see someone fairly new grab this concept so quickly!

Glad you got something out of this too, Tony!   Without the help of a dozen others on this forum when I was trying to learn about this stuff, I'd have the darn veil over my eyes STILL.   

Now we can mess around with how to bias Vref to maybe cause some kind of clipping on one side of the signal - assymetrical clipping, provided whatever device you're using has a pleasant clipping characteristic...and away ya go!

[antonis]

Now we can mess around with how to bias Vref to maybe cause some kind of clipping on one side of the signal - assymetrical clipping, provided whatever device you're using has a pleasant clipping characteristic...and away ya go!

Shouldn't be enough just to bias Vref in a level different than Vcc/2..?? 

[teemuk]

Which we have to do in our world of unipolar power supplies instead of a V+/V- supply...

I believe both (unipolar and bipolar) are conceptually the same: We provide the opamp two power supplys, one more positive than the other, and bias its inputs to an arbitrary DC reference point between the voltage potentials of each power supply. Usually we bias exactly "in between" but offset bias in practice won't matter any more than offset bias in general. It can be even preferable. However, such offset must usually be isolated from other stages.

Anyway, with unipolar power supply we might have one power supply voltage in the region of, say, 30 volts, and to cope with usual limits of the opamp we may bias the other 30 volts below that, conveniently that happens to be 0V, our ground. Then we generate an external bias voltage supply that works as dc offset reference point for our opamp circuit. Usually to provide 15 volts, which is half of 30V-0V. This supply must be able to efficiently sink all current you "dump" into it. With small signal opamps this isn't anything unreasonable and in case of pedals a compromise between efficient sinking and battery life is made. The DC offset should again in practice be isolated from other stages that operate around different dc offset point.

Alternatively we could ground the positive supply rail, provide -30V to the negative and bias to -15V. Or we could provide 130V to the positive power supply pin, 100V to the negative and bias to 115V. And so on... Concept remains the same. 15 volts, by the way, is just an examplatory value. The supply rail voltages, and their voltage difference, can naturally be any value that still keeps the opamp operational.

With bipolar power supply we continue the same concept: Two power supply rails, one more positive than the other and dc offset biases somewhere in between. Ground reference now becomes handy dc offset reference point. In theory we now should be able to directly couple gain stages without DC offset's interferring with each other, in practice the gain stages can generate dc offset on their own, and to avoid coupling this offset to other stages we need capacitive coupling.

[ashcat_lt]

As a matter of fact, I've seen at least one schematic that referred to the ends of the battery as +/-4.5V, and the Vref (generated by voltage divider as normal) was called 0V.

There's ultimately no real reason that the bottom of the battery needs to be the ground point for anything.  You could connect the chassis to Vref if you really wanted to and save a couple capacitors, but the pedal might not play well with more traditional pedals on the same power supply.

[karbomusic]

I believe both (unipolar and bipolar) are conceptually the same

That makes perfect sense. So wouldn't the fact (as ashcat may be saying also) is that since for pedals -4.5 (conceptually and typically) = ground, thusly, we have to make the adjustment to create the middle point? Whereas with a true bipolar supply ground is a separate thing so we don't have the problem by default. Or is still some midpoint that needs to be set, seems (from a newb perspective) that it wouldn't need to by virtue of the fact that GND isn't part of the - "leg"

[GibsonGM]

Of course you guys are all correct.   I've been trying to keep this thread a bit simple (introductory) so as to nail the point home solidly!  That Vref is just a point for the signal to ride around, and isn't actually "ground" as pertains to the sort of common agreement that builders/amp makers and the like have come to.   

You can call 8.2V ground if you want, in a 13V supply, work around that, and be fine.    But question: would you want or need to?  For a Tube Screamer?     It seems to me that we can overcomplicate things pretty fast if we start drifting into the finer points (maybe trigger a "what is ground" argument, LOL).

On a 9V supply, 4.5V typically (but not always altho for this discussion is convenient) is what could be called the Signal Bias Point, or Vref as we've been discussing.    Since your jacks, pots, enclosure etc. will be grounding to 0V, it's not 'Chassis Ground'.    I don't even use the term "Ground" with it, normally, as it seems to be a bad habit.  It's just the reference where the AC signal crosses zero - it's where you set your device to work from.  Reference is one thing, ground another.

What works for me is to just keep the ol' Vref associated with AC ONLY, and circuit ground is 0 volts (for 99.9% of the things that I do, anyway, YMMV).   If a noob can understand this, inside and out, THEN might be the time to start pondering the "Other" ways the same thing can be achieved.   Separating the world of DC from AC can be a major PITA is if you're new (ask me how I know).   Bad enough that in reality, you're making the AC ride on a DC level that you later block with caps.



<<Antonis, you're correct - by shifting Vref, you shift the reference point and therefore can push the signal into 'territory where there's no more amplification possible', and cause clipping, on either pos. or negative side (or both >>

[antonis]

I don't even use the term "Ground" with it, normally, as it seems to be a bad habit. 

Consequently, using the term "Virtual Ground" should be a nasty habit..

[GibsonGM]

I think so!   I never call it that, but do get 'sucked into it' sometimes! ha ha.   Just reference voltage.  Sometimes you see "rail splitter" applied to this, which is (to me) a much more accurate and easy to understand term!   

I want to pass a plank out of a job site thru a window...I have to pick it up 3 feet off the ground and move it forward to do so.  That is my reference height    That's all we're really doing here. 

[duck_arse]

(open the window first)