Virtual ground connected to enclosure?

[armdnrdy]

As some of you know....

Dino (DIGI2T), me, and a cast of other characters are working on a MN3007 version of the Mutron Flanger.

We're getting close to building the prototype. (parts in the mail)

There is one thing that is bugging me a bit, so I turn to you great forum members for some knowledge so I may understand this anomaly.

The power supply section outputs 15 volts to "power" ground. There is also a "virtual ground" circuit that feeds the signal/delay sections and a few points in the LFO section.

The buffered virtual ground circuit sits at 1/2 VCC (7.5 volts) in relationship to power ground.



Now all of this is good with nothing out of the norm.......except that the virtual ground is connected to the sleeve of the input/output jacks thereby making a connection to the enclosure! (non isolated jacks)

This connection is depicted in the schematic, and verified with voltage readings by Dino in the actual unit.

Now this is the part I can't seem to wrap my mind around......7.5 volts on the sleeve and enclosure!

This means that 7.5 volts will be "sent" through the sleeve connection to your guitar and amplifier and any effect unit in between!

Can someone explain how this works?

Here is the schematic for reference:

[armdnrdy]

I just found some info in "Electronics for Guitarists" by Denton J. Dailey which outlines the mixing of virtual ground circuits and "standard" 0 volt ground circuits.
It seems as if there is a potential for short circuit and damage to the virtual ground op amp when sharing a common wall wart supply.

Check out pages 128-130:

http://books.google.com/books?id=HccWz4_ym7YC&pg=PA129&lpg=PA129&dq=virtual+ground+connected+to+input+jack&source=bl&ots=JcYgyTR6QJ&sig=ZslYhyjaOsIjeQYom8_d5qgfjzY&hl=en&sa=X&ei=jOs1Utv1B7DXiAKC04DQDg&ved=0CF0Q6AEwBQ#v=onepage&q=virtual%20ground%20connected%20to%20input%20jack&f=false

I was thinking that the original Mutron Flanger utilizes an AC power cord to a transformer. This transformer would provide isolation and the only return path for the virtual ground.

The build we are working on has the potential to share power with other effects.....so.....I believe that the input/output sleeves and the enclosure should be connected to "power ground" 0 volts.

Any thoughts?

[Minion]

I would think that the virtual ground at the input of the opamp (R73 and R74 in your first schematic) shouldn't be connected to the sleeve or the enclosure , it should only be used to bias the inputs of the opamps , the sleeve and enclosure should be connected to the Power ground not the virtual ground ......

[armdnrdy]

I would think that the virtual ground at the input of the opamp (R73 and R74 in your first schematic) shouldn't be connected to the sleeve or the enclosure , it should only be used to bias the inputs of the opamps , the sleeve and enclosure should be connected to the Power ground not the virtual ground ......

Yes....this is what I think as well.
As I stated....I believe that the original Mutron design "worked" because the transformer created an "isolated" return path for all aspects of the power supply that wasn't "shared" with any other effect.

[R.G.]

"Ground" is whatever you say it is - except you only get to call one net "ground".

Remember, there are three kinds of "ground" - reference ground, which tells your circuit what Real, True Zero Volts is; then shield ground, which keeps radiated nasty signals off your signal lines; and finally "sewer ground", which is what returns used electrons to the power supply.

Reference ground is needed so your signal can tell how the signal is moving. "Sewer ground" only happens when one combines the power supply return lines with the reference ground. This is an unfortunate side effect of trying to make single-sided power supplies. When you mix sewer ground and reference ground on the same wire, you can get ugly things happening. Sometimes you get away with it.

The use of single power supply circuits has confused issues so that many people thing one side of the power supply *is* reference ground. This is not necessarily true, and it leads to some conflicts, not least being the difference between attaching the negative side of the power supply to ground in one pedal, the positive side to ground in another pedal, then trying to connect both of these grounds with a piece of wire (i.e. the shield of a guitar cord) and slamming the power supply.

It is perfectly OK to call some voltage between V+ and V- on a unipolar power supply "reference ground", and attach it to the input jacks and enclosure. But you can NOT then successfully connect either of the power supply leads to "ground" either intentionally by internal wiring or unintentionally by external connections or paralleling something else to the power supply that attaches one or the other power supply  lead to this third-party-circuit's "ground", and then hooking the two ground together. This is just another version of the positive/negative ground tragedy.

[tubegeek]

This is just another version of the positive/negative ground tragedy.

Also known as "The tragedy of the commons."

[armdnrdy]

R.G.
Thank you for the reply! I was hoping that you would "magically" appear for this one.

It is perfectly OK to call some voltage between V+ and V- on a unipolar power supply "reference ground", and attach it to the input jacks and enclosure. But you can NOT then successfully connect either of the power supply leads to "ground" either intentionally by internal wiring or unintentionally by external connections or paralleling something else to the power supply that attaches one or the other power supply  lead to this third-party-circuit's "ground", and then hooking the two ground together. This is just another version of the positive/negative ground tragedy.

Just so I have a better understanding, why does connecting the rail splitter ground to the sleeve work (assuming it doesn't cause problems with some pedals) in this Musitronics design? Is it as I surmised...not sharing the power supply among multi effects?
The transformer providing some form of isolation?

Since the "clone" circuit might be powered by a supply supplying more than one pedal, I'll connect the sleeve/enclosure connections to power ground so this build will "play nice" with other pedals.

[R.G.]

Just so I have a better understanding, why does connecting the rail splitter ground to the sleeve work (assuming it doesn't cause problems with some pedals) in this Musitronics design? Is it as I surmised...not sharing the power supply among multi effects?
The transformer providing some form of isolation?

The transformer provides complete isolation. It's secondary is completely floating. That is, the transformer secondary and rectifier/filter/regulator generates a completely floating 15V supply. The rail splitter generates a signal ground that is halfway between the two sides of the power supply. Notice in the lower right corner of the schematic, there are two different symbols for "ground", signal ground and power ground. As long as these are not connected together with a wire, all is well. The closed-triangle symbol for power ground must always remain NOT connected to the three-bars signal ground symbol net, otherwise the the rail-splitter opamp IC11B has its output shorted to the negative power supply for the chip.

There is an inside and an outside to this circuit. The inside of the circuit uses the negative power supply rail (closed triangle) for both reference and power return/sewer. Presumably they did the wiring well on the boards so that this does not cause noise issues. The outside of the circuit is connected to the "signal ground", which performs the same function as a "Vbias" in many pedals. However, instead of tying the signal ground to the power supply minus side with a wire like most pedals do, the designer wired it to the chassis and only connected the "signal ground" to the minus power supply with a capacitor, so they could be a fixed DC voltage apart. It's a confusing way to set things up. Perhaps the idea that someone would ever try to use this without its own power supply never occurred to the designers.

It has the side effect that you cannot use the same power supply for this pedal and other pedals without fixing the grounds somehow.

Since the "clone" circuit might be powered by a supply supplying more than one pedal, I'll connect the sleeve/enclosure connections to power ground so this build will "play nice" with other pedals.

That won't work. What you have to do if you want to power this pedal and others with a single power supply, you have to
(1) Go through the schematic and everwhere you find the three-bars ground symbol, replace that with "Vbias". This step is to help keep your head straight.
(2) Check each place there is a "Vbias" from step 1 and see if it needs any work; at least two places do, those being the input and output jacks, which need pulldown resistors. There may be others, I haven't checked exhaustively.
(3) Hook the jack sleeves to the negative side of the power supply.

This in effect converts the power supply back to +V and combined sewer/signal ground, and renames the previous "signal ground" to "Vbias".  Now you can share the sewer/signal ground with other circuits.

[armdnrdy]

Thanks for the guidance R.G.

Here is the redraw with the (confusing) virtual ground symbols replaced with Vbias symbols, input/output jacks connected to the negative supply, input/output pull downs added, and new power supply depicted.



 

[R.G.]

Looking...

R37 isn't needed.

Places where a capacitor is connected to Vbias could be connected to ground/V- if you have noise issues with the finished product. This ought to be the same, but the more places connected to Vbias, the more it must be **pristinely** quiet and well bypassed. C2 and C8 are examples, but for different reasons. C2 is to bypass high RF, C8 bypasses current peaks from the envelop detector. Normal signal connections to Vbias don't matter all that much. This is a nit.

If your power supply is fed with 18Vdc, the negative side of the 18Vdc must NOT be shared with another pedal that can tie it to signal ground. This is one version of the "univibe problem".

I didn't catch any, but any place there's an electro cap that's not NP, it should have a DC voltage across it the correct way, not nearly 0V.

You're working hard on this.

[armdnrdy]

Thanks for the reply R.G.

Yes...I am working hard on this one....sometimes good things don't come easy!

After I posted the above redraw, I took a look at a few somewhat similar single supply circuits (Wow! say that fast 10 times) and I did see that there are other things I'll have to change. I noticed certain caps that probably want to go to ground, the feedback and rate pots, T5 trimmer, etc.

I was wondering if you had noticed the link that I posted earlier. The exact sleeve ground connection issue is outlined in this article and an unbelievably easy fix is given as well.

I'll post it again. Check out just above page 128-130:
What are your thought on this?

http://books.google.com/books?id=HccWz4_ym7YC&pg=PA129&lpg=PA129&dq=virtual+ground+connected+to+input+jack&source=bl&ots=JcYgyTR6QJ&sig=ZslYhyjaOsIjeQYom8_d5qgfjzY&hl=en&sa=X&ei=jOs1Utv1B7DXiAKC04DQDg&ved=0CF0Q6AEwBQ#v=onepage&q=virtual%20ground%20connected%20to%20input%20jack&f=false

[samhay]

If the enclosure is only connected to 'ground' at the jacks, then as long as the input and ouput are not DC coupled, the question become: whether (aside from potential noise issues; the elephant in the room) connecting the jack sleeves and pull-down resistors to 'sewer' ground is a good idea?
Is the plan to power this with DC and use a charge pump to generate the bi-polar supply?

[armdnrdy]

Sam,

The schematic posted shows the power supply that I intend to use.

The supply basically provides 15 volts to the existing circuit in a different fashion than the original. ( No bipolar supply)

[R.G.]

I was wondering if you had noticed the link that I posted earlier. The exact sleeve ground connection issue is outlined in this article and an unbelievably easy fix is given as well.

I'll post it again. Check out just above page 128-130:
What are your thought on this?

It's pretty much what I told you. In figure 3.51, he shows using the minus side of the power supply as ground in a single-power-supply pedal to the left, and using the negative side of the power supply as *both* the minus-side power for a bipolar circuit, and the rail-split middle voltage as the ground for the circuits inside the right-side effect. The only difference is he calls the internal virtual signal ground in the right-side circuit "ground" instead of "vbias".

He also notes the issue with offset grounds from the same single power supply in figure 3.50.

[samhay]

Sorry - big schematic, little phone.
One thing - if I'm reading it right, your 'ground' is a diode drop above the DC power supply 'ground'. Is this going to play nice with other pedals? I don't know.

[R.G.]

Right - that was the issue with the full wave rectifier that I alluded to here:

If your power supply is fed with 18Vdc, the negative side of the 18Vdc must NOT be shared with another pedal that can tie it to signal ground. This is one version of the "univibe problem".

[armdnrdy]

Right - that was the issue with the full wave rectifier that I alluded to here:

If your power supply is fed with 18Vdc, the negative side of the 18Vdc must NOT be shared with another pedal that can tie it to signal ground. This is one version of the "univibe problem".

I'm using 1N5819 Schottky diodes to lessen the voltage drop.

Can you point me in the direction to some other configuration that will solve that problem?

I could always drop the AC/rectifier option, or the DC option. I wanted to include the AC option for builders because a 12VAC wall transformer is much easier to source than an 18VDC. On that note, some people already have 18VDC on their pedal boards.

[samhay]

Not a great solution, but would a half-wave rectifier work?

[armdnrdy]

Not a great solution, but would a half-wave rectifier work?

It would with heavy filtering.

[samhay]

You're right - it's only going into a regulator. I guess if it's quiet enough, then it's good enough.